{ "cells": [ { "cell_type": "markdown", "id": "49017e4e", "metadata": {}, "source": [ "# Why do today's numbers look bad, and how long until I know the real result?\n", "\n", "**Part one** builds a synthetic paid-search account where I know the truth, shows the\n", "drop every dashboard shows on its most recent days, and then measures the thing that\n", "causes it: how long conversions take to arrive after the click.\n", "\n", "The measurement is a survival fit on the click cohort: every click that has not\n", "converted *yet* is a censored observation, not a failure. Off that one fit come two\n", "readouts: the share of a day's eventual conversions that has arrived so far (which\n", "tells you what to multiply by), and the number of days after which essentially every\n", "conversion is in (which tells you how long a reporting window has to be).\n", "\n", "**Part two** runs the same fit on your own data, locally." ] }, { "cell_type": "code", "execution_count": 1, "id": "45c98430", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:43:25.449908Z", "iopub.status.busy": "2026-10-24T17:43:25.449452Z", "iopub.status.idle": "2026-10-24T17:43:33.148152Z", "shell.execute_reply": "2026-10-24T17:43:33.144254Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "today = end of day 56; 420 clicks a day\n" ] } ], "source": [ "import json\n", "from pathlib import Path\n", "\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import pandas as pd\n", "from scipy.optimize import minimize\n", "from scipy.stats import chi2\n", "from scipy.optimize import brentq\n", "\n", "for d in (\"figures\", \"data\"):\n", " Path(d).mkdir(exist_ok=True)\n", "\n", "SEED = 20260906 # one seed; every number below regenerates from it\n", "rng = np.random.default_rng(SEED)\n", "\n", "PAPER, INK, MUTED, LINE = \"#fdfcf9\", \"#1f2124\", \"#565c66\", \"#e6e2da\"\n", "GREEN, BLUE, WARM = \"#0a8f5f\", \"#2a78d6\", \"#b06a2c\"\n", "FIGDIR = \"figures\"\n", "\n", "plt.rcParams.update({\n", " \"figure.facecolor\": PAPER, \"axes.facecolor\": PAPER, \"savefig.facecolor\": PAPER,\n", " \"axes.edgecolor\": LINE, \"axes.labelcolor\": INK, \"text.color\": INK,\n", " \"xtick.color\": MUTED, \"ytick.color\": MUTED, \"axes.grid\": True,\n", " \"grid.color\": LINE, \"grid.linewidth\": 0.8, \"font.size\": 11,\n", " \"axes.spines.top\": False, \"axes.spines.right\": False,\n", "})\n", "\n", "# ---- the account, and the truth we get to know because it is synthetic ----\n", "WINDOW_DAYS = 56 # eight weeks of click history; \"today\" is the end of day 56\n", "CLICKS_PER_DAY = 420 # a mid-market paid-search campaign\n", "TRUE_CURE = 0.062 # 6.2% of clicks ever convert. The other 93.8% never will.\n", "TRUE_K = 0.85 # Weibull shape on converter time-to-convert (k<1: front-loaded)\n", "TRUE_LAMBDA = 4.0 # Weibull scale, in days\n", "REFERENCE_HORIZONS = (1, 2, 3, 7, 14, 21)\n", "MIN_TRUST_FRACTION = 0.30 # below this share reported, the platform refuses to multiply\n", "LOOKBACK_QUANTILE = 0.99 # the lookback readout: the day by which 99% of converters converted\n", "N_BOOTSTRAP = 400\n", "print(f\"today = end of day {WINDOW_DAYS}; {CLICKS_PER_DAY} clicks a day\")" ] }, { "cell_type": "markdown", "id": "01710eda", "metadata": {}, "source": [ "## The account\n", "\n", "Eight weeks of clicks on one campaign. Each click either converts eventually or never\n", "does, and the ones that convert take a while about it. Both facts are set here, so\n", "every estimate below can be scored against them.\n", "\n", "A click that has not converted **yet** is not a click that failed. It is a click whose\n", "window is still open, and the whole method turns on keeping those two apart." ] }, { "cell_type": "code", "execution_count": 2, "id": "2e7d959c", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:43:33.156266Z", "iopub.status.busy": "2026-10-24T17:43:33.155623Z", "iopub.status.idle": "2026-10-24T17:43:33.202390Z", "shell.execute_reply": "2026-10-24T17:43:33.199295Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "23,520 clicks | 1,437 will convert eventually (6.11%)\n" ] } ], "source": [ "def simulate_clicks(n_days=WINDOW_DAYS, per_day=CLICKS_PER_DAY, cure=TRUE_CURE,\n", " k=TRUE_K, lam=TRUE_LAMBDA, seed=SEED):\n", " \"\"\"One row per click: the day it happened, whether it will EVER convert, and\n", " the delay it will take if it does.\n", "\n", " Nothing here knows about \"today\" — this is the account's whole future, which is\n", " exactly what a real analyst never gets to see.\n", " \"\"\"\n", " g = np.random.default_rng(seed)\n", " n = n_days * per_day\n", " click_day = np.repeat(np.arange(n_days), per_day).astype(float)\n", " click_day = click_day + g.random(n) # spread within the day\n", " will_convert = g.random(n) < cure\n", " # Weibull delay among converters, by inverse CDF\n", " u = g.random(n)\n", " delay = lam * (-np.log(1.0 - u)) ** (1.0 / k)\n", " delay = np.where(will_convert, delay, np.inf)\n", " return pd.DataFrame({\"click_day\": click_day, \"will_convert\": will_convert,\n", " \"delay_days\": delay,\n", " \"convert_day\": click_day + delay})\n", "\n", "\n", "clicks = simulate_clicks()\n", "eventual_converters = int(clicks.will_convert.sum())\n", "print(f\"{len(clicks):,} clicks | {eventual_converters:,} will convert eventually \"\n", " f\"({eventual_converters / len(clicks):.2%})\")" ] }, { "cell_type": "markdown", "id": "e3967775", "metadata": {}, "source": [ "## What the dashboard shows today\n", "\n", "Now cut the account at the end of day 56. A conversion is visible only if it has\n", "already happened. Clicks from three weeks ago have had three weeks to convert; clicks\n", "from yesterday have had one day.\n", "\n", "This is the shape every account manager has stared at on a Monday morning." ] }, { "cell_type": "code", "execution_count": 3, "id": "9e9b97b9", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:43:33.210786Z", "iopub.status.busy": "2026-10-24T17:43:33.207956Z", "iopub.status.idle": "2026-10-24T17:43:33.311660Z", "shell.execute_reply": "2026-10-24T17:43:33.310521Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " day_index clicks reported eventual reported_share\n", " 48 420 24 27 0.888889\n", " 49 420 20 25 0.800000\n", " 50 420 15 23 0.652174\n", " 51 420 15 25 0.600000\n", " 52 420 18 27 0.666667\n", " 53 420 11 19 0.578947\n", " 54 420 6 23 0.260870\n", " 55 420 3 16 0.187500\n" ] } ], "source": [ "TODAY = float(WINDOW_DAYS)\n", "\n", "def observe(df, today=TODAY):\n", " \"\"\"The account as of `today`: what has been reported, and what is still in flight.\n", "\n", " Returns the click cohort in survival-analysis shape — elapsed time and whether\n", " the conversion has been SEEN — plus the per-click-day reported counts.\n", " \"\"\"\n", " seen = df.convert_day <= today\n", " elapsed = np.where(seen, df.convert_day - df.click_day, today - df.click_day)\n", " return pd.DataFrame({\n", " \"click_day\": df.click_day,\n", " \"day_index\": df.click_day.astype(int),\n", " \"elapsed\": elapsed, # days from click to conversion, or to now\n", " \"observed\": seen, # True = converted and reported; False = censored\n", " \"will_convert\": df.will_convert,\n", " })\n", "\n", "\n", "cohort = observe(clicks)\n", "by_day = cohort.groupby(\"day_index\").agg(\n", " clicks=(\"observed\", \"size\"),\n", " reported=(\"observed\", \"sum\"),\n", " eventual=(\"will_convert\", \"sum\"),\n", ").reset_index()\n", "by_day[\"age_days\"] = TODAY - by_day.day_index - 0.5\n", "by_day[\"reported_share\"] = by_day.reported / by_day.eventual\n", "\n", "print(by_day.tail(8)[[\"day_index\", \"clicks\", \"reported\", \"eventual\", \"reported_share\"]]\n", " .to_string(index=False))" ] }, { "cell_type": "code", "execution_count": 4, "id": "77442fbc", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:43:33.351334Z", "iopub.status.busy": "2026-10-24T17:43:33.324724Z", "iopub.status.idle": "2026-10-24T17:43:34.817161Z", "shell.execute_reply": "2026-10-24T17:43:34.813265Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "last seven days: 88 conversions reported, 158 eventually — 55.7% of the truth\n" ] } ], "source": [ "def days_ago_axis(ax, marks):\n", " \"\"\"Label the day axis as time before today. The bars sit at the account's own day index,\n", " which means nothing to a reader; a reader knows yesterday, a week ago, three weeks ago.\n", " `marks` are (days before today, label); day index 55 is yesterday, so a mark d days back\n", " sits at WINDOW_DAYS - d. Older days stay on the left, yesterday on the right.\"\"\"\n", " ax.set_xticks([WINDOW_DAYS - d for d, _ in marks])\n", " ax.set_xticklabels([lab for _, lab in marks])\n", " ax.set_xlabel(\"when the clicks happened\")\n", "\n", "\n", "fig, ax = plt.subplots(figsize=(10, 4.6))\n", "ax.bar(by_day.day_index, by_day.eventual, color=LINE, width=0.82,\n", " label=\"what those clicks will eventually deliver\")\n", "ax.bar(by_day.day_index, by_day.reported, color=BLUE, width=0.82,\n", " label=\"conversions reported as of today\")\n", "days_ago_axis(ax, [(56, \"8 weeks ago\"), (42, \"6 weeks ago\"), (28, \"4 weeks ago\"),\n", " (14, \"2 weeks ago\"), (7, \"a week ago\"), (1, \"yesterday\")])\n", "ax.set_ylabel(\"conversions\")\n", "ax.set_title(\"The same clicks, twice: what has arrived, and what is still coming\")\n", "# the legend sits in a band cleared above the bars: at the lower left it printed over them\n", "ax.set_ylim(0, by_day.eventual.max() * 1.24)\n", "ax.legend(frameon=False, loc=\"upper center\", ncol=2)\n", "ax.set_xlim(-0.8, WINDOW_DAYS - 0.2)\n", "fig.tight_layout()\n", "fig.savefig(f\"{FIGDIR}/fig_dashboard_illusion.png\", dpi=170)\n", "plt.show()\n", "\n", "last7 = by_day.tail(7)\n", "print(f\"last seven days: {int(last7.reported.sum())} conversions reported, \"\n", " f\"{int(last7.eventual.sum())} eventually — \"\n", " f\"{last7.reported.sum() / last7.eventual.sum():.1%} of the truth\")" ] }, { "cell_type": "markdown", "id": "20574c17", "metadata": {}, "source": [ "The recent days are not worse. They are younger. Nothing about the campaign changed\n", "on day 50. The only thing that changed is how long those clicks have had.\n", "\n", "The rest of this notebook measures exactly how much of a day's eventual total has\n", "arrived at each age, so the two can be told apart." ] }, { "cell_type": "markdown", "id": "2ac6ebb8", "metadata": {}, "source": [ "## The fit: one curve, fitted with the unconverted clicks kept in\n", "\n", "Take every click and record two things: how long it has been since the click, and\n", "whether the conversion has been seen. A click that converted on day 3 contributes\n", "\"3 days, converted\". A click made yesterday that has not converted contributes\n", "\"1 day, not yet\", a **censored** observation, which carries real information (this\n", "click has already survived a day without converting) and must not be scored as a\n", "failure.\n", "\n", "Two curves come out of that cohort.\n", "\n", "**Kaplan-Meier** is the nonparametric one: no assumption about the shape of anything.\n", "It is the honest picture of the data and it is what I check the parametric fit\n", "against. What it estimates is the *unconditional* curve, which flattens out at the\n", "share of clicks that ever convert.\n", "\n", "**The mixture-cure model** is the parametric one. It splits the population in two on\n", "purpose: a share `c` that will eventually convert, and the rest that never will.\n", "Among the converters, the time-to-convert is a Weibull. The reason to bother is that\n", "the conversion-time curve you actually need is the **converter-conditional** one (of\n", "the conversions this day will ever get, what share has arrived?), and that curve is\n", "buried inside the unconditional one until the two components are separated." ] }, { "cell_type": "code", "execution_count": 5, "id": "fb00274a", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:43:34.836447Z", "iopub.status.busy": "2026-10-24T17:43:34.835838Z", "iopub.status.idle": "2026-10-24T17:43:35.205690Z", "shell.execute_reply": "2026-10-24T17:43:35.204445Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Kaplan-Meier flattens at S = 0.9383 -> 6.17% of clicks ever convert (truth: 6.20%)\n" ] } ], "source": [ "def kaplan_meier(elapsed, observed):\n", " \"\"\"Product-limit estimate of the unconditional survival curve, with Greenwood's\n", " running variance factor.\n", "\n", " Returns (times, S, var_factor). S(t) is the chance a click has NOT converted by\n", " elapsed time t. Because most clicks never convert at all, this curve flattens\n", " at 1 - c rather than falling to zero — the flat part is the never-converters.\n", " \"\"\"\n", " order = np.argsort(elapsed)\n", " t, e = np.asarray(elapsed)[order], np.asarray(observed)[order]\n", " times, S, vf = [0.0], [1.0], [0.0]\n", " n_at_risk = len(t)\n", " i = 0\n", " s_run, v_run = 1.0, 0.0\n", " while i < len(t):\n", " tt = t[i]\n", " j = i\n", " while j < len(t) and t[j] == tt:\n", " j += 1\n", " d = int(e[i:j].sum()) # conversions exactly at tt\n", " if d > 0 and n_at_risk > 0:\n", " s_run *= (n_at_risk - d) / n_at_risk\n", " if n_at_risk > d:\n", " v_run += d / (n_at_risk * (n_at_risk - d))\n", " times.append(tt); S.append(s_run); vf.append(v_run)\n", " n_at_risk -= (j - i)\n", " i = j\n", " return np.array(times), np.array(S), np.array(vf)\n", "\n", "\n", "def km_at(times, S, t):\n", " \"\"\"Step-function lookup: the KM estimate at elapsed time t.\"\"\"\n", " idx = np.searchsorted(times, t, side=\"right\") - 1\n", " return S[np.clip(idx, 0, len(S) - 1)]\n", "\n", "\n", "def greenwood_ci(times, S, vf, t, z=1.96):\n", " \"\"\"Log-log confidence interval on S(t). The transform keeps both ends inside\n", " (0, 1) however close the estimate sits to either boundary.\"\"\"\n", " idx = int(np.clip(np.searchsorted(times, t, side=\"right\") - 1, 0, len(S) - 1))\n", " s, v = S[idx], vf[idx]\n", " if not (0.0 < s < 1.0) or v <= 0:\n", " return s, s\n", " se = np.sqrt(v) / abs(np.log(s))\n", " lo = s ** np.exp(z * se)\n", " hi = s ** np.exp(-z * se)\n", " return float(lo), float(hi)\n", "\n", "\n", "km_t, km_S, km_vf = kaplan_meier(cohort.elapsed.values, cohort.observed.values)\n", "km_plateau = float(km_S[-1])\n", "print(f\"Kaplan-Meier flattens at S = {km_plateau:.4f} \"\n", " f\"-> {1 - km_plateau:.2%} of clicks ever convert (truth: {TRUE_CURE:.2%})\")" ] }, { "cell_type": "code", "execution_count": 6, "id": "8f136062", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:43:35.214810Z", "iopub.status.busy": "2026-10-24T17:43:35.214454Z", "iopub.status.idle": "2026-10-24T17:43:35.589409Z", "shell.execute_reply": "2026-10-24T17:43:35.588171Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "eventual-conversion share c = 0.0617 (truth 0.062)\n", "Weibull shape k = 0.856 (truth 0.85)\n", "Weibull scale lambda = 3.982 (truth 4.0)\n" ] } ], "source": [ "def cure_negloglik(params, elapsed, observed):\n", " \"\"\"Negative log-likelihood of the mixture-cure model.\n", "\n", " A converter seen at time t contributes c * f_W(t); a click still unconverted at\n", " time t contributes (1-c) + c * S_W(t) — it is either a never-converter, or a\n", " converter whose time has not come.\n", " \"\"\"\n", " logit_c, log_k, log_lam = params\n", " c = 1.0 / (1.0 + np.exp(-logit_c))\n", " k, lam = np.exp(log_k), np.exp(log_lam)\n", " t = np.maximum(np.asarray(elapsed, float), 1e-9)\n", " z = (t / lam) ** k\n", " log_f = np.log(k / lam) + (k - 1) * np.log(t / lam) - z # Weibull density\n", " S_w = np.exp(-z)\n", " ll_obs = np.log(c) + log_f\n", " ll_cen = np.log(np.clip((1 - c) + c * S_w, 1e-300, None))\n", " obs = np.asarray(observed, bool)\n", " return -(ll_obs[obs].sum() + ll_cen[~obs].sum())\n", "\n", "\n", "def fit_cure(elapsed, observed, start=None):\n", " \"\"\"Maximum likelihood on (c, k, lambda). Returns a dict of the three parameters.\"\"\"\n", " x0 = start if start is not None else [np.log(0.06 / 0.94), np.log(0.9), np.log(4.0)]\n", " res = minimize(cure_negloglik, x0, args=(elapsed, observed), method=\"Nelder-Mead\",\n", " options={\"maxiter\": 4000, \"xatol\": 1e-6, \"fatol\": 1e-6})\n", " logit_c, log_k, log_lam = res.x\n", " return {\"cure\": float(1 / (1 + np.exp(-logit_c))), \"k\": float(np.exp(log_k)),\n", " \"lam\": float(np.exp(log_lam)), \"negloglik\": float(res.fun),\n", " \"converged\": bool(res.success)}\n", "\n", "\n", "fit = fit_cure(cohort.elapsed.values, cohort.observed.values)\n", "print(f\"eventual-conversion share c = {fit['cure']:.4f} (truth {TRUE_CURE})\")\n", "print(f\"Weibull shape k = {fit['k']:.3f} (truth {TRUE_K})\")\n", "print(f\"Weibull scale lambda = {fit['lam']:.3f} (truth {TRUE_LAMBDA})\")" ] }, { "cell_type": "markdown", "id": "82688b55", "metadata": {}, "source": [ "### Is the split real, or is the model inventing it?\n", "\n", "The cure share is only estimable when the curve visibly flattens: the flat part *is*\n", "the never-converters. If the curve is still falling at the right-hand edge of the\n", "data, then \"everyone converts eventually, slowly\" and \"most never convert\" fit the\n", "data equally well, and the split is not identified.\n", "\n", "The test is a profile likelihood on `c`. Fix `c` at a grid of values, re-fit the\n", "Weibull at each, and read off the interval of `c` values the data cannot rule out. If\n", "that interval runs all the way to 1, the cure model is not supported and the honest\n", "move is to fall back to a single Weibull, whose lookback then reads long, and the post\n", "says so." ] }, { "cell_type": "code", "execution_count": 7, "id": "f378eb8c", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:43:35.600163Z", "iopub.status.busy": "2026-10-24T17:43:35.597640Z", "iopub.status.idle": "2026-10-24T17:44:00.643612Z", "shell.execute_reply": "2026-10-24T17:44:00.642428Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "profile-likelihood interval on c: [0.059, 0.065]\n", "cure share identified — the parametric fit is supported\n" ] } ], "source": [ "def profile_cure(elapsed, observed, grid=None, alpha=0.05, centre=None):\n", " \"\"\"Profile likelihood interval on the cure share c.\n", "\n", " The grid is fine near the estimate and coarse out towards 1. A uniform grid over\n", " the whole unit interval is coarser than the interval it is trying to measure on a\n", " cohort this size, and collapses to a single point — measured, not assumed.\n", " \"\"\"\n", " if grid is None:\n", " c0 = centre if centre is not None else 0.06\n", " near = np.linspace(max(c0 * 0.5, 0.002), min(c0 * 1.8, 0.999), 45)\n", " far = np.linspace(min(c0 * 1.9, 0.998), 0.999, 25)\n", " grid = np.unique(np.concatenate([near, far]))\n", " def prof_at(c):\n", " logit_c = np.log(c / (1 - c))\n", " def nll(p, logit_c=logit_c):\n", " return cure_negloglik([logit_c, p[0], p[1]], elapsed, observed)\n", " r = minimize(nll, [np.log(0.9), np.log(4.0)], method=\"Nelder-Mead\",\n", " options={\"maxiter\": 1500, \"xatol\": 1e-5, \"fatol\": 1e-5})\n", " return r.fun\n", "\n", " grid = np.asarray(grid, float)\n", " prof = np.array([prof_at(c) for c in grid])\n", " cut = prof.min() + chi2.ppf(1 - alpha, 1) / 2.0\n", " idx = np.where(prof <= cut)[0]\n", " lo, hi = float(grid[idx.min()]), float(grid[idx.max()])\n", " # The grid points inside the cut sit up to one grid step short of the true edges, which\n", " # narrows the interval (about 0.1 points a side on the worked cohort). So each edge is\n", " # found between its last point inside and its first point outside.\n", " if idx.min() > 0:\n", " lo = float(brentq(lambda c: prof_at(c) - cut, grid[idx.min() - 1], grid[idx.min()], xtol=1e-5))\n", " if idx.max() < len(grid) - 1:\n", " hi = float(brentq(lambda c: prof_at(c) - cut, grid[idx.max()], grid[idx.max() + 1], xtol=1e-5))\n", " return lo, hi, bool(hi < 0.999)\n", "\n", "\n", "cure_lo, cure_hi, cure_identified = profile_cure(cohort.elapsed.values,\n", " cohort.observed.values,\n", " centre=fit[\"cure\"])\n", "print(f\"profile-likelihood interval on c: [{cure_lo:.3f}, {cure_hi:.3f}]\")\n", "print(\"cure share identified — the parametric fit is supported\" if cure_identified\n", " else \"cure share NOT identified — fall back to a single Weibull\")" ] }, { "cell_type": "markdown", "id": "3f2627f8", "metadata": {}, "source": [ "## Readout one: what share of a day's conversions has arrived\n", "\n", "Among clicks that will convert, the share that has converted by elapsed day *N* is\n", "\n", "$$F(N) = 1 - S_W(N) = 1 - e^{-(N/\\lambda)^k}$$\n", "\n", "That is the number the dashboard needs. Multiply a young day's reported conversions\n", "by $1/F(N)$ and you have an estimate of what that day will finally show.\n", "\n", "The multiplier is honest only while $F(N)$ is large enough to invert. At $F=0.30$ the\n", "multiplier is already 3.3, and every bit of noise in $F$ is amplified by the same\n", "factor. Below that the platform refuses to give a number at all, which is a better\n", "answer than a confident one." ] }, { "cell_type": "code", "execution_count": 8, "id": "44ef9405", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:44:00.680655Z", "iopub.status.busy": "2026-10-24T17:44:00.680198Z", "iopub.status.idle": "2026-10-24T17:48:09.077601Z", "shell.execute_reply": "2026-10-24T17:48:09.074203Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "400 bootstrap re-fits done\n", " days_since_click reported_fraction reported_fraction_lo reported_fraction_hi lag_factor lag_factor_lo lag_factor_hi trustworthy truth_fraction\n", " 1 0.2640 0.2433 0.2821 3.787 3.545 4.110 False 0.2649\n", " 2 0.4258 0.4035 0.4460 2.349 2.242 2.479 True 0.4258\n", " 3 0.5438 0.5220 0.5649 1.839 1.770 1.916 True 0.5430\n", " 7 0.8022 0.7844 0.8201 1.247 1.219 1.275 True 0.7999\n", " 14 0.9468 0.9360 0.9568 1.056 1.045 1.068 True 0.9450\n", " 21 0.9842 0.9788 0.9890 1.016 1.011 1.022 True 0.9833\n" ] } ], "source": [ "def reported_fraction(N, k, lam):\n", " \"\"\"Converter-conditional share reported by elapsed day N.\"\"\"\n", " return 1.0 - np.exp(-((np.asarray(N, float) / lam) ** k))\n", "\n", "\n", "def lag_factor(N, k, lam):\n", " \"\"\"The multiplier on a young day's reported conversions.\"\"\"\n", " f = reported_fraction(N, k, lam)\n", " return np.where(f > 0, 1.0 / np.maximum(f, 1e-12), np.inf)\n", "\n", "\n", "def lookback_days(k, lam, q=LOOKBACK_QUANTILE):\n", " \"\"\"Readout two: the elapsed day by which share q of converters have converted.\"\"\"\n", " return lam * (-np.log(1.0 - q)) ** (1.0 / k)\n", "\n", "\n", "def bootstrap_fit(elapsed, observed, B=N_BOOTSTRAP, seed=SEED):\n", " \"\"\"Resample CLICKS with replacement and re-fit. The resampling unit is the click,\n", " so censoring comes along for free.\"\"\"\n", " g = np.random.default_rng(seed)\n", " n = len(elapsed)\n", " rows = []\n", " start = [np.log(fit[\"cure\"] / (1 - fit[\"cure\"])), np.log(fit[\"k\"]), np.log(fit[\"lam\"])]\n", " for _ in range(B):\n", " idx = g.integers(0, n, n)\n", " f = fit_cure(np.asarray(elapsed)[idx], np.asarray(observed)[idx], start=start)\n", " rows.append((f[\"cure\"], f[\"k\"], f[\"lam\"]))\n", " return pd.DataFrame(rows, columns=[\"cure\", \"k\", \"lam\"])\n", "\n", "\n", "boot = bootstrap_fit(cohort.elapsed.values, cohort.observed.values)\n", "print(f\"{len(boot)} bootstrap re-fits done\")\n", "\n", "\n", "def band(fn, lo=2.5, hi=97.5):\n", " \"\"\"Percentile interval for any readout computed from the bootstrap parameters.\"\"\"\n", " vals = np.array([fn(r.k, r.lam, r.cure) for r in boot.itertuples()])\n", " return float(np.percentile(vals, lo)), float(np.percentile(vals, hi))\n", "\n", "\n", "def predict_total(reported, ages, k, lam, g):\n", " \"\"\"One draw of a day's (or a week's) eventual conversion count.\n", "\n", " TWO sources of uncertainty, and the first version of this notebook carried only\n", " the first. The curve is estimated, so F is uncertain — that is the bootstrap. But\n", " even with F known exactly, WHICH of a day's conversions have arrived so far is a\n", " coin-toss per conversion, and dividing by F magnifies that toss by 1/F. An\n", " interval built on the curve alone came out far too narrow: the coverage check two\n", " sections down measured 35% against a nominal 95%, which is how this was caught.\n", " \"\"\"\n", " f_bar = float(np.mean(reported_fraction(np.asarray(ages, float), k, lam)))\n", " if f_bar <= 0:\n", " return np.nan\n", " n_hat = reported / f_bar\n", " resampled = g.binomial(max(int(round(n_hat)), 0), min(f_bar, 1.0))\n", " return resampled / f_bar\n", "\n", "\n", "horizon_rows = []\n", "for N in REFERENCE_HORIZONS:\n", " f_hat = float(reported_fraction(N, fit[\"k\"], fit[\"lam\"]))\n", " f_lo, f_hi = band(lambda k, lam, c, N=N: float(reported_fraction(N, k, lam)))\n", " horizon_rows.append({\n", " \"days_since_click\": N,\n", " \"reported_fraction\": round(f_hat, 4),\n", " \"reported_fraction_lo\": round(f_lo, 4),\n", " \"reported_fraction_hi\": round(f_hi, 4),\n", " \"lag_factor\": round(1 / f_hat, 3),\n", " \"lag_factor_lo\": round(1 / f_hi, 3),\n", " \"lag_factor_hi\": round(1 / f_lo, 3),\n", " \"trustworthy\": bool(f_hat >= MIN_TRUST_FRACTION),\n", " \"truth_fraction\": round(float(reported_fraction(N, TRUE_K, TRUE_LAMBDA)), 4),\n", " })\n", "horizons = pd.DataFrame(horizon_rows)\n", "print(horizons.to_string(index=False))" ] }, { "cell_type": "code", "execution_count": 9, "id": "b0ee987f", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:48:09.100381Z", "iopub.status.busy": "2026-10-24T17:48:09.099845Z", "iopub.status.idle": "2026-10-24T17:48:11.222316Z", "shell.execute_reply": "2026-10-24T17:48:11.221314Z" } }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA9sAAAHnCAYAAACork4jAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzdd3RU1d7G8e8509N7AiGE0HvvTToWBEFAsWHBLnrVa/e1d6+994qKogj2LiJdAZEivZNCejJ95pz3j0kmGZJABhMp/j5rZZHZp8yeYWfIw26K3+fWEUIIIYQQQgghRINRj3QFhBBCCCGEEEKI442EbSGEEEIIIYQQooFJ2BZCCCGEEEIIIRqYhG0hhBBCCCGEEKKBSdgWQgghhBBCCCEamIRtIYQQQgghhBCigUnYFkIIIYQQQgghGpiEbSGEEEIIIYQQooFJ2BZCCCGEEEIIIRqYhG0hxL/Cnj17admmE08+/dw/+ryFhUVcf8Mt9B80jJZtOjHt7PP/0eevbumy5bRs04k5H889YnU4Xsl7K8JxJD6Ppp19PkOGjf7Hnu+fdMONt9KyTad6nz9k2OgG+SxuqPs0hDkfz6Vlm04sXbb8X10HIY42EraFOApV/uL+7HMv1nnOkGGjGT7yxH+wVuJwPPDQI3zxxVecdeZUHv/fQ1xx+SVHukr/iMow8W8On6WlpTz59HPyi6cQx7Ann36Ob7/74UhXQwhxjDIe6QoIIcTx7NdFSxgyZBBXz7ziSFdF/MNKS8t4+pnngSvo36/vka6OEOIwPP3M85w+cQJjRo880lU56k08bTynjjsZk8l0pKsixFFDeraFEKIR7d+fT1xs7JGuxnGrrKz8SFdBCCEEYDAYsFgsqKrECyEqyU+DEMeZ1avXcNHFV9Cj9wDad+rBqLHjeObZF/B4PMFzfl+5ipZtOvHW27NCrr1gxmW0bNOJZ559IaT8mmtvoGOXXrjdHg6mct5ccXExN996B336DaFzt96cd/7FbN+xE4Dvf/iRCZOm0rFLLwYMHs7zL75S4z51zYOra56jruvM+XguU844m67d+9KxSy9GjR3H3fc8EPK6K/28YCGTJk+jQ+ee9O47mNtuvwuHw3HQ11bd1q3buPo//6VP/6G079idE4aP5YGHHg0JfpXvha7rfDx3Hi3bdKrXsOp3Z33A9AsuZuDgEbTr2I0+/YZw+ZXXsHHT5nrXD+DNt95l5JhTaN+xO0OHj+GZZ1/A5/PXOK+83M7jTzzNpMnT6N13MO06dmPIsNHcefd9FBcXB89zuz306juI004/o9bn+/qb72jZphNvvzOr1uMQ+Ht66+1ZjBt/Ol179KNzt94MG3Ei11x7A3l5+w/5mirbxYa/NnLhjMvp3msA3Xr2Cx7fn5/PXffcz5BhowPvXf+hXHv9TezZszfkPm63h2eefYExJ55Kp6696dq9LyPHnMKNN9+Oy+UKnteyTSduuPHWGvWoz/zsOR/PZejwMUCgZ6zy7/9gc2YP5z12u9089czzjBo7jvadetCj9wBmXHIFa/5cG3LtweYIhzvPsqCgkPseeJjho06ifcfu9Oo7iKnTzuWzz78MnpObm8cDDz3KqRMmBz6LOnZnxOiT+d/jT4W8xxD6fs79dD4njZtI+049GDB4OI8+9iR+f812W5stW7Zy9X/+y6AhI2nfsTt9+g1h8tSzmf3hnJDzdF3ng9kfcdrpZ9Cpa286de3N5Kln1zlUeMnSZVxw0aV07xV4HaPHnsqLL71a73odypdffcOpEybToXNP+g08gXvuexC73R5yvGWbTrwz6/1ar7/40itp36kHhYVF9Xq+/fv3c91/b6Znn4F06NyTM6adV6O9AGiaxptvvcvJp06kQ+eedO3el7PPvYCFvy4OOe/W2++kTfuulJaWBsvyCwpo1bYzLdt0Yl92drDc4XDQrmM3brz59nrV9dN5nzPx9DODf0+TJk8LaWeH8ufadZwz/SI6de1N914DuPzKa2p8HtSlsl0CIZ/htc0N37Z9B5dcdlXwc+2CGZexY+fOGueF2/Zq4/V6eeKpZxkybDTtO/Vg9NhTeXfWB7WeW9+fw7y8/bTt0I2rrr6u1vu8+vqbtGzTiR9+/Pmgdavts6S+n7dCHK9kGLkQRzGny1XnL1CapoHBEFK24JeFXHLZVURFRXHOWWeSnJzETwsW8sRTz7Jy1Wpee+UFVFWlW9cuREVGsmjxEqafdzYQ+Ad8xYrfUFWVRYuXMvOqy4HALwdLli6jd6+eWCzmetX7/IsuJSkxkZkzLycvbz+vvf4W0y+4mOv+M5P7H3yUs6ZNZfLpE/n886/432NP0iy9KeNPPeWw36cbbrqVT+bOp1PHDlw84wISExPZtWsX33z7Pdf+5yrM5qp6L1iwkHfefY9pZ07l9EmnsXjJUt6f/REoCvffe+chn2v9+g2cedZ0fH4/55x9JhkZzfjtt5W8+tqbLF68lI9mv4vNZmPatKkMGjSA6/57M31692LamVMA6Nmj+0Hv//Krr9O9W1fOOWcaCfFx7Nixk9kffcyUqWfz2bw5ZGY2P2QdH370cV56+TU6d+7Ef6//D26Ph48++oTvfvipxrm5ubm8P3sOY8eM4pRTTsRitvDHmj95/4OP+O23lXz6yWxMJhMWi5lJEyfw2utvseGvjXRo3y7kPrM/nIPVauW0CafWWa/nX3iZx554mmEnDGXq1NMxm0zs25fNgoW/kpuXR0pK8iFfW3Z2DtPOPp8xo0dyw/X/YX9+PgD7srOZMvUc7A4HU6dMIqtFJjm5ebz33mx+/XUxn86dTXrTpgDcdc99zP7wYyaMH8f0884BYPeePfz00wIcDidWq/WQ9TiUvn16c/utN3HfAw8zZvQoThw7CoCIiIg6rwn3Pfb7/Vw44zKWLF3OsBOGct4509i/P59Z789m6pnn8sZrLzKgf7/anuqw7d23j6lnnEtObi7jTz2FC6afg8frZf36v/jxpwWcOu5kAP7auJGvvv6W0aNGMGXyRHQ9EF5eePEV1q3fwBuv1lyP4v0PPiInN5czppxOQkIC33z7PS+8+ApRUZFcfunFB61XUVExZ517AbqmM+3MKTRr1ozS0lI2btzEsuW/ccbUycFzb7zpNj75dD6jR41gwqnjAPjm2++57IqruffuOzj7rKr/7Pjwo0+45bY76NSxA5ddchExMTH8vnIVjz72JOs3/MXTT/7vb72fP/60gNdff4uzzz6TKVMmsXjxMt586102bNjIu2+/hsFgYPSoESQnJ/Hhhx9z7tnTQq7Pycnl5wULOfmksSQkxB/y+ZwOJ1OnnUeXzp249j8zyc8v4I033ubCiy7j5x+/ISoqMnjuDTfdxtxP59OrVw/+e/01OOwOZn/0MedfeAmPPfpgsB0OHNCfD2bPYcnS5YwdE2jnixYtQdf1wL8ni5YyZfJEAJYt/w2v18eggf0PWdcnnnyGZ557kXbt2nL1VZej6zqfzv+ca669gV2793DlIda/WLtuPWeeNR2DqnLeuWfRpEkav/66mGnnnI/T4Tzk87dq1ZLH//dQjc/wA+Xk5nLmWecxauRwbrrhOnbs3Mnb78zi0stm8tUXn4b08obT9upy/Q238PkXXzFgQD8uumA6RcXFPPX0czRp2qTGufX9OUxJSWbkiGF8//2PFBYW1WhLH374MU2apDHshCGHrN+B/onPWyGOan6fW5cv+ZKvo+tr0aKFemZW60N+DT1hePAaj9uhDxpygt6uQ2d969ZNIfe7/r836JlZrfWPP54TLLvgwov0Tl266x63Q/f73PqSxYv0zKzW+rXXXae3adtBLyst0v0+t7527Ro9M6u1/txzzx2y3tddd72emdVav/mWW0PKX3nlFT0zq7XesXNXfeeObcFyh71M79Wnnz5x0uSQ8wcOHqpPPWNajfvv3LFNz8xqrT/2+OPBss8+m69nZrXWL7n0Mt3tsoec7/O6dJ/XFXJtuw6d9e3btoScd+550/XWbdsHX/PBviZPOUNv0bKNvmzZkpDyJ554Qs/Maq0/9dRTIeWZWa316667vt5/97XVYcOGdXrrtu31226//ZDXb926Sc9q1VYfP2Gi7rCXBcsL8nP1Pv0G6JlZrfXZsz8IljsdZbrLWV7jPu+9N0vPzGqtf/bZ/GDZ5s0b9RYt2+j/d8cdIefu2rVdz2rV9pCv8+RTTtVHjhpz2D8XAwcP1TOzWutvv/12jWMXX3Kp3rV7zxp/tzt3bNM7du6qX3f9f4NlXbv31M+bfv4hn6+uv7vKn8/q72NtZbW110N9hfMev//++3pmVmv9pptvrnGPNu066icMH6l7Pc5D1mX27A/0zKzW+qJFCw9Zv/MvuEjPzGqtf/HF5zWOVT6X3+fW7eUlIY8rvx555BE9M6u1vnLlbzXeu959++uFBXkh9xs5aozep9+AQ9brm2++0jOzWuvz5n160PO+/jpw3suvvFzj2IUXzdA7demulxQX6n6fW8/J3qO3bd9Jv/iSS4OfI5VfL730kp6Z1VpfvOjXw2rLlX8fLVq20Vf+viLk2J133lWjLT1c8b6tXr0y5Nynnnqq3n93U8+YpmdmtdafefbZkPJPP52rZ2a11t99951g2cKFC/TMrNb6eedfEPw3wu9z63m52XrP3v30Lt166qUlhcGyFi3bhHw+Xf/fG/TRY07UTxk3Xp959TXB8rvvuUfPzGqt52TvOWhdt2wJfI6deNIpenlZcbC8rLRIHz32JL1l63Yh/5ZU/ttT/R6Tp5yhZ7VqW+M9u/mWW/XMrNa1/htT29fBPsMrP5Pmzv0kpPzZ557VM7Na6z//9ONhtb26vir/Xi6+5LKQNrlj+1a9XYfONdpCOD+HP//8k56Z1Vp/6aWXQs5dunRxvT/Havssqe/nrXzJ1/H6JcPIhTiKnT7pNN5+89Vav5KSEkPOXbtuPXv27GXihFNpkZkZcuyaqwOLc339zXfBsoED+lNeXs6aNYHhg4uWLCUuLpZLLr4Ij9fLit9+B2DxkqWB8+vRE1FpxkXnhzzu17cPACNHDKdZs/RgucVipnu3rmzfsaPe9z7QvPmfA3DbLTdiNIYO1lEUBUVRQsrGjB5J8+YZIWWDBw3E6/Wx+xDDCwsKClnx2+8MHjSA3r16hhy7eMYFRETYQt7jw1HZ86nrOmVl5RQWFpGUmEhWVhar//jzkNd/++0PaJrGxTMuCBmJEBcXV6NXDMBsNgcXs/H5fJSWllJYWBTsEV39x5rguS2zWtC/f18+nfd5yPC/Dz/6BE3TOLOOnp9KMTHR5OTmsWzZikO+jrrExcVy5hmTQ8pKy8r4/oefGHbCUKKioigsLAp+RURE0L1bNxb+uqiqHtExbNq8hQ0b/jrsejSWcN7jyrb2n6uvqnGP8aeewo4dO9m4cVOD1a24uJifF/xCv759OHFszeHw1XvwrFZr8LHX66W4uJjCwiIGDxoIwB/V2lWlKZMnEVttfQNVVRk4sD95eftDhlXXJiY6BoCfF/wSMpz5QHM/nY/VamXcySeFtJPCwiJGjxpJeXk5q1atBuCrr7/F7XYzdcrpFBUVh5w7YvgJAPxSrV0djsGDBtCtW9eQsiuuCPTYVv8smXbGFFRV5YPZVUPidV3nwzmfkJXVot4L8KmqykUXnHdAHQJ/J5VTfao/98wrL8NQbQRVYmIC5559JqWlpSxesixY1r5d2+BjCPy7MXBgfwYO7M+SpaHlbdu0Jjn54KNYvvsu8Dl26SUXYbPZguURERFcMuMC/H4/333/Y53XV35WDx0ymC6dQ4d9z7zysoM+d7hSU1NqjMyqek93BMvCaXt1qfx7ufyyGSH/tmVkNGPC+Jqjw8L5ORw8aACZmc2Z/eHHIfeY/eEcDAYDU6dMOsQ7Ubuj+fNWiH+CDCMX4iiW2TyDwYMG1HrMYrGEPN69ew8A7dq1qXFuetOmREVFsXPX7mDZoIr7Llq8hB49urF48RL69+tL2zatSUtNZdHipZwwdAiLFi0lNjaGzp061rvezTOahTyOjY2ptRwgNiaGoqLiet/7QNu37yA2NiYkxIdTNwgEOIDiQ9Sj6j1uW+OYzWajefMMdlV7jw/H8uW/8fRzL7By5eoa89kymtWs+4Eq/45bt25V41ibNq1rveaD2R/x7qwP2LR5Cz6fL+RYcXFJyOOzp53BVVdfx1dff8vE08bj9/uZM2cubdu2oVfPHget2w3XX8tlV17NtHPOJykpkT69ezFwQH9OPfVkYqKjD/naAJo3zwj55R8CbUDTNObN/zz4ny8Hqh4E77j9Zq674RZOGX86TZs2oU/vXgwdMoiTTxpb4+fqSKjve7x79x7i4+NqHX5f+Tmwc9duOnRo3yD12rlzF7qu06lTh0Oe6/f7eeXVN/j4k3ls3xH4+6nuwHYFgcBwoLi4uOD5kZGRNY5X6tu3N1MmT+KjOZ8w/7Mv6dSxA7179+SkE8eETN3YunUbLpeLAYOH13mv/IICALZs3QbAjEvq3kkgP7+gzmP1UdvPaVJiIvHxcezcWfVZkp7elKFDBvPZZ19w2y03EBERwa+/LmbPnr3cctN/6/18qSkpNdp4fHwcQMjncPCzrm3Nf08qP/927toVLBs0aACvvvYm2dk5uNxu9u3LZtDAAZjNZl5+5XU2btpMYmICGzduZvq5Zx+ynrsO8lnbrm2g7GCftZXH2tTy/qalpRJdz8+b+qit3cZXtNuiau08nLZXl4O9rjata36+h/NzqCgK086YwkOPPMbyFb/Tt08vSktL+fKrbxk6ZDBNm9Qcpl4fR/vnrRCNTcK2EP8SB/bwtmvbhuTkJBYtXsr508/ljzVrufP/AotBDRzYj0WLlgTmcf/2G4MGDgxrddEDw1AltY7yQ9W1ks/vq7U8HAerg47+t+//d1Qu5pOR0Yzrr7ua5hkZWK1WFEXh3vsexOE89DzDcL3x5jvce/9DDBo4gHvuup2UlBTMZjN+v48LLroM/YBfzirnj34wew4TTxvPgl9+JTsnh0suvvCQz9W9e1d++v4rFi1ewpKly1m+/De++vpbnnzqWT547y1atWp5yHvYrLYaZboW+Hs75eQTQ+bm1mXkyOEs/Plbflm4iKXLVrBs2Qrmzf+cp595njkfvkdiYsJBr69tobmG9Hfe47rU9TMFNNhCX9U98NCjvPHmO5x04hguveRCEhMTMZlM5ObmcsNNt6HpNX/WDAf5jNFrOf9ADz94L5fMuICff1nIb7+t5KM5cwPrRZx7NnfeEfhs0zSd6OhonnvmiTrv07biP6W0inb14AP3BOf7Hyi1HusMNJSzp03l5wW/8PkXXzN1yiQ++HAOZpOJ0yedVu97qIaDfI7X4z2uy6CB/Xn1tTdZtHgJLpcbo9FIv759MBoNmM1mFi1aQlJSIrquM3Bgw64jcKQZ1IP8m1LtPQ2n7TWUcH8OJ58+kceffIbZH86hb59efDo/MMLmrEOMWjqYv/t5K8SxTsK2EMeJyqHRmzZvrXFsX3Y2ZWVlZDYPHWo4oH8/vv76W35e8As+ny84VHzggP7M/fQzfvjxZ+x2B4MG1X8IeUOIi42luKRmz9fuXXtqlGVltWDL1m3s3buP9PTafyFuKJU9GJs3b6lxzOVysXv3HjKbH3oBs7rMm/85Pp+PN159scZQ96Li4nr1AmRWXLdly9Yav7jVVu9P5s6jWbN03nrj5ZD/UNmypWY7AjCZTEyZPInnX3iZrVu3BRftmnha3QujVWez2Rg1cgSjRo4AAov6XXDRZbz40qs8+sgD9brHgTIzm6OqKi6Xq86RIAeKiYlh3CknMe6UkwCY9d5s/u/Oe3hn1vv85+orgcCIh1rb4e76jV44WMA9mPq+x82bZ7B12zb25+eTnJQUcmzTpsDfdWV7qByeXVJLj3JlL+KhZGY2R1EU1q/fcMhzP5k7jz69e9UIFj8vWFiv5zpcrVq1pFWrllx0wXRcLhcXzriMt96ZxUUXTqdZs3SysjLZum0bnTp2CPbo1iUrKzAdJzYmpt7tKly1/ZzlFxRQVFRMj+7dQsqHDRtKkyZpzP5wDiNGnMD3P/zI2DGj67UwWrgqP382b95K9+6hw9w3VeyMUP2zrk/vXphNJhYtXorL5aJb187BxdZ69ezOosVLSUpKxGAw0K/voYe8Zwaff0uN3vVNmzeH1PGg9a/l/c3JyaWsrOyQdWho4bS9ulR/Xd26dgk5tnlL7Z/v4fwcJiTEc+LY0Xz19bfc+X+3MHv2HJqkpTFs2NDDqm+l+nzeCnG8kjnbQhwnOnXsQLNm6Xw6bz579+4LOVa5ldeB8ywHDRyAx+vlmWdfoGnTJmS1yAyW67rOE08+E3g84J8N21lZLdi2bRs5ObnBMk3TePX1N2ucO2F8YEXXBx56tNYeuvr0iNVXYmICfXr34peFi2rMOX31tTex2x21zmWtr8oekgPrPOu92fUerjp69EgUReGVV98I2aqtuLi41q2DKnu6qg8v1HWdpw/Y/q26M6dORlVVnnnuRX76+RdOPmkMMTExh6xbQUFhjbLOnQLzKYuqbTMWrvj4OIadMISffv4lZH5odZWrlvv9fkpqCdCdOwemSVTf7qxlVharVv2Bs9qIArfbzVvvvFevelXOv68t4B5Kfd7jypWfn30udGXvHTt3Mv+zL2jRIpP2FSuaR0VFkpKSzJKly0LaV1FRMR/N+aRedYqLi2PYCUNZumxFrfNlq7chg2qoMVLE6/XyQi1b/TWE4uLiGkNkrVZrcJh2ZfuaeNp4AB565LFaPxsq2wnAKSediMVi4alnnq91zrjL5aK8/OBzyQ/l10VLanyWPP/8ywCMHRP6WWIwGDhjyumsWv0HDz38P7xeX50rZP9dlW3r+RdfDnlfCwuLeGfW+8TExDBwQFUPdUREBN27d2PxkqUsW76CgQOr/nNi4MABrFjxG78uWkzXLp2Jjo465POPHj0CVVV55bU3cLvdwXKn08nLr74RXKW9LomJCfTu3ZNfFv7K2nXrQ449+/xLh34DqomMjKj1P93CFU7bq0tlm3jhxVdD7rF79x7mzf+ixvmH83N41plTcblc3HPfg2z4ayOTJ0+sc7TaoYTzeSvE8Up6toU4ThgMBu69+/+45LKrOG3SGZw1bSqJiQn8vOBXfl7wC0OHDKqxiEvlcL7NW7YyZXLV4iepqSm0btWSzVu2kpaaSsuWWf/oazn/vHP47PMvOfu8Czlr2lR0XefLL7+ptafw5JPGMmH8OObND+zHOnbMKBITE9i9Zy9ffvUN8z6ZXa8gWF93/t8tnHnWdM4+70LOnlax9dfvK5n/2Rd0aN+Oiy6cftj3Hjt2FK+98RYXzLiMaWdMwWqz8ttvK1n46yIym2fgq8dw36wWmcy48Hxeee0Nppx5DuPHnYzH6+XDDz8mJSWlxn7WJ584locffZzpF17CSWPH4HK5+Oa77/F6vXU+R7Nm6QwZMoj5nwV+uTvzjPr9wj/mpFPp1rUr3bp1IS0tlZLiEj6ZOw+ASRMn1OsedbnvnjuZOu1czjv/Ysafegpdu3RGURX27d3HzwsW0qVzJx595AHsdjv9Bg5j5IhhdOzQnuTkZPLy8vjgwzkYjcbgdjwA0887m2uuvYFp55zPpNPGY3c4mDt3fr3ne8bHx5GZ2ZzPv/iK5s0zSEpKJMJmY+TIuudsVqrPezxp4njmzf+Md959n337shkyZBD5FVt/6brO/ffeGfIzM/28s3n0f08y/YJLGDN6JIWFhcz+8GMymjWr93/m3HPX7Uw+YwOXX3kN4089he7duuD3a6xfvwGf388Tjz0MwEknjWHWe7O54qr/MGTwQIpLSpg3/wus1saZo/nJ3Pm89sZbjBkVWADRZrPy59r1fPjRx3Ro346OFfPWTzpxDGeeMZkPZs9hw18bGVMxZD83N4+169bz84KFbNrwBxCY23v/vXdy0y3/x6ix45g0cQLNM5pRXFzC1m3b+ebb73nphaeDi5MtXbacs865gNMnTqj3KI1OHTtw7vQZnH32mTRt2oRFi5by7Xff07t3TyZNHF/j/DOmTubZ51/ik7nzadEis8G3dqs0cEB/Jp42nrmfzuescy5gzOiROByBrb8KCgp57NEHa8yhHzigH8tX/AYE/sO20uCB/Xns8acot9uZdFrN11SbFpmZXHn5JTzz3ItMmjyNCePHBbf+2rhxE9dfd80h1+m4/dabOPOs6Zx19vmcc8604NZf69ZvICG+/qMBunfrxqLFS3nxpVdp2rQJiqIEt7gLRzhtry6DBvbnpBPH8NXX33LO9IsYPXIExSUlzHpvNq1bt2Lt2nWhz3kYP4d9+/amTetWfDJ3PqqqcsbU08N+rZXC+bwV4nglYVuI48gJQ4fwway3eea5F3nrnfdwOZ2kN0vn2muu4tJLLqox7zq9aVNatMhkx46dNfY9HTRwAFu2bjsi8+t69OjGE489zHPPv8Qjjz5OYkIiEyeO5/SJExg1tuY/zo//7yH69unF7I8+5vkXX0FRoElaE0YMP6HB9/Ds2LEDn8x5n6eeeZ6PP/mUsvIyUlJSuOjC6cy86vKQlXPD1bNHd156/mmefu4Fnnz6OcxmE7179WT2e29zx133sWfvwVdLr3TzTdeTmprCO7Pe59HHniA1NZUpp0+kR4/unHf+jJBzL55xAQCzP/qY+x54mPj4OEaNHM5/r7uGHr0H1vkcZ0+byoIFC2nbpnWNldnrMuOi8/nll0W8O+sDSktLiIuLo2OH9tx2600MGVz3c9VHWloqn336ES+98jrfffcjX3z5NWaTidTUVPr06cnUyYFfGK1WKxddOJ0lS5axbNkKysvLSUxMpHv3rlwy44KQlaFPHXcy+/fn89bbs7j/wUdo2qQJZ007g06dOnDOeRfVq15PPvYw9z3wMP97/CmcTifp6U3rFbbh0O+x0Wjk9Vdf5IWXXuWzz79k4cJFWG02evfqwcyrLq8xzPSSGRditzv4+JNPWb58BS1aZHLtfwIrma9affBf8iulpzdl/twPee6Fl/nhx5/54suviI6KpnXrVpx37lnB82675UaioqL4/Iuv+PHHn0lJTeHUcScz6bTxjD6xflMOwtG/Xx82/LWRBQt/Zf/+/ei6TpO0Jlx2yUXMuOj8kJ65B+67m/79+vLB7I945bU3cbtcJCYl0rZtm+C6FZUmTZxAy6wsXnntDT6a8wklJSXExsbSPCODiy86n/bVFvCyV/Ryp6al1rveI0YMo22b1jz/wsts2bqNmJhopp97Ntdfd02tvYmpqSmMGD6Mb7/7njPrsT7B3/How/fTuVNHPpzzCY/87wlMJiNdu3TmgXvvYsiQQTXOHzRoAE8+/RwRETZ6VBt63qVLZ2JiYigtLQ1rV4tr/zOTrKwWvPX2LJ58+jkA2rdry5OPP1LjP45r07VLZ96f9SaPPPo4b709C6PRyMAB/fhg1ltMO+f8etfjnrv/jzvvuo/nX3iZ8ooRDocTtiG8tleXJx57mFYts/h47jweevh/NMtoxtUzL8dmtXLjzbeHnHu4P4fTzpzKPfc9+LcWRoPwPm+FOF4pfp/7yK4IJIQQ4phTOdf6jttv4fzp5xzp6hyX5D0+ttxz34PM/XQ+P33/VXAV9cZw9X/+y7fffs+ihT/K4lKiUVTOqX75xWeC62sIIQ6PzNkWQggRtjffnoXNZqv3wmgifPIeH1t+WbiIKy+/tFGDdk5OLt98+x1jx46WoC0ahaZpvDPrfZqkpTF82AlHujpCHPNkGLkQQoh6yS8oYPHipfyx5k8WLFjIxRddEFzlWjQMeY+PXd9/U/se7w1h9eo1bN22jVnvzUbTdC67dMahLxIiDLt372HlqtUs+OVXNm3azF133nbYC6MJIapI2BZCCFEvW7Zs5T/XBeYAnj5xQnC+r2g48h6L2sx67wM++XQ+6elNeeSh++hQscq8EA1l2fIV3Hjz7cTHxzHjovM556wzj3SVhDguyJxtIYQQQgghhBCigcmcbSGEEEIIIYQQooFJ2BZCCCGEEEIIIRrYvz5s67qOpmnouoymF0IIIYQQQgjRMCRs6zr7c3bj9/uPdFXEcUTTNHL37UTTtCNdFXEckXYlGou0LdEYpF2JxiJtSzSGxmhP//qwLYQQQgghhBBCNLQjvvVXTu5+Pv/qB7Zs3cHuvdk0bZLKo/ffesjrdF1n/hff8d2Pv1JaVk6L5umcO20SbVpn/QO1FkIIIYQQQggh6nbEe7b37M1m1Zp1pKYmk940rd7Xzf/iO+Z8+hUnjx3GjddeSlxcDA/+73ly8/IbsbZCCCGEEEIIIcShHfGw3bN7Z557/F6uveoisjKb1esaj8fLvC++45QTh3Py2BF07tiOqy+/gMjICD7/+odGrrEQQgghhBBCCHFwRzxsq2r4Vdi0ZRtOp4v+fXoGy4xGI317d2P1H+sbsnpCCCGEEEIIIUTYjnjYPhz7snMBaNokNaS8aZNUCgqL8Hg8R6JaQgghhBBCCCEEcBQskHY47HYnJqMRs9kUUh4VGYGu65TbnSSYzbVe6/V68fp8wce6HljiXdM02T5ANJjKtiRtSjQkaVeisUjbEo1B2pVoLNK26qZpGitWrjnoOW1atSAhPq5GeU5uHjt376vzOpPRSM/unWs99sefG3C53cHzenTrhKIo9a94I9F1HV0HLfilo+kEy/SKx5oOfr9GapylQZ//mAzbf8e8z7/j43lfBR9bLGYevvNqivKzj2CtxPGqIG/vka6COA5JuxKNRdqWaAzSrkRjKcjbi67r7N6bS2lZebA8LTWJpIS4GueXlJaze29Orffasn03MVGRjBjat8ax3XtzmT33m4PW5bILJhMVGVGj/LOvf2Hjlh3oeuBxxR/Bx5mZGYwaOQy/Dn6tIvRp4HC5+fjDD4MXHHgdQJ8hI0lIaVIVJDXw+TXmvfnGQevacdCpJDZrUy2ABq7N2bKS3Wt+rvM6gyWC5sMuDb2u4tr8pbPxO4uC56q2BCydJ4MpEq/fhx8NNCOgBK/z6X7cuhNN1zFp0egQDMIu3Y4XJwpmTP6YQDiueP3lSi4aPkz+BAy6req1Ky4chmx0XcfmaYmuV71vbmMuPkMhBi0aqzd0nbBy6zp0xYvV04K/7u510PcuXMdk2I6MtOH1+fB4vCG92+V2B4qiEBVpq/PaCeNGc/KJw4OPdV2jvDif+KQmGI3H5NshjkKaplGQt5fElPTDWpdAiNpIuxKNRdqWaAzSrv4d3G4PK/9Ye9BzOndoS3R0VPCxX9Px+XXuuOdRdPSqUFTxZ2WgHD72FJpntcGn6fj84NN0vH7YuHY1q5f8gGowUJyfW+P52vY7ifT2Gfg0vSK86vg0KNi1k+2LP62znhHNe7OwIA81phl+TcevBa71Fml49gWeR6Gqt1ZTdDzGQJ0v+0jDb9DxVVzn0otxYyc5P5e04ryQ58mL0fCrOnF2lb+Ko3l1rx8An1qOw7IZ0Im3d6RDXtV1+6M1iiI1olwKTYsNALy40I3dqlEUuRBNdRHl7IrFm0yWOR0Ah9nDjtRidHQ67a6afvv5Jgu5eSuxWzZi9TYjzj4IgBhHJInmdP5okY3H6KdVTgIJ5VX/gVBs0/k893lAJyP/ipDX5E3xk53gIbk0ko67U3AaUsnZZwNFZ3fSi/gN5SQXTyDS3T54jcuUTU7CLABa5N8Ucr/CqKWURv5GhKsdKeWnhRzbnTSnjvvlkxP7YeB+uaH3s1vXBu9nLQkN20VRPwfvV+LwExthoKEck+mycq52dk4umc2r3qx92bkkJsRjrmMIOYDJZMJkqgromqZRTmChNvmHQDQ0aVeiMUi7Eo1F2pZoDNKuDs/a9RvJzsmr83j7dq1p2iQNj0/H66/8gpdfeoWcnJyQntTg9zq069aX9r2GVrsOfH6dgty9/PnT7MCwW6rOr/59Yt/z0M1RgeDrDwRLr6MU3+9vVTyXjqaApoBJqwqlOelnUmyJxo0d3W/G6I8DoPPewJDl/dEaPkMgfNq8Vdfd/+U2cuLXAzqxjn7B8oTyPBRyybdpxCSpZOYHwpHDHNhG+M2iVTh/W0uso19IL6bBX8j+Djq6Ar23pIUE5+0pxeyNW4a1rJCE7PEh7/XupA/x9/KQsX8gsY6MYLnDnM+2JoGdkFrkGEDxB48VRq2gNPI3Ck0tcBlPC7nfX+nz8RmdZOwfQKSrKiz6DCUUxnwLQIyjKzuSqq7Ljl9FQcwmYuzN8BgD4dhlSgag3PYnfkMZVk8zLL5UtqdMqTi+l9y4dwGwuSeHvF6vYSMuyw5UvSo3lUa0oTSiDSURz+M3ONkX35eSiA7B4y7TXvyGwP109JD7ldlaoKn7KbM2ZXvKREJVnqfXUV7zfnVfU/cxRVdRNHOt91N1KwZ/NKpWs1PW5EtE1awouhmnVye2lmc8XMdk2G7buiU2m5WlK1YHw7bP52fF73/QvVvHI1w7IYQQQghxPPL5dTw+HXfFl7fizzK7C69fR1dNuH2Bc4LnefyU2R0smfcSrvISFDUQDCtjQlLPKRjis/D4A/fz+nU8jhL86z5AcReFPL9GRU+qAkVRIymM7Bo85jUU4TMU0zx/L6mlJSHX7U704zHopJaobFuVz5vb8iquKaYkYgm6opGxvzuty6ueb0eyn70JfhLKVDrsC0SGhRvteI0q+xLexqeWkFg2lhhHMzIsgd/HS2wuVrUKTBs4YW3LYNgp8xnJi15GaeQKIlxtSSkJhLHtSZMA2JQ+D6/RQbP9A4l1ZAbrUGrzUhQdmP4Z4+iDUrG2c5m1JYVRrSmI/Ytoe1M0huKwNEFXAvUsiXgBv6GUSFdVUARwWJIojQgspLwt5Yzg/QAKo37CY1qO0e+npsB5TnMiBq1FsNRtrJrfqysGlFqu8asmyq0tQo/osRj8BlymNAxaSrVyMyZvMgoKuqIecF0xVrcbVW9W434R7tZoqguDFl11LwUsehSxzp6oikKEGQyKgqooqCoY1dZEeqxEKskkxRoxKKAooKoKVmUYmu4lKa4Z0YoZVQVVUfCRRro+CYOq0izTilFVUFUFgwrlWj8cegciYqJISo9EVcBQcV2xfzooOrFN44g02FBVBVUBTbfh0P+LUVGIax2L0aCiKoG6e/VT0ZWTMalGrAZzoLziOrgNg6JgMKgoCsG6G9RkVKVb4DVWnK9U3E9VJqEop1e9zopzAt9fi6oogEZabMPG4yMett1uD6vXrAMgv6AIp9PFshWrAOjQrjUxMdHc9/Az5BcU8uQjdwJgNpuYcMpoPv70K2Kio8ho1oTvfvyV8nIH404cecReixBCCCGEaHxev47bq+Pyari8gVBb+b3Lq+PyBL53eDRcPh2Xp+qc8rJynI5SvH5wFOfhzNuChlo13DiuA+7IFrgrgrXXp+Px6+jOXUQ6V6PqChZfVW9gvGM9uTF+iiIzcJrHYKrotQXwqaXYrb8S41xH5zJjIET5A4t67UwKhFm1aClR2Skhr29fwvv4O+TTbYdCpLtdsLzM6mB1y+0AdNoZHXJNmW01pZHLsVszKd9/YsixLU0/wWssp8Q2gEh3l2C5pjgpjwgspuU2jWBr8hnBY3lxv1MUtQ6v2gSzd0Tg9RhsFdc50AwOdMWL3xDBjuTJgXsYc4BAL/eO5IkoIVEjED51pWpRM7u1OQCKHoPqV3CbUoJlgXMLMHuaoGCgei+m1xiNQhuiXCpWmqDHNCeqIvQZVAWf1hNN95AcmUJslDEQMlXwKYnEKWNQFQMtm1kwGgwYFFBVKNN74qAFkVGxNGkaWRFIA9cV+M5CUXQS0xOJNEViUMGoKOhE4dBvxGwwEN8pBqOqYKioh6JMwqBOwmwwYDQEgp1RVTAYwKDcGLiHQcFYUWejqmBU01HV7hgqyxUwGBQMChgNzVGU8VUBNxgwFVTlUgwVwdJQETKr9KR2WXWUAzQ/yLG2dZQ3Pcg1qQc5dvRojAX3jnjYLi0t48nnXg8pq3z8fzddTceYaDRNw+8PffHjTxmNDnz+9Q+UlZaT2bwZN//3ClJTkv6pqgshhBBCiGo0Tcfp0XB4NMpcfnLzdXY7nTh8VJQHjleGYqdXC4Rgj0Zp/l6c5aWBock+Ha+m4/WB2xiDy5SM06sFeou9gUWUVM1NlGsjFt8KvAYTcXZryLDR3Fgb+xISMPpjiXJ1CpYnlv1OXuyPOC06rXKMJJdV9W6W2TQWN1uDX1FoXjAzJCzao35nY/pGUotV+mwN3RFnXYYPh3U7SSV7QsK2X3VSFP0nRdHQcQ8ouoEtqeegKyr5Mcsoil6FzeUiynXA+6h48Bv87Ew6iWhXVe+1x5gHBMJ2mTU9ZA9fRTehaCY0xYbT0iTkfiZfU1StHJ8hDa8xJhD0VAWTMZJ4T1eMioHkeBtmNRqTQcFoUIhWIEVpQnRMAhmJWRVBMnBdtv9UdMVPWmJT4kxRweCoY6VEuwqzqpLWNgGzQQ3cT1XwK+NR1XFYDUZsJmNF0FQwGsBkuAmTgYrAWRFCDQomQzomtTsmI5hUBYNRwaSCUdGxFxmITz45MEVBqRpYrCigkBn4s9bVsDvUUgaQXkc5HDwsHhtBUhwZRzxsJycn8v6bzxz0nDtuuaZGmaIonDZuDKeNG9NYVRNCCCGEOC5oFQtFVc7f9fg0ylwadreG0xPoAbZ7At87PRpOr47dXRmKK8+p6D32BEJysBe52vduX6D30aeWoKluDP4oDLqfCPc+LN588qNbU277E13xEWvvj4KBaOc2Mgvmsych0NMb41TpsDfwK6oJKI3sxspWO/BY9pNYehLReqBn1uQvI875Pb909AJw4iozxmrzhO3WSIqj1mFzZ4WEbV0xUhgNdqtOapGZ5DIfhZGd8RhicZrL8RmWV5znR9GrflX2K5EAOE2x5MT0CHl/Vf0PjD4fil61sJKqgE2NIMbTFoNiZH/7k7AarTQzKpgMCklqF0qVBGJjEslKtgVDrtEAOd7TQPWT1jSNeHNMIHiqCihROLkRi9FIYvdoLCYjpooAajJOwqSejsWoBMoMCiYDFY9nBh4bA/cxGZRgQFWUrigoFeG0sgwU0lGUiv++qBiKGzymhIb5UAcLrQ1D0zQcSuD1hPbiCnF0OeJhWwghhBBChO73qh+wH6xfCwRau0un3B0IyeXuQA+y3R0IxvaKx46K8OxwV/UkOyuGVFeGad8BoyXdxhw01YHJl4hRq1oeyK/aKYlYhqZ4aVLYA6NWsWqyIYqC6M2URizHrDQhxTkBgITyP7B5cljZ8i+cFhetsjNom12CQXfjNUSxPyaZouifAYh29MKgG1B0HwAOs87+WB2vwYTdXBXmPMbYwABiRUdXfMFyTTHhNqQCe1A0hV0JozBqVXNofYYcIrx7iSSV5GgDZqMSCJ5pfXArOprqIr19O+LN6aQYwGxU0BUPrX1tsBpMtGqZgM1oqgip4GcCKOOINJuJMVsxGRXMBgWzScFqHB/83mJUsJoUjKqKooJC58A804qwqiqVwbZpMLyqFaG2KvCmBOaQBq+hWk+t9KQKcayoV9i+8ebbw7rpIw/dd1iVEUIIIYQ4llT2GGsHhGNN19Eq9sz1aRp2l06pK9CbXO7yB/50a5S7dMrdfuxunXKXFgzJTq9eNey6omdZq7bwrsdQgKY6MPpjMWoxwXK/4qAkcgma6iahbCSRrgKsvkIA8mK2kxO/iQh3HK1zAis7+9QIymwtKYj5Co8pj4TS0cQ4A3M84+1rcRsd7E5eAUDm5vVYfYFexLzovuTHGPAZSzBoVVs6Rbj3EufchNnnwWkBi28fhoreYZO/HJVIbN50DIqJpGiFKKMRc0IrtCbTSTbmYjQWEZ+cRKvM7sFwbDYqdPO1QlU1mtiSSbLFYDEqWIxJmAzXcK0RIswGzAYVi0mpCtWGyvmsVb22akVgVZQJVQGWqsWSAj246Sgo1QJuzTCsBoOvEELUrV5he/36DSGPc3PzKCwqIi4ulsSEBAoKCykuLiEhPp60NPnfNiGEEEIc3YK9yFogGPtrCcl+PbD/r9evU+rwU+LUKHVplFd8VfYu2yt6kO3uQFC2V+ttruxp1vGjqS4U3YiqV/S+6hpGfw5262+oehFNC5tit3bEYQkMwy2z/klJ5GJMxkT67YFI995g/Re1d1McqdNht5Hm+THsShyH2xSPz+ijNPI3AOLKBxPn2Eii/Q8A3CYfe5P8qHoB6UXfA2A3p1Nma4miWyq2zAkkeqtJoWnR93gNOoXRKkYN1Iqwr1rjaN00htYtsijWk0m0xNC+axxWo0JJdl88ZS1pp3qxqCox7d0oXaOwWSz07NWLqEgrilI1pLtyeHJlIFarBdnK7wPnJFRbOTg0/Fa/XoKwEOJoUq+w/fn8j4PfL/hlIf935728+9T/GNC/ar+7xUuWcvOtd3D9dTXnVwshhBBCNDRNqwrJ/ooeZr8WCMh+vVqvc0Wgdnp1ypwaJU4/pc5AYC5z+wN/VvQ0l1WGaXe1UO2p6lLW8eMx5uJXXVg9zVCpWpXaad5BuW0Nqj+CxPJRIXXNiX8ft3kv8WXDyNwfRUrpMsz+UtxGne+6BbYiaptTiM+QFgzbuuLDZywOLHylR6NUW43Z6Ad00BUdk2an5f45FGRMQYlNwUMHrKqFHpkR2CKa4C90YFChtdFLapGLSN1EamYEqqKQmJzKtSNSsRiuw2pWsVb0DKuKwjtvdgJdpwtgNpnod/EgMrOyKrbnqQzDHQJBV60Mub2qeonRcBfvIzI+HTW4pU/NUC2EEMersOdsP/TI41x7zVUhQRtg4ID+XDPzCh56+H8MO2FIg1VQCCGEEMe3ytBc9WegV7n6n8HvdfD7NcrdOqVOjdKK4FxQYidv//7g3OXKucl2j47TreHw6pSqKYGuzwo+tQxVW0Nq2VJSSwzo1Vay3pbupSBao4kjBoyXVdVV8ZKd+A4AgzdEE+vwBo/tTvSRG+8j0qXQac869sUNpygqsJK0qlsD16su0kr+wKg5gEBoji9XMPkU1MQOtEnLIDopkgizgpNO5PtjSbTG0L5zOhaDjtWoYLOonKX6ibIYiTCrgWHTFitWq7WiJldWe3ezgNHBHmRD9bCrVvUeV36vKJV78cKN/7mkxnnhhGNNU9gPRFjUwIrRQgjxLxN22N6xYyexsbG1HouLjWXnrt1/u1JCCCGEODZVBma328vS5SspLS+vGJ5NtT91klPTaNO2HX4dfL7AfOZSp5/8Yjsb/vitav9kb6BH2u3T8JTmorlK2J48pcYCX0b/CqJdvxDhVoh3VAU7q6Kzsp0Xj1Envmw6Zn9a8JjLvJP82F/ZH2cgtYSQnmOnRac0UifO7gsOnwYCQ8B1UHQzflULuSbGqdAiT8XqUVDQGdDKSotO8USaVez6VKxmaBIZw19LfyF33y5aZLWgc5duPNS0rv1p44FWADUCcuW+wTVCc7XyymAd2PNXepCFEOKfFnbYbtO6FS++/Cr9+vYmMjIyWF5ebufFl1+lTetWDVpBIYQQQhw5uq7j02Db9l3k7i+oWgxM03G63GzZvJUpZ54ZMoQb4IfvvuPH776u+8bJXShKi6HYEVgcrJLRV0xa6Ze4jTrxdhWzvyoklkZprM8Er+F1UkouDLldXuxmtjT10Xy/jci9VQt26eiU2vLRVYixO8FfdY2q2VA1K6pqxdVsCJakNkRYVaIsKr2UrbiVUlo2Tad7WiJRVpVoi0qkRcVifoJoixGnvRy9Wtg+kMVixWyuHGYeURF8FTpNGoehIiQb1NCQbDggLKuqDLcWQohjVdhh+847buWCCy9l4JCRDOjfN7hA2pKly9E0P2+89lJj1FMIIYQQB1Fud5CTkwdAi8xmGI2h/8Rrmsab786peFS1anblomB+t52BJ4ygabMM/H4dvwa+ivD84D13UV5eVuvzmqMS2PXxOhymFIodGsUOPyUOP4ZihQx0fCrkR7fE5rWEXLfT5iJf/wCrIYNYqqam6YqZxe38aKpGp10tibdHB4+V2sooidyGoheio6NUG/atKmmYtBJI7Iy16SiirWogIFtVrO7fiTab6d6tLc2iY4myqkRZDERZ07EYh9Xxjjav+81WwKgq2OJjAmFZVarCs0JFQK4M0hU90dK7LIQQ/zphh+1ePXvww/df8vobb7NmzZ9s3bqdlJQkzpo2lQvPP5fk5OTGqKcQQgjxr1FYVILf58NitRATHVXj+JZtO1i9Zj15+wtYvmI10dFR5BcUBo8//sg9RMfE4td0fFogOHu8Pr77ceFBn9eU0pHokgQK7f6QL6ffhoFA2LabQ4c8l/nMLN27ArvFRKSrPWZ/xe8Btrb8lmUmN+FD0DeTmXd9SDgujvwep2Utih76q4jfEIGiR6HqbnxNehJl6UCMVSXGZkAxl9PU+ydJ1mhG9E0kLsJIjM1AjFXFZj7vIGF2ZN0vWgFDRY9zjeBcrac5NExLaBZCCHFoYYdtgOSkJG664bqGrosQQgjxr+Pz+fD5A2Obd+zcw90PPBk8NnzoAC658KzgUG6/puPzw9Lf1vLFl98Ez3NXC9rx8QnsL9dw4kPXdewenfwyH/vLvGR2HRbcs7n6n25voAf7peUqDstWfIYyLN70YDg2xp5KSdM/KIpaj8UbSXLpKSGvIS/2JXzGYkz+uKqwrShV+y8rfowmD0m2COIiDMRFGCgxtKeISFo2bcaQtERibYFAHWNTsZrvwWqs7VeUBA7a4wzBXufK3mSDGgjLlWXGiscGCc5CCCEa2WGFbYCSkhI2btpCdnY2w04YQmxsLG63G5PJJCtOCiGE+Nfzer2sXb8Jj8fDX5u20aZVCwb27xVyzroNm3jkiRfxeLy13sPtV9mZ78F/wGJg7Tr35Isvv6Fbz944PRqxqVn4zHG4TUkUuow8+r2L/LJy8sv9uLzV5xR3B8Bu2YDHVoDNkIXVmA62wFGPIZ99SS8CkJF3NQY9cMBnjMZnMOIzFmHQbMG7WU0KcTYVjzEDv5pAz+bxdIuLqQjUKpHWBMr1G8iMi6NJZPQBvx8kAQPq9V4qB4Rko1rV81xbiBZCCCGOBmGHbV3X+d/jT/HW2+/idLpQFIVPP5lNbGwsl115Dd27deWamVc0Rl2FEEKIo97Gzdv45rsFLFm+MqTcZB5LUlo2iUnJ+DTw+XU0SwJnnjOd2bPewe12B88dMGgwJ46bQJFDY81uN/vLfOSX+8kv85Nf+X3r61ibF0jhWokHRTei4AYC93GYt1Aa+RsGLZLk0lND6lIa8Ttu816UMgNWb3qwPNgTrUNynItmkXEkRBpIiDLgMXSjSE+iVWwqg9KbEBehYjVVhudLD/KORNZ5xBAMzwoGQ9VQbmO1IdyVQ7uFEEKIY03YYfvxJ57mnXfe49abb2TgwH6MHF01lGzUiOHM/uhjCdtCCCGOC5qmsX3HbhYuXoHRaAg5dtbUCaiqis8fmBft88PaDRt56qnna9ynRVZLdmfn002JoMRREZB1HReR+KIyGXTGDewv08gr9ZFnh8+KNd54YW9weyufWo7PUIyqWTH7k6rqp7jZk/Qimuqiaf5FocdUFy7LTgz+mGCZ2aiQGGlAsWTgVC10zUhiUHIsCZEqTa2l2KKb41XvJj0mDpMa+nohDmhbr/fNoILRoAR7oWuEaFXBaJAALYQQ4vgWdtj++JN5/Pf6azhr2lT8fn/IsebNM9gl+2wLIYQ4xjmdLt5+72N+Xri0znMGjzwJDZXqOz8ZbXFkNG/O7l276N6zFy3adCQ2vQO5pX7ySn28scRJXmk5uaU+9pf58Fb7Z9SnluMy70RT3MQ4e4Y8V2H09zisG4l29CKxbFSwXNHN6ErgJrqxlLTIVJKijSRFGVCsrdnlG0lmdBITxqSSFGUk0qJULCJ2TuiL0TUi/WXYDSZQEut8zZXDuY2GqsBcGaKNEqKFEEKIEGGH7aLiYlq1qn0vbU3z4/PVPu9MCCGEOJp4PF6cLhclJaWUljlp2bolXr+Oz6+zv6CcrLadQsJ2YlISHTt3CT72+UFRdEqcGtnFPrJLfGQXG8m3dsDd5WQ+LYmgfJEG5IU8b0nECnyGQqKVHphJCZZ7jfvJj/0cdAPRzh4hK3ebtGjQFWIi/AxKtZEcbSQp2kBSlIF93gtpFhtN5+Q0Ik3Was+UCLQJ6z0xGQNhuXqvtNGghPRICyGEEKJ+wg7bWVmZ/LpoMYMG9q9xbOmyFbRtE94/7EIIIURj03Udu8PN8pVrKC4p5/MvvsbpdAaP9xswkJG2BCIjA3OWzbYYomPjOfGUcbTt0BHVlkhOia9aqPbx9Qd5ZJf4DliADKADDv9myq3rMRnjibcPDTlaZluJz1iM2dsEsy8QthUg2ZJICTHEmuKZ2NtC87goUmIMpMQYsVnOxGY6t5ah3QBdaimrqXJodyA8B743VW53pegU5kF6vCxyKoQQQjSUsMP2heefx62334XJaOSkE8cAkJOTw8pVq3nr7Vk8+vD9DV5JIYQQoj68fj3ky+cDT0Vv9btvvcWGdWtrvW7H9u0UFJWSazezp8jLniIfe4tM7C3qwN51Xjy+7BrX+NRyCqO/x2coIq3oLFTdUlUPQyEO61+YvU2CYTsh0kBKjAGLsS2K0cWQzKYMSU8mpaKX2mRQgAfqeGWH/uc6JEQbAkO9jWpVuK57D+rA3HQhhBBCNKyww/bk0ydSXFLC008/z/MvvgLApZdfTUSEjeuunckpJ5/Y4JUUQgghKmlaZZiGguJSPv/iG35d+GvIORMmTaZv/9Btpdq17xAM22nN26BbYjFlDiXHbuKPIh8/fOIHcms8X1HUAhzmrUQ7uxLj7B0sV3UTDutGALyGImz+NFJjjDSJM6LbWrNXc9I2Pp1pbdJIiTFiNlaG3fMO63UbQoJ0IEBX//5gYVoIIYQQ/7zD2md7xoXnM+2MKaxcuZrCoiLi4mLp0aM7MdHRDV0/IYQQ/1L+ilDt8emB3ulqvdQAc2a/z6rff6v1Wk3XyS72srPAy65CH7sLvewtzGBvi5m4fAbWaoAT+EsD3LhMeymOW4iu+GlSdHbIvXxqKV7Tfjze/ZgMBAN1k9goVnuGkh4VxyknZNEhObHa4mDJVO5pXW9KIDybVAWjEUxqVbA2GSRMCyGEEMeasMP2J3PnMXzYCcTHxzFkyKCQY8XFxfz40wImTZzQYBUUQghxfNO0QJiuDNVeX+B7vwbLly7B6XAwZNjwGnOJIyIigt+bzFaadhtNmTWLvUU+Hv/NgHtFTo3nyo/+Do8pj/iyYdi8zasd0XFZdoIOOj5SYyw0izeRHm+k3NQPr9qGvk2yGJLRDDUk9E4N67UqSlXvtMlQEa6DgVrCtBBCCHE8CTts33jz7Xz84Szi4+NqHNu9Zy833ny7hG0hhBA16Hpg6LfHp+P2aXir9VTn5ebwx6pVeL1eXC4nv69YHrwuOjqG7du2cvLp57CvTA30Vhd42V3QGlVdxqbU6fgNEayqmFbtNO+lPHodBi2KhPJhIXVwm/bhNe3Ha8wjnkzS44ykJ5hIjMlivXsEHRLTmDSxCbGW6qt69wr7tRoNCmZDRQ91MFxLoBZCCCH+TcIO27p+4KqrVUpKSomMjPxbFRJCCHHs03Udty/w5fHpuL2BgL0/Nxe/XyM1LS2kp7qwsJCdO7azfdvWGvcqKytlR9RQ5r5TECzT8FBu24SnXQviyxUM1f5p8hqKsNvWYfIlQPkw4iNVMhJMZMSb2M0ArBY/o5p3pG/T9AOGZk8K6zWqFXOmzUblgEAtQ76FEEIIUc+w/fOChSxYsDD4+NXX3yIpMTHkHLfHzZIly+jYoX3D1lAIIcRRrXIYeCBQB768vkD6LS4uwuN2M3/uJyFB+o57H8BiseDwaOzM97Ldm8GeQjf+xC6UeC1omobLlESpzUKpzYlBM2HzVj2ngkph9I+g6ES6OmHzNifaqtI8wYQtpjXbtUI6JjTjkklNibFV3y5rTNivz6ASCNQVodpcEbBlz2khhBBCHEy9wvb27Tv44aefg49X/PY7ZrM55ByTyUTbtm3473XXNGgFhRBCHD00rVpvdUWPtddf+4in2268vs77PPltAdsLFbJLfIH7Km68cQOx+JqCreq8/OivKI9YQ4SrPbaSTABsJoXmiRFgakesxcKkrikMa96UuAi1okc5BegU9murHqorA7XZoKBKqBZCCCHEYahX2L7g/HO54PxzARg6fAwvPf80HaQHWwghjmvBoeAVPdYuj0Z2Th7vvPkaJpOZypHS8QmJnH3e+SHXOjwaRpMZn9cDgGawovpdbEk5C6c5GWW7P3iux5jHvsQ3AMjIuwaDXjVfOtGYhk/ZStukaK4YnESLZBMp0YaKUH3VYb0uVSUYpk1GBUtFj7X0VAshhBCiIYU1Z9vtdtO2TWvKysobqz5CCCGOEK9Px+XTcHt1XN7A0HAqOq1Liot55IF7a7/O62NDtpstuR625HnYkuthb5EPW9x4dAw4zWmgKBRG/YTd+jExjp7EOvoFrzf6EkBXAJ0erR30b5pGiyQzLZJMWE3pGNTxh/eCKrbSshireqnNRqXa9lxCCCGEEI0nrLBtsVhYvuJ3LrpwemPVRwghxD/ArwV6rF0+HbdXw+3T0bS6z/9rw3osVitulwsAg9FMfPdJ7Cn0saJc4eeP8gDwqWXYrRvwRZWSyKiQe2iqC7+hFL8lh64JFlqlmMlKNpGVZKZcuZEWMSlEGC2H9XpUlUAPdUVPdWWwloXKhBBCCHGkhL0a+ZDBA1n462IG9O936JOFEEIcFSp7rV0VvdaVC5gditOjsTnXw05jZzz6d5THtGFHzEh0NLz5+ahaLEZzdPB8v1pGUfRPAMTZBxNptNEq2UzrVBNEDABLB4Y0a0WbuJQDnimj3q/FWBGkgz3WRtlSSwghhBBHn7DD9uTTJ3L7/92N3W5n2AlDSUpKrNFz0LlTxwaroBBCiPB5fDpOj4arYq61/yC91gBul4utW7ewfU8emzZtobi4iNKmp7DDEYtWmcuTLwyenx3/Lh5zNvFlw4JDws1GhbZJzfAa0mgZ3YwrhsTRNS2p2lzo+LBfh8mgYDZVBWuLrAIuhBBCiGNE2GF7xiVXAPDurA94d9YHIUFb13UURWHLxj8broZCCCEOqTJcO706Lq920CHhlfyazjuzPmDzn7/Venxf0Trykorwqw5Si6eEHDP7UvCYs0lJKuPiNvG0STWTkWCqCMK3H9ZrMBoULBKshRBCCHGcCDtsv/fOG41RDyGEEGHw+nQc3oph4Qfpud741wbmz/2YXr37MmT4aDbneVi31826vW7+ynbTYscfoOgURepEeBQiPArFtnYYdDd2SwrltoUAWCxOOqbG076JmXZpFqJjJhJrnUqyLfaw6q+qYDWqwXBtMUmwFkIIIcTxJeyw3a9fn8aohxBCiIPwazoOdyBR7yn04tPqDqYFBfksXvgLSxcvCpb98MNPPL+pA17/AedG9WR9xjLKbR7iS4cQ6xyIAmQkGBmTZmSlbwu9U7M4p2M6cZbIaldaqTeFYKCuDNgyx1oIIYQQx7uww3alTZu38NtvKykuKSEuNpbevXvStk3rhqybEEL8a+l6YCEzh0fD6dHx+HTQNSIBn1+HWlbZ9vp13p/9MRtXL6lxbHdcG3JtP+AzlJJSMjFYnhc7AANlwGpapvuZ2SmZtmlmIi1qxRkXh133yuHgVmPVsHBZFVwIIYQQ/zZhh22328P1N9zM1998h67rmM1mPB4PiqJw0oljeOzRBzGbzY1RVyGEOK5VDg13unWcXg39EAuG+zWdHfle/tjt4s89btbvdRK/30GMQUcHNCWJ/OjeFEe0x23KoTTybQBMTgddmyTQsamFTukW4mMmE205m0hTGL3VlQ7otbaaZB9rIYQQQgg4jLD9v8ef5Keff+G+e+7glJNPIjo6irKycr748ivue+Bh/vf4U9x68w2NUVchhDiu6LqOs7L32q3j9R88Xeu6zp5inSV7y1mzx8PaXWU4XF78BlvFGQprMkuwWz0klvYn2nUCALE2lfZNMllNe3qktGDGqU1Ijoipduf6722tKGAzBYaCWyt6rVWZay2EEEIIUUPYYfvzL77ihuv/w5lnVK1MGx0dxZlnTMHpdPHyq69L2BZCiDp4/VVDw52eQ/deOzwaf+5xs2qnk5U7XeSW+kHfT4Tne1Q2EKFNwGVKxmsMLFSm6pGggDmhgBmt4+iWYSUjwVgxjPuqsOtrUMFiUrFVhmuTeuiLhBBCCCFE+GG7uLiEVi2zaj3WqlUWxcUlf7tSQghxPHF5NexuDYdHx+s7dO/1jnwvK3e6WLnTxV/ZbnxayAk0K3qWXzt4ARi15jMK4k8jMyOFLhlWEhJGEhkxmB7JWZgNprDralDBag6Ea5tJxWSUXmshhBBCiMMRdthu1TKLufM+Y8iQQTWOfTrv8zqDuBBC/Ftomo7DE+jBdngOved1qdPP6t2B3utVO10UOQIXuEy7sUdsxOJLI8rVmUjXTrLy56KjYPKB1aPQdcwEzho0pFoojqn7iWphUMFmDsy1lnAthBBCCNFwwg7bV115GVddfR179u7lxDGjSUpKpKCgkK+++ZZVq/7g2acfb4x6CiHEUc3n17F7NBxuDadXh0MMD99T5GXFdifLtwV6r7Vazrdb/6IsYiVedyYtTd1I2TMXAAWFkX+aefCB/6Gq4Q3rlp5rIYQQQoh/Rthh+8Sxo3nx+ad4+pkXeOChR9F1HUVR6NihPS8+9xQjRw5vjHoKIcRRx+vTKXcHhoh7DjE83K/p/JXtZvk2F8u3O9lX7AuUKy6Ko37Bad5GWtE5GLUo0P1E+/NpGtuebGMxI9t24aruTfjw/e6sWb2afgMGMH7i5HrVUVECPdc2k4LNrGKWcC2EEEII8Y9Q/D73Ifpf6uZwOCgtLSMmJpqIiIiGrNc/RtM09ufsJjElHaPxsLcdFyJEZbtKTssIu+dRHN3cXg27R8fu1g45/9ru1li108WK7U5+2+Gi3K2ho6EQaBNmXzEGfxmbm76Pz6jTdVcUWWVW/K4yAO558BEMBkPVDXWNSH8udkMqKHW0KwWsxkCwtpkVrLKgmagH+cwSjUHalWgs0rZEY9A0rcHb099KlzabDafThc1mO/TJQghxjKpc4Mzu1vEdYnuuYoefJVudLN3iYO1eN5rPhc2TR2FUOYXxK1F1C6nFU7B4C2he8DkWrQxdVVF1SC724Pd6q+5VXERiYtIh62cyKNgsChGmwNxr2YpLCCGEEOLIO6ywvXDhIp5+9nn+XLsen8+H0WikS+eOzLzqcoYOGdzQdRRCiH+c26tVDBE/dMAuKPexZKuTxZudrN/nRtc9xDg2k+7cTLRrBwBOc2fc5r0ouoFB7QwMatmWjqnXMX/OeyTm76estJRzL7wIi8UKQExsDNHRtS92pihgs6hEmAM92CaDhGshhBBCiKNN2GH7ozlzueW2O+jTuxe33PRfkpISyc8v4Kuvv+XCGZfzwH13M3XKpMaoqxBCNCpP5Rxsl4b3EAE7r9THki1OFm918Fe2BwCbOxvFPIu9iX7iHUaiXVVDwIekZdGuaQQT2nSjT0oqRjVwbPqFM+pVN5MxsKCZvwyaJ5pCh5cLIYQQQoijTthh+5nnXuD0Safx8IP3hpRPP+9sbrjpNp59/kUJ20KIY4bXH5h/XeY69Bzs7GIvi7c4WbLFwZZcJ5piQKGqV9nkL6XYrOMxQXa8RnqRgax2nThxzCiaZTQPq16VC5tV773WNI39ZaAo0pMthBBCCHG0CztsFxQUMu6Uk2o9Nn7cyXz51Td/u1JCCNGYNC3Qg13u0nB5Dz1E/NdNThZsKGRXbhkppUuxetdT2iKB4shy0vdfjKqYUYCM9HSyTP0xx/k4b/QYmiekhVWv4NzritXDJVQLIYQQQhy7wg7b3bt3Zd269QwZPLDGsbXr19Ota+cGqZgQQjQkXddxenTKKrbqOtg+2GUuP4u3OPllo4PCtV+SYP8TG9Cu4rhXhaKoHDQVUtL3MLltL/q3shEfaQA61L9SFSuHR1pUIsyy57UQQgghxPEk7LB9w3X/4Zprb8Dt8TBm1AgSExMpKCjgm+9+YO7c+Tz1xKMUFxcHz4+Li2vA6gohRHg8Pp1Slx+7S8Ov1X2ey6uxfJuTXzY5WLXThU8DtzEXQ9JacuJ8dNxb9XFpwsSgyC4MaNOGU1p2J84SVe/6KApEWFQiK7bmMsjK4UIIIYQQx6Www/bpU88C4OlnnueZZ18Ilut6oJto8hlnh5y/ZeOff6d+QggRNk0L9GCXuzTcBxkmruk6f+5x8+N6O0u2lBNdvJL86N7B4x5TDgUxLgx+lfaxUxneqz0DWkUSHRHeR6fRoBBhVoiwyPBwIYQQQoh/i7DD9sMP3iu/KAohjkpOT2Chs/JDDBPfV+zlm1W5LFuXTYlTw+Rbh275A4tXxeRri9cY2HKrS0xndto2MaFVL6a164HVaK53XUwVw8MjzQoWk/p3X5oQQgghhDjGhB22J58+sTHqIYQQh8Wv6ZS5NMqcB9+uy+7WWLTZwY9/5FC6eQHxjvUkAonAonYeiqJ0XGaddmW/0WfwaQxtG0GTOBNwXb3rYjYqRFkDQ8Rl/rUQQgghxL9b2GG7UklJCRs3bSE7O5thJwwhNjYWt9uNyWRCVaUXRwjRuJwejVLXwRc7qxwm/v16O0u3OHH7NWAdqaoVrxqBSXPgNsbRpNiBy1xOh4yOPDz+vLBG71hMClEWlQhLYHsuIYQQQggh4DDCtqZpPPbE07z19rs4nS4UReHTT2YTGxvLZVdeQ/duXblm5hWNUVchxL+cX9MpdwVC9sH2xC6y+/l+fTkLlv+Ft2ALObFD0FQPe5Nex28oRVPHoxgsZLbuyIQ+WXTJMGExqqhK/f6j0GhQiLQoxFgN0oMthBBCCCFqFXbYfuLJZ3jnnfe49eYbGTiwHyNHnxI8NmrEcGZ/9LGEbSFEg3J7AwG73KWhH6QXe/UuF9+utbNpzTLSin4gFgWXMZ52jr/Y2ORijP5o/IZSOnX0cNuAicTYDPWvhAKRFpVoa2CbLiGEEEIIIQ4m7LD98Sfz+O/113DWtKn4/f6QY82bZ7Br1+4Gq5wQ4t9L13XK3RqlzoOvKF5Q7ueHDXZ+Xr4Bc/avaGoRhUllbEvRGLjRhNVXRFHycE7rFUNm+hm0S4klPTKx3vWwmhQirSpRFlW26RJCCCGEEPUWdtguKi6mVatWtR7TND8+n/dvV0oI8e/l8+uUujRKnX60OvbF1ivmYn+5ppxl25xomk5ayRZU3YOu29meGviPwKKOQzln0Cj6toysmE8dV686mAyBhc6irDIPWwghhBBCHJ6ww3ZWVia/LlrMoIH9axxbumwFbdu0aZCKCSH+XVxejRLnwRc8c3o0ft7o4Ms/ythV6MNjKMBvKsXmySIndggJ5auJ98fRXnExqGV7pncZSqwlsl7Pr6oQZQkEbKts1SWEEEIIIf6msMP2heefx62334XJaOSkE8cAkJOTw8pVq3nr7Vk8+vD9DV5JIcTxSdd17G6dEqf/oEPF9xV7+XJNOT+uK6PJnllE+kowNIllX+pejL5YmhVcSq8WNsZ0HkXvFjaMYfRGR5gDATvSooS1CrkQQgghhBAHc1j7bBeXlPD008/z/IuvAHDp5VcTEWHjumtncsrJJzZ4JYUQxxdNqxwqruGrY29sTddZudPFF6tK2LJxHQn2tWS5dwWPN8/fz9Y0iLGaePDMCDqmJNX7+U0GhWhbYB52OMFcCCGEEEKI+gorbOu6TklJCeeefRbTzpjCypWrKSwqIi4ulh49uhMTHd1Y9RRCHAe8/kAvdpmz7lXF3T6NnzY4mLeqjH3FPmzuBZRHL8drVIjKC3xk6aqJ0eOvZHqGl+5pWfXqkVaqrSZuk9XEhRBCCCFEIwsrbHu9Xvr0H8pLLzzDiOEnMGTIoMaqlxDiOOLx6RQ7/JQfZD52icPPl2vK+XJNOaWuqpXR9seWkx+rYfVAG1dzxp48nkE9Wtd7yLfZGOjFjraoqLKauBBCCCGE+IeEFbbNZjNpaalomv/QJwsh/vVcXo1iu4bDU8ey4sCeIi/zfi9izfKFaP4yMCVARI/g8TTDCZjMRZzXbShTzu6PST303tiKAlHWQC+2LHYmhBBCCCGOhLDnbJ979jRee/1thgwehMViaYw6CSGOcQ6PRrHdj6uORc90XWf9Pg+frixl1eb9tMt5jZJUH1vS/DQpTCK1xE10y8Gc1iuGQa2bYTTcUq/nNRkUYmyBkC292EIIIYQQ4kgKO2zvy85m+44dDB46in79+pCUlIhC1S+1iqJwx//V7xdjIcTxxeHWKHLUvbK4ruv8tsPFRytK2Zjjod2+V2in2QHQFPAaIT+hjCt7tGfsoLR6DxWPMKvERKhEyFxsIYQQQghxlAg7bP/40wLMZjOYYc2fa2scV5CwLcS/jd2tUWT34/HVHrL9ms6SLU7m/FbKtnwPLtMuIjQjpoqgDdCG9vRvncWMLkOJMUcc8jkVBaKtKrE2Ayaj9GILIYQQQoijS9hh+5efvm3wSuzdl8Obs+awefN2rFYLQwb15YzTx2E0Hrx6ZeV2Zs/5jNVr1lNebic5OZExI4cyesTgBq+jEKKmQ4Vsn1/nl00O5vxWyt4iHwD7Y+fhsG4kxt4Xb14sJl8J19x6Hylxtno9p9GgECtDxYUQQgghxFEu7LDd0MrtDu575BnSUlO4duYMCouKeff9uXg8Hi44d+pBr33yudfYl53LmaefSmJiPKvXrOf1t2ejqgojh8lK6UI0Fodbo/AgIdvj0/lhg51Pfi8lr9QPuk6CfQ2FkV2weNNxWDfSLNXOXWfdQoTVVK/ntJoUYiMMRFpkqLgQQgghhDj6HfGw/f1Pv+J0urh+5gyioiIB0DSN19/+kAnjxpIQH1vrdcXFpazfsJnLLjqbE4b0B6Bzx3Zs3b6LJctWStgWohE4PIGe7LrmZHt8Ot+sLeeT38sosHtxm5bRvGQPKWV7AUgr+ZVeE6+gZVZ3eqa2PPQTKhBlUYm1qVhkVXEhhBBCCHEMOeJh+4816+nSsV0waAP079OD196azZ9rNwSD9IH8/sD2YxERoUNPI2xWXG5341VYiH8hl1ejsLzu1cW9fp3v19v5aHkpBfbAz2ZBzFuU2/Kw+FVSygK916ruZXL/5od8PlWFGJuBGKuK0SBDxYUQQgghxLHniIftfdm5DDsgUEdGRhAXG8O+7Nw6r0tMjKdr5/Z8+vm3NElLITEhntV/rufPtX9x5aXTG7vaQvwreHw6hXY/Dnft+2T7NZ2f/rIze3nFcPEKzQq+ItpVyNrmoAMZmZkMPWE4HTt3OejzmQwKMREqMVa13iuRCyGEEEIIcTQ64mHb7nDU6J2GQOAutzsOeu21M2fw9PNvcMNtDwCgqirnnzOZfn2613mN1+vF6/MFH+t6IERomoam1R4ohAhXZVs6VtuUz69T5PBjd9Vef03X+XWzk/eXlbK32Ivdug6LoRkmfzwA1ub9id70F8mlZk4bcTInDB8RuFCv/X4Wk0KszUBExXxsXdfR9dp70f/NjvV2JY5e0rZEY5B2JRqLtC3RGDRNQ1Ubdtpi2GF7wS8LGTpk8BHvddJ1nRdfnUVO7n6uumw68bGx/LnuL95+7xMiIyIY2L9XrdfN+/w7Pp73VfCxxWLm4Tuvpig/+5+quvgXKcjbe6Sr8LdEHvBY13UWbdd5c7nGjsJAWV7cXJyWLUQ6O9LHeCoX9FUZmJXEF9/258SRAzEaDeCve5QKAH6wu8B+8LNEhWO9XYmjl7Qt0RikXYnGIm1LNLTUppkNer+ww/aFMy4nLS2VyZNOY/LpE8nIaPa3KhAZEYHD6apRbrc7iIqse6/dVX+sY9mKVTx87y00z2gKQMcObSgpLePdD+bWGbYnjBvNyScODz7WdY3y4nzik5occqsxIepL0zQK8vaSmJLe4P9D1hh0XafUpVHi8FPXfxKv2ePmrUUlbMnzhpRHujrgNG+lcxMLj41Kw6gacADDT56MG6htBYUIs0pshCx6Fq5jrV2JY4e0LdEYpF2JxiJtSzSGxhgpEXa6/ObLeXw0Zy4ffDiH5198hb59ejF1yumcOHYMFos57Ao0bZJaY262w+GkuKSUpk1S67xuz74cVFUlo1mTkPIWmc346ZcluN2eWutjMpkwmaq2GtI0jXICQ9Dlh1U0tGOhXZW7NArtGj6/DqhwwKCVXQVe3lpUzG87XDjN2zEYojH7kwBIiTFwVZ+BfPHBN/RtY2Dr5i20a9+hzueKsKjES8j+246FdiWOTdK2RGOQdiUai7QtcbQLu3W2bt2KW27+L4sX/sgLzz1JVFQUN958G/0GnsAdd93Ln2vXhXW/bl07snb9RuzV5mcvXbEKRVHo0rnuX9qTE+PRNI1du0OHj2zfsZuYmOjDCv5C/Ju4vBp7i7zklfoqgnaoQruf534o5Jr3cvhth4uC6O/Ijf+QoqhfSIw0cNnweJ6elszPb91FpFth3do/+X35slqfK9Kq0izBRFqsUYK2EEIIIYT4Vzjs33oNBgOjRo7gpReeYcGP39KhfTtmvTebiaefycmnTuSjOXPrdZ9RwwdjtVp47OlXWLN2Az8vXMp7sz9l1PBBIXts3/fwM/znxruDj7t37URSYjxPPvc6CxevYO36jbz34TwW/LqMsaOGHu7LEuK45/Xr5Jb62Ffkq3W/bKdH472lJVz2VjbfrrOjVZxic2cBkJ6g8My5ybQw7OS+/7s55Nqzzjs/5HFlyE6NMWI2yuriQgghhBDi3+NvTVLesmUrH875hE/nfUZpaRknnzSWMaNH8vOChfzfHXez+o813H/vnQe9R1RkBLffOJM33/2Ix55+BavVyvChAzlj8riQ8zRNw++vGkdvs1m57caZzP74M97/cB4Oh5Pk5ETOnTZJwrYQtdA0nWKnRrHDH9iP6wB+Tee7dXbeX1ZCgdOBrrgxEvgPL6MKUzt0o0ubTPo3bcnOHdt5983XQ66/5f/uCn4fYVFJiDRIwBZCCCGEEP9ait/nDmt/Hbvdzmeff8lHc+byx5o/yWrRgjOmns6kiRNISIgPnjf30/mBYeWrVzR4pRuSpmnsz9lNYkq6LJAmGkxlu0pOyzgq5hIF5mX7ax0uDrB6l4tXfylid6EPu2UDBTHfYvamklp8BkPaRHDuwDjSYqt+PsrLynjw3rsA6DtgIGPGnoQtIkJCdiM72tqVOH5I2xKNQdqVaCzStkRjOCq2/uo74AQAThw7mptuvJ6+fWpf9btjxw7Ex8fXekwI8c/w+HTyy3y4ahkuDpBT4uP1hcUs2+YMlqlaBJrqwmfJ4dSMFexfsQZX++m4LU2xWK0ARERGcusddxMZFRV4LCFbCCGEEEKIEGGH7ZtuuJbTJpxKTEzMQc9r17YNv/z07WFXTAhx+DRNp9Dhp9Sp1Tpk3OnRmPNbKZ+uLMOraSiooOtEeLJpYdBpY29OxOYcdvgXA/DKC88x47IryGrZCgis/hkZFYXVpJAQZcAqi54JIYQQQggRIuywfd65ZzdGPYQQDaTcpVFQ7sNfy1aBuq6zYKODtxaVsN9RRnHUQjzGAtKKppGo76Pp/o+qnV3VSz3j0qqgDWAyKiRGGoiwSMgWQgghhBCiNvUK208/83xYN7165hWHVRkhxOHz+nUKyvw4PLWkbGBngYcXfypm/T534HxTIWURqwHo0HYPtw7tywuPfUNZWSnNMprj9XoZOmw4HTp1xmKxAGBQIT7SQIzN8I+8JiGEEEIIIY5V9Qrbr7/5Tshjr9eLy+UCwGKx4HYHfnm3Wq2YzSYJ20L8w0ocfgrtfvQ6hozPXl7K/NVl+DUw+UpJLltOgn0tfyb04sR2WVzatT8GVeXaG28OBusQCsRFGIizqaiqzMsWQgghhBDiUOoVtlf/viT4/Zo/1zLz6uu56srLOOnEMURFRVJebufLr77huedf4umn/tdolRVChPL4dPaX1b5ftq7rLNvm5JUFxeTYi9EMX9Jj1z6UasPD72/fgy7dugcf1xa0I62Bxc9MBgnZQgghhBBC1FfYc7bvuvt+Lp5xAVMmTwyWRUVFMnXKJFxuF3fefR+ffjy7QSsphKipuKI3u7YF0HJLfbzycxErdrjwK06yE1/Gb/CRVGakeUFgCHhCQiKRkVF13t9sVEiKlsXPhBBCCCGEOBxhh+0Nf20kIyO91mOZzTPYtGnL366UEKJuXr9OXmntvdl+TWfeqjLeX1aKxxc4btBtpJaYKYj2YfTD2JNPof/AQZjNtQwXR+ZlCyGEEEII0RDC7rJqlp7Oe+9/iH7A5FBd13l31gekN23aYJUTQoQqdfrZU+itNWhv2+/hhtm5vLE4D7evapG0GKvKBV1nMGydmfP6jWPosBG1B20FYiJUMhJMErSFEEIIIYT4m8Lu2b7xhmu5cua1jBh1EiNGDCMxMZGCggJ+/PFn9u7L5rlnnmiMegrxr+bz6+SX+3G4a6407vHpzF5ewie/l1Fu3kxhwlekFHfA7B/NmE6RnDswFptRI807jkFDTqj1/laTQlK0EbNR5mULIYQQQgjREMIO26NHjeDTjz/gxZdf4/vvfyJv/35SkpPp1q0Lzz3zBB07dmiMegrxr1Xu0sgv96HVsqPXur0unv2hiH3FPtB1jP5v8BmdFMSt44muoxjUJaHiTANDh42ocb1BhcQoI1FWmZcthBBCCCFEQwo7bAN07NiBp5+UVceFaEx+LdCbbXfVTNkOj8bbi4r56k876Drx9rWkF/+A16Bj0FXa7tNpM6r2OdmVYmyBVcZlKy8hhBBCCCEa3mGFbSFE43J4NPLL/Pj8Nedm/7nHxdPfF7LXno/PnEefbV8Fj5n8Ct12mrj1zrvrXGncZFRIllXGhRBCCCGEaFSHFbZ37NzJxx9/yvYdO3G73TWOv/LSc3+7YkL8G+m6ToHdT6mjZm+226vx9uISPv+jHKd5O3mJn6DoKkWRKcTb9wMw+Yxp9OjVu/abKxAfYSAuQkVRpDdbCCGEEEKIxhR22P5jzZ9MO/t80ps2ZfuOHbRv15aysnL27N1LWloqLTKbN0Y9hTjueXyBLb0qt+yq7q9sN099V0huQRmKasbsbYKqWbEYjHTr1YpEeyYTJ0/FYKh9FXGrSSE52ohJFkATQgghhBDiHxF22H74kcc45aSxPPTgvbTt0I2HHryXzp068vvKVVxz7Q1ceslFjVFPIY5rpU4/+eV+OCBne3067y8r4dPfisjKnUUHXxF/NZmByRjJFa0u5MwezYkyW+u8r6pCguyZLYQQQgghxD8u7LC94a+NXHbpDFQ1MN+zchh5r549uGbmlTzy6BMMHTK4YWspxHFK03T217EI2o58D49/lYt722wcKbvQFSOg0C5/NpdfdwsZiQdfAC3KqpIYZcAgC6AJIYQQQgjxjws7bCuKgslkQlEUEhMT2Lt3H7169gAgLS2V7Tt2NnglhTgeub0aeaV+vAcsgqbrOrN/2sySBd8R5dzK7x29lNt0fAYffbcYueGG/xAfV3fQNhoUkqINRJhlATQhhBBCCCGOlLDDduvWrdi1azcD+vejZ4/uvPr6W7Rr1xaT0ciLL79K8+YZjVFPIY4rdQ0bL3b4efq7QtZs9ZGkRhGFQoe9BtY093Fuj8lMv2ToQe8r23kJIYQQQghxdAg7bE87Ywp79+4D4L/XXcP0Cy7hlFMnARBhs/HcM080bA2FOI5oFXtnlx8wbHzTxr/I9cbzyhIfxU4/ijEau7U5dmsmI/p14IkBcURbDt6bnRxtwCa92UIIIYQQQhwVwg7bE08bH/y+detWfPv1fFau+gO320X37t1ISkxs0AoKcbzw+HRyS314D1ht/K+/NvLO66+wPfFU9iTtxhtbTHLJBExJ7fjPmAS6ZtS9ABpIb7YQQgghhBBHo8PaZ7u6yMhIhgwe2BB1EeK4Ve7S2F/mQz9g2Pj23HLeef1lACK9n1IS5QMgI24HD40dRLS17lXEDSokxxhlbrYQQgghhBBHocMK24WFRbzy2hus+XMt2dk5vPj807Rt05o33nyH7t260qNHt4aupxDHrPxyH6WO0GHjuq7z+eItLJ33YrDMrwwj1pVH34xkHjxhIKY69swGiLSqJMlK40IIIYQQQhy1wg7ba9et55zzLiI6Oop+fXuzbNkKPB4PALm5ubz+5ts80+OxBq+oEMeinGIvLl9oIPb4dB5+8iUceZtQCRzzqpHEtezPrLEJZCSY67yfokBStOGgPd5CCCGEEEKIIy/ssH3f/Q/To0c3Xn7hGRRFYe6nnwWPdevWlc+//LpBKyjEscjjDfRku7x6ICFXyC728vCXBeRqUezq6KVttoEoZ1u6jJzGBYPiMBnr7qm2mBRSYoyYDNKbLYQQQgghxNEu7LC95s+1vPDck5hMJvx+f8ixxIQECgoKG6xyQhyLylx+9pf6iDygfNFmB898X4jTq5OfXEi5TeePFiYe63kWw9vFHfSeMREqiZEGFEWCthBCCCGEEMeCsMN2hM1Gebm91mP7srOJj4v7u3US4pik6zoFdn9gfna1hdB8fp03fi3m8z/Kg2Xx5Sdhsnm5d9AkhjSPq/Oeqgop0UYiLLIImhBCCCGEEMeSsMP2kCGDePb5lxg4oD8xMdEAKCi4XC7eeOsdhg0b0uCVFOJo59d0ckt8gWHj1RQ7/Dz86Q60P14mIe4ESm2t8RkiGdc5kQuHXIf5IMPGbWaF5GgjRhk2LoQQQgghxDEn7LB90w3XMeXMcxgx+iT69+uLoig89sTTbNm6FQWF6669ujHqKcRRy+3VyC314/OHBu0NuTp3fZ2HJ38W2zq5GLDxRzxRLbj0xESGto2o+4YKJEQaiIuQRdCEEEIIIYQ4VoUdttPSUvli/se89sbbLFq0hObNMyguLmbCqadw0YXTiZNh5OJfpK79s79dZ+eN7/KJL/mRNW3y8RphXabKK1PakZFgqvN+JoNCcowBq0mGjQshhBBCCHEsCytsu91uHnr4MU477VSuveYqrr3mqsaqlxBHvSK7nyJ76CKBXp/OK78U8c2f5XTOfguAnttMbEv189iYKw4atCOtKslRBlTZO1sIIYQQQohjXljdZxaLhY8+/gSX09VY9RHiqKfrOrmlvhpBu9Du57ZP8vhmrZ2mRd8HyzPUJrw7/lY6ZbWs/YYVe2enxhglaAshhBBCCHGcCHsYec8ePVi1+g/69evTGPUR4qjm13RySny4D1gIbWueh/s/y2enfw2umJ0YfQPRbYkkFvzC7bfeUOeWXSaDQmqs8aALpQkhhBBCCCGOPWGH7WuvuYprr78Jg8HAsBOGkJSUxIE5QuZti+OR1xcI2t4DFkJbssXB498WUqrsJj/hcwDUtAzuOqELKWmn1Bm0IywqKdEybFwIIYQQQojjUdhhe/IZZwPw0COP8fCjj9d6zpaNf/69WglxlHF6NHJLfWhaVZmu68z5rYx3l5QAYCGdaEcPYqIdvHDKUDKtZdjV2mdqxEcaiI+U1caFEEIIIYQ4XoUdth9+8N46e+qEOB6VuzTyynxQrUPb49N57odCft7oCJY1K/yWMU07c9mkPhgVBfxlNe6lqpASbSTCIquNCyGEEEIIcTwLv2f79ImNUQ8hjkolDj8F5aELoRU7/Dz4RT6r87ehmd1EeFqR6PiTeMcGspdv4IlN33PDLbfWuJfJqJAWY8Qk87OFEEIIIYQ47oUdtoX4tygo91Hi0ELKdhV4uXf+fna4N5EXPxdFV+myfxBNCpcEz7n0ypk17iXzs4UQQgghhPh3qVfYPvGU0+o9dFxR4MvP5v6tSglxpO0v81HmDA3aa/e4eOCLfOxuHYvSFKM/GqtiIDl3MRD4+Tj7vPOJiY0Fvera2AiVxCj5fy0hhBBCCCH+TeqVALp06kiNJceFOA4F9tD243CHBu1fNzl44tsCfBXFBt3GCH006p/zMGqBn43xE0+nY+cuIdclxxiJtknQFkIIIYQQ4t+mXing0UceaOx6CHHEaZpObqkPpyd0a695q8p45dccQMeADYCTukRy4eBBvJz/O9n79tJ/4CD6DRgYvMagAn6IlIXQhBBCCCGE+FeSLjchCATt7BIfbm9V0NZ0ndcXFjN3zV5yEz5C1WykFp3B9IGJnN4rGkVROPPsc0lITESttsWXyaiQGm2iaP+ReCVCCCGEEEKIo8Fhhe3S0lK++vpbtm/fgdvtqXH8zjtqrsQsxNHKr+lkF/vw+KqCtsen8+S3BSza4sRj3o/XUIiiqoxos5lJPTODaxgkJSeH3CvCrJISYyBknzAhhBBCCCHEv07YYXv7jp1MmXo2Ho8Hh9NJQkI8JSUl+Hx+YmNjiI6KlrAtjhk+f6BH21staNvdGvd/ls+6fW4AItytaJfdhKTS/eTbv+GFTeu48ppra9wrxqaSFB34kdI0CdtCCCGEEEL8m4U9ofT+Bx6mW7euLF/6C7qu8/qrL7L+z5U8/r+HiIyM5NlnHm+MegrR4Lz+QI929aBd7PBz+yd5rNvnxuItINq5lc57n6JNdgHx9sCPy9Bhw2vcKzHKEAzaQgghhBBCCBF2OvhjzVoeeuAezGYzAF6PF4PBwITx4ygqKuaeex/go9mzGryiQjSkyqDt81cF7f1lPu6Yu58NnkVYjWV03rO2xnX/veU24uMTqgoUSI0xykJoQgghhBBCiBBhh22Px0N0VBSqqhIXF0tuXl7wWNu2rXn0sY0NWkEhGprXFxg6Xj1o7y3ycsfc/ezw/UlR7M+gQ66hP+m71mKz2chskcW0c84LuY+qQlqsEatJgrYQQgghhBAiVNhhOysrk7379gHQsUMH3n3vAwYPGojRaOD9Dz4iJSX5EHcQ4sipLWhvy/Nw16c5uMoKiDC1xWrNJNEazePTpxJjPafW+xgNCk1ijZiMsv+8EEIIIYQQoqaww/a4U05i/Ya/mHjaeK77z0ymX3gJPXoPQFEUdF3n0YdlT25xdPL6dPYVe/FrVWXr97l58t3vycj/EZ9qY2PahZwSfzY3npiK1WSo9T5mo0JarBGjQYK2EEIIIYQQonZhh+0ZF54f/L5Hj258/eWn/PLLr7hcbgYM6Ee7tm0asn5CNIjagvbKHU4e/XQ9TUp+ABSMmpP+ifu44ZRBGNTag7TVFAjaah3HhRBCCCGEEAIOc5/t6po2acKZZ0xpiLoI0Si8Fdt7VQ/ay7c5+eC1+8huWcqOBJ2Bf5lp2u4krj1nEKpSe5COsKikxhiCe2wLIYQQQgghRF1kryJxXKtt1fFlW528/v48jCYHpTYdvwrOZmlce+7IOoN2tE0lWbb2EkIIIYQQQtSTpAdx3PL5dXIOCNpLtjp49KsCjLZ2tMldQd8tJpoMPJU7zxxR533iIg0kRNY+f1sIIYQQQgghaiNhWxyX/Fpg6Li3Imi7XS5+36Pxv68L8GugGeMojOzMmBPHM6lvYp33SYo2EGOToC2EEEIIIYQIj4Rtcdzxa4Gh415fIGiXlpTw8P33sKb5hZRZ/yLG0QcUIyeNn8z4HtG130SB1BgjkRbZQ1sIIYQQQggRPgnb4riiaTo5JT48vqqh46+88S4eg05+9Ms4rH40xc2NvU9lXLfag7aqQlqsEatJgrYQQgghhBDi8ISdJhb8spD5n30RfLwvO5tzp89g4OAR3HDjrTgcjgatoBD1pes6uaU+3N6qoP3LRjuF+7Zh8kNimY6iG5ncsUOdQdtoUGgaZ5KgLYQQQgghhPhbwk4UTzz1LLm5ecHHd919P1u3bmPcuJNYsPBXnnjq2QatoBD1lVvqx+mpCtqLt9iZ99ZTACgopBSfwLWtZ3L1gB61Xm8yKjSJM2I2ytZeQgghhBBCiL8n7LC9Y8cu2rdvB0BZWTm//PIrt992E7fefAM3/vdavvn2+wavpBCHsr/Mh8NdtZH28s3FfPHyHVh9BcGyM8aPYnrvVrVebzEpNI0zYjJI0BZCCCGEEEL8fWGHbb/fh6oGLlu+4jd0dE4YOhiA5hkZ7N+f37A1FOIQCsp9lDmrgvafe1w8/sVafm2nUmoLlPc9/QbGdo2t9XqbWaFJrBGDKkFbCCGEEEII0TDCDtstW2Yxb/7nOBwOPpj9ET179CAyMhKAvP37iY+La+g6ClGnEoefEkdV0N6Y7ea+z/LZkbCc4igny9r46Tb5Zib0S6v1+giLSlqsEVWCthBCCCGEEKIBhR22Z155GfPnf07XHv34ZeGvXHHZxcFjC375lU6dOjRoBYWoS7lLo6Dcj8/nY97cj9mwu4S75+Xh8uoklI3E6ItjQvq5TO2bXOv1UdZA0FYUCdpCCCGEEEKIhhX21l+jRo7g268/Y936DbRv346sFpnBYz17dKN9u3YNWkEhauP0aOSV+QC4787b8Xq9LF+yGH/KNDCnYvIncEXz/3Dh4MRar4+xqSRFy853QgghhBBCiMZxWGmjefMMmjfPqFE+7cypf7tCQhyKxxfYSxsd/lq/Hq/Xi1/RUXRwmVMBOLlrFBcOjqu11zomQiUpSoK2EEIIIYQQovEcVuIoKSnh5wULycnJxe12hxxTFIWZV13eIJUT4kA+fyBo6zqUlpbyzpuvoaGzsqWPcktrkkr9jOwQzcUnSNAWQgghhBBCHDlhp46FCxdx5cxrsTscWK1WTCZTyHFFQcK2aBSaFgjaPn9gL+2H77sbgJw4jdw4DdhE34R9XDVyAKoEbSGEEEIIIcQRFHbyuP+hR+nStTOPPHgf6elNG6QSe/fl8OasOWzevB2r1cKQQX054/RxGI2Hrl5hUTEfzPmM1WvW43K5SU5KYOKpYxk8sE+D1E0cHXRdJ6/Uj8cXCNo5ufuDxxLKk4kv60CTOIVHTxlQ6xZeMTYJ2kIIIYQQQoh/TtjpY/fuPdx2y40NFrTL7Q7ue+QZ0lJTuHbmDAqLinn3/bl4PB4uOPfgc8CLiku4497HadIkhYvPPxObzcqevdl4fb4GqZs4ehSU+3F4Alt8abrOrFUq69Jnkl74DXsST6Z3tJl7xydjMtQM2tGyGJoQQgghhBDiHxZ2AunUqQPZ2TkNVoHvf/oVp9PF9TNnEBUV2K9b0zRef/tDJowbS0J8bJ3Xvjd7HokJcdxy/RWoamAXs84dZTX0402xw0+pMxC0dV3n9YXF/LS5AINiY0/iyTSNM/J/pyZhM9fcyS7KqpIsQVsIIYQQQgjxDwt7n+17776Dt95+l18W/oqvAXqQ/1izni4d2wWDNkD/Pj3QdZ0/126o8zqH08nSFasYPWJIMGiL40+5S6Ow3A8EgvbsRfuYvf4P9iS9iMO8hfgIlbsmJBMbYahxbYRFJSVGgrYQQgghhBDinxd2Epk89Sy8Xh8XzrgcVVWxWiyhJygKa1Ytq/f99mXnMmxI/5CyyMgI4mJj2JedW+d1O3bswefzYTAauPvBp9i8ZRtRkZEMHdyPqZPGYTTWDF/i2OLyVu2lXVZWykP33o1PMWLvALrqoSxqKbePGUBqbM1mbDMrpMZIGxBCCCGEEEIcGWGH7YsuPL/WLZUOl93hICLCVqM8MjKCcrujzuuKS0oBeOX19xl+wgAmn3YSW7ft5KO5X6IoCtOmjK/1Oq/XGzKnW9cr5gFrGpqm/Z2XIhqQz6+TW+wFDRx2Ow/dG1h53Kj76LyzJetbGHh4yGRaJ5tBD/17s5gUkqMM6LqOrutHovrBtiRtSjQkaVeisUjbEo1B2pVoLNK2RGPQNK3BR0yHHbb/c/WVDVqBw1UZojp3asu50yYB0KlDW5wuN1989QOnTzgRs9lc47p5n3/Hx/O+Cj62WMw8fOfVFOVn/zMVF/VmBewOJ/ff91ywrNSaRV78eJ7sp3JCcy/4axn94IcC1z9Xz4MpyNt7pKsgjkPSrkRjkbYlGoO0K9FYpG2JhpbaNLNB73fYE1p1XWf79h0Ul5QQFxtLVlaLw+rxjoyIwOGsmYzsdgdRkRF1XxcZ6A3v1KFtSHnnjm359LNvyMnNp3lGzRXTJ4wbzcknDq/2OjTKi/OJT2pSr63GROMKbPHlw+nRsdvtPHDf//Argf9YKY7oRnb8CGYMjaV3uyjsB1xrNCikxRox1rIi+T9N0zQK8vaSmJIuawqIBiPtSjQWaVuiMUi7Eo1F2pZoDI0xUuKw0uU7s97nmWdfoLCwCF3XURSFxMQEZl55OeecfWZY92raJLXG3GyHw0lxSSlNm6TWeV160yYHva/X66213GQyYTKZgo81TaMcUFVVfliPAvllPpxeBRSFrdt34k/syKrYP7BbzMSV92N81yhO7R5T4zqDCk3iTJiMRz5oVyftSjQGaVeisUjbEo1B2pVoLNK2xNEu7LD9/gcfctfd93PquJMZd8qJJCUmkV+Qz+dffM1d99yPyWTkjKmT632/bl07Mu/zb7HbHURW9GQvXbEKRVHo0rlDndclJyWQ0awpa9dtZOyoE4Llf677C7PZRHp6WrgvTRxhJc6qLb40XeebPWn8GdmCnLhVoLjoELuPi4bU3NpNVSEt1njUBW0hhBBCCCHEv1fYYfv1N95h+nlnc8ftt4SUjxo5goSEeF557c2wwvao4YP55vsFPPb0K5x26hgKi0p4b/anjBo+KGSP7fsefob8gkKefOTOYNkZp4/jsadf4a1ZH9OjWye2bd/J51/9yKknj6q5Sro4qjk8GgUVW3wBvL+0lF83OzHSlpTiKUTE5PHUxGEY1NBArSiBoG0xyf9qCiGEEEIIIY4eYYft3Xv2MGL4sFqPDR92Au+9/2FY94uKjOD2G2fy5rsf8djTr2C1Whk+dCBnTB4Xcp6mafj9oePoe/XowszLpvPJ/G/4/qdfiYuNYfLEk5lwyuiw6iCOLK8/ME8bHfJyc5n7zSK+LO4TPJ5ubM2jpw4iwnxAoFYgNcaIVYK2EEIIIYQQ4igTdthOSUlm1arVDB40oMax1av/ICUlOexKpDdN47YbZx70nDtuuabW8gH9ejGgX6+wn1McHTRNJ6fER+V6BK+9+hr5dge2hAyc5jTMRoXbTk0iObpmU02KMhBhkaAthBBCCCGEOPqEHbanTp7EM8+9gMfj4aQTx5CUlERBQQFffvUNr7z6BjOvurwx6imOU/nlfry+wGrjO3PLyLfns6i9h+SSn1D0M7luTCJtUmtOCYiPNBBjM/zT1RVCCCGEEEKIegk7bF95xaWUlJbyyqtv8MJLr1bdyGDgvPPO5sorLm3QCorjV4nTT7kr0KXtdPt4+bG72Jzhw2mBPUk5/KelyoDWNbd/i7apxEdK0BZCCCGEEEIcvcIO24qicNstN3LFZZew+o81lJSUEhcXS7euXYiPj2uEKorjkctbtSCapmnc9383AdBpjxG/qtIpayrn9U2vcV2EWSUpSoK2EEIIIYQQ4uh2WPtsA8THxzF82NCGrIv4l/BrOnmlfgiMHufp1+YEj6m6QoekG7hvTAqKErryuNmokBJjqFEuhBBCCCGEEEebeoXtr7/5joED+hETE8PX33x3yPNPHCurgYu65ZX68PkDSXvlDieLS1oRG62TUracPW2u5fFTEjEfsGe20aCQFmtEVSVoCyGEEEIIIY5+9QrbV868lk8+eo9u3bpy5cz/Z+++o6MouwAO/3aTTe899IQECAFCL6F3BFQEpaPoh0jvoIiCItJBOlIUpCvSkV5FadJ77ymk97673x+BhSUJZENCCNznHM7Jzrwzc2f2zZK7bxv83LIKhYIbV8/nSnDizRMRryYxJT3RDopKZfzum9x13YFTzDtE2PszrpUzjlb61VKhAFcbI4yNJNEWQgghhBBCFAzZSrb/3r8LZ2cn3c9C5ERCsoao+PRx2gkpGn7cGso9sy2kmAQRYreBseWH41Mo48zjrjbGmMpa2kIIIYQQQogCJFvJduHChTL9WYjsSlVrCYlNA0Cr1TJrdwTx907haNuUULsttHT8gJblrTMc5yhraQshhBBCCCEKIIOzmICAQK5cvaZ7nZycwtx5Cxg89Ev+XLchV4MTbwatVsvD6DQ06at8sfZELLdP7qRQ1H4sUqxoaNyLb5v4ZZj4zMZcia2FzDwuhBBCCCGEKHgMTrZHjhrDxo2bda8nTZnGrDnzuHnrNqO+/Y7lK1fnaoCi4AuLU5OSlj5O++z9JFYfDscp7hQAnpFb+aqVC6pnJkQzN1HgKEt8CSGEEEIIIQoog5Pty5evUK1qFQDS0tJYv2EzI4YPYfOGPxjQvy8rV/2e60GKgis2SU1sYnqTdkS8mgk7bgGz0CjSk+927dtnSKpVRgpcbIxliS8hhBBCCCFEgWVwsh0XH4+1dfrY2jNnzxEXF8e7rVoCUK1KZe7ff5C7EYoCKyVNS1hs+oRoao2WKdseEqSczy23VE56pI/frlmumN4xSiW42RpjJEt8CSGEEEIIIQowg5NtNzdXTp85C8DOXXvw8iqJi4szANExMZibmeVuhKJA0mi0PIxJQ5vegM2aYzFoTszELVqJUgMeIUYMGTFS/yAFuNgYZ+hSLoQQQgghhBAFTbZmI39a+4/aMn3GLLbt2MmlS1f45usvdftOnzlLyZKeuRqgKJjC49WkPhqnfepuImv/i0Hl0hnH2DM0DD/LD6OnYvnMLONOVkZYmMjM40IIIYQQQoiCz+Bku/cXn+Pm6srZc+fp2rkTH7Zro9sXHR1Dh/btcjM+UQDFJWl047TDYtP4aWcEWiBF5UC4Q11GdOiUIdG2sVBiYy4TogkhhBBCCCHeDAYl28nJySxfuZq6dWrzQZv3Muz/8YcxuRaYKJhS07SEPlpPW63RMnlHMCHqB5hRGID/NXDB08VE7xhzEwVOVgZ/7yOEEEIIIYQQry2D+uyampoy/afZREVG5VE4oiDTavXHaS/7J5wz4QsJtl9BtMVx6pWyoHk5S71jjB/NPC6EEEIIIYQQbxKDB8iW9SnD9Rs38yIWUcBFxD9ZT/vE7UQubp9OskkkKMBelUzvRvb6y3kpwNXGSGYeF0IIIYQQQrxxDE62v/3mK35duoxt23eSmJiYFzGJAighRUN0wpP1tBdsPImpOhn/qyr87hgzoVHLDJOfOVsbYaqSCdGEEEIIIYQQbx6D++92/fgzUlNSGTBoGADm5mYoeLq1UsG508dyLUDx+lNrtITGpI/T1mi1zNgWiEvQBgCUWgWdmo6iTFE7vWOszZVYm8mEaEIIIYQQQog3k8HJ9v8+667fFVi89cLi1KjTG7XZfDqW6DOrsHq0z6JMK1pUctIrb2KswMlKEm0hhBBCCCHEm8vgZHvQgL55EYcooOKSNMQnpWfaN0NS+PnEKWK9H1D5tjGWKVYM7NJA78sZpRJcbY3lCxshhBBCCCHEG+2lBswGBgVx8tRpEhISciseUYCkqbWExaV3H09K1TBlZwghVjuJM9dyoagJH302ECvTZ8dpG6MykkRbCCGEEEII8WbLUbK9es0f1KrTkLr1m9Kh08fcun0HgF59BrBk6fLcjE+8xsJi1WgedR//5e8ogiK1uEZ9iGlKUTqX6UcVb/3u4zYWSixNZUI0IYQQQgghxJvP4Mzn1yXL+H7seD5o8x6/LVmE9vGiykCN6tXYtmNnrgYoXk+xSWoSUtIz7aM3Evj35GUATNJcaGjWnc/8i+iVN1UpcLSUcdpCCCGEEEKIt4PBY7aXLV9Fv7696Ne3F2q1Wm+fp2cJbt+6k1uxideUWqMlPC79vY9KUPPHil/wSLhDqFUVYl3qMaSZo97a2QoFuNjIOG0hhBBCCCHE28Pglu3ghw+pXLlipvuMjVXEy/jtN154XHr3ca1Wy/RV+wiyuoFGocU57iQ969vjaqv/HY6M0xZCCCGEEEK8bQxOtgsXLsTZc+cz3Xfm7Dk8PIq/dFDi9ZWQoiHu0ezj+68kcD5+K5eKqjlcKhUzv440KGOhV97aXImVmYzTFkIIIYQQQrxdDM6COrb/kLnzFvD72nXExcUBkJaaxr79B1m0eAmdOrbP9SDF60Gr1RIem959PCQmjYX7I1CpQaEFG3UZBnxQWa+ruMpIxmkLIYQQQggh3k4Gj9n+vMenBAYGMeqb7/jm2+8B+KhjVwC6du5Ity6dcjdC8dqITNCQqtai0WqZuTsC+9A9OMQb4xqtpEOnzthaPJVYK8DFxgilUrqPCyGEEEIIId4+BifbAGNGf82n3bvx7+EjRERGYWdri79/TTxKSBfyN1VqmpaohPRW7a1n4rj4IB7f+AsA2CQqqVPGXq+8vYURpirpPi6EEEIIIYR4OxmcbMfFxWNlZUmxYkUpVqxoXsQkXkOhsWmghcCoVH47HI4WBVfdPqVk5GY+69ZBr6ypSoG9dB8XQgghhBBCvMUMbnqsXqse/QYMYeeuPaSkpORFTOI1E5ukJik1vfv49N2B3LZfSJTVUVKMrej2xRBKlyqpK/t4mS8hhBBCCCGEeJsZnGx/NWIID0NC6Nt/MNVq1mP4l6M49M9hNBpNXsQn8plGoyXi0ZraW06EcjJqN2qjGKIt/qOhr4JyRcz0yjtaGckyX0IIIYQQQoi3nsFNkB9368LH3boQEBjI1q3b2frXdtZv2ISjowMt32nOu61bUqVypbyIVeSDiAQ1ag3s//sox7eupaqRGReKFsHMuhK96uoPI7AwUWJjLt3HhRBCCCGEECLHM1gVLlSIL3r+jy2b/mT3zq106dyRnbv20LHzJ7kZn8hHKWlaYhI0XLxwnj1b1wJgok6iUGRpvmtQH3OTJ9VHqQQna0m0hRBCCCGEEAJyOBv508LCw/nnn3/5559/CQkJxdraOjfiEq+BsNg0EuLjWbVsqW7bfYcW1KxeBb+i+t3HnayMMZbu40IIIYQQQggB5DDZjomJYfvO3Wzdup1jx/9DpVLRqGF95s/9jAb16+Z2jCIfxCapiYlP5sfvRxNupUGVBlEWrTBxKUv32nZ6ZS1MlViZyTJfQgghhBBCCPGYwcl2j559+OffI6DVUru2P5Mn/kjTJg2xtLTMi/hEPng8KdqD+/dIs7TltEcIycYKnGI0TG5sj4XpM93HraT7uBBCCCGEEEI8zeBkOyEhgTHfjuSdFs2ws7PLg5BEfns8KdpDbSEu2jdDy0YghYbFS1K5uLleWek+LoQQQgghhBAZGZxsr1qxNA/CEK+LlDQtMYka4pM1LDgYiRGFcYv8ApVlKP3blNArK93HhRBCCCGEECJzORqzrVarOXP2HMHBD0lOTs6wv+0H7790YCJ/7PvnJJ6lyrL8cAyR8elrpyswpn9tX71lvaT7uBBCCCGEEEJkzeBk+8LFS/TuO5CgoGC0Wm2G/QqFQpLtAupBcCS/Lf0NgNtO7cAsfR3tikVNqV/aQq+so5WRdB8XQgghhBBCiCwYnGx/O3os1tZWTJ30K15eJVGZqPIiLvGKaTRadu77F4A4Uw033f7BIbYJVgoXejdyQKF4klhbmCixNpNWbSGEEEIIIYTIisHJ9vUbN5gz6ydq1KiWF/GIfBKZoGbPrp1o0XKuuIYk03uEGm3mf6UG4Wb7pJooFOBoLYm2EEIIIYQQQjyPwbNbeZQoQVxcXF7EIvJJapqWyPg0ABQocImqgCrNibKKFrSpbKtX1tHKCJV0HxdCCCGEEEKI5zI42f7m6y+Z//Mibt68lRfxiHwQHqfm7KlTutdJqiYUDv+UUY0q6Y3LNjdR6E2SJoQQQgghhBAicwZ3Ix8z9kfCQsNo0aoNLi7O2NjY6O1XKGDblg25FqDIWwnJGhJSNPz5++onGxVKmpezpLSb6ZNNCnCyztHk9UIIIYQQQgjx1jE4eyrvWzY98xIFnlarJTxOTVKqmgi3atgHHyfQrgnWZkq61tLvPu4g3ceFEEIIIYQQItsMTranTB6fF3GIfBCdqCFVrWXtiRhOuFxH6WSLQ6wbg/1t9bqLm6kU2Er3cSGEEEIIIYTItpfqF5yUlERMTCw2NtaYmZnlVkziFUhTa4mMVxMYlcqqcxdJsXsIWgXFnJQ08bV8UlABztJ9XAghhBBCCCEMkqMsau++A8yaM49Ll66g1WpRKBSULVuGQQP60bBBvdyOUeSByAQ1arWGBfvCMEkujltEV1JUAYx4ryzKp4YJ2FsYoTKW7uNCCCGEEEIIYQiDk+1du/fSp98gKlWswKiRI3ByciQ0NIxtO3bSs1c/5s2ZQdMmjfIiVpFLklM1RMamMObrL0kwccfKujrgQVuf0pR0MdGVUxkrsLMweMJ6IYQQQgghhHjrGZxsz5o9j3dbt+SnaZP0tn/avRuDh37JzFlzJdl+zT2MSmbM118CYJEShFZhjK25ki7PTIrmbG2EQibDE0IIIYQQQgiDGdxsefPWbdq2eS/TfR+0eZebt26/dFAi78Qna/h5/gK0aAm0V5OmUBFvVpSP/W2xMn1SHWzMlZippFVbCCGEEEIIIXLC4GzKzs6WW7fvZLrv1q072NnZZrpP5D+tVsvRk5e4fesm9x01nPJMY195SzycFTQq+2RSNCMlOFjK7ONCCCGEEEIIkVMGdyNv1bIFU6fPwMzMlHdaNMPGxoaY2Fi2b9/J9J9m0aHDh3kRp8gFMYkali/7DYAEUy1owSTNlZ6NnPUmRXO0MkaplO7jQgghhBBCCJFTBifbw4cOJiAgkK+/+Y5R336PsbExaWlpaLVamjdrwrAhg/IgTPGy1BotkQlqPMpU5PTFm9jG++AeUZjaJZwpW8hUV87CRImVmXQfF0IIIYQQQoiXYXCybWpqwvy5M7ly9Rr/nThJTHQMtra2VK1amTKlS+VFjCIXpC/1peW8cT2C7IqSaOKGNcb0quv2pJACHK2l+7gQQgghhBBCvKwcrbMNUKZ0KUmuC4jUNC0xiRr+vpbA1eAUMC0MQNvK1jhbP6kC9hZGqIyk+7gQQgghhBBCvCyD+wtv/Ws7Cxf9mum+RYuXsG37zpcOSuSuiAQ1yaka5hw5Q4z5CbSocbQyom0Va10ZlZGsqS2EEEIIIYQQucXg7OrnBYsxMTHJdJ+pmSk/L1j80kGJ3JOUqiE+ScOvG/7mjvEOImz2EmG9l+61bTF9amkvR1lTWwghhBBCCCFyjcHJ9u07dylVyivTfd5eJbl9587LxiRyUUScmsvXbnPn5GasE+xRaI0oZ1GDuqUsdGUszZRYmEirthBCCCGEEELkFoMzLFNTE8LCwjPdFxISipFRjoeBi1yWkKwhKVXLisVzMNYoqHUjgCIhvelfx1vXiq1QyJraQgghhBBCCJHbDE62q1erxs8LF5OQkKC3PSEhgYWLfqVmjWq5Fpx4OeHxav45dlr3+qGNP/6eTvgWfrLUl51MiiaEEEIIIYQQuS4H62wPpF37LjRo3IJ3WjTDxcWFkJAQtu/YRWpqKrNmTDU4iIDAYJau/JPr129jZmZK3drV6dCuNcbG2Q9v2879LF+9nkp+vowY3MvgGN40sUlqUtO0bF+3Qrct3KYa39W21b02lknRhBBCCCGEECJPGJxslyzpycb1a5gxcy47du4mKioKOzs76tT2Z0D/3pQoXtyg88XFJzBu8mzcXF0Y3L8HEZFRrFi9gZSUFD7t1j5b54iKimH9pu3Y2Fi/uPBbQKvVEhmv4eL9BC4WScM0FRxjS9G8vDVF7FW6co5WMimaEEIIIYQQQuSFHA2wLlG8ODOmT86VAPbs/4fExCSG9u+BlZUlABqNhl+X/cH7rZvjYG/7gjPAqj82UrlSecLCInIlpoIuJlFDapqG+b9N4LaXGhRQOKYY42vY6MqYqRRYmkqrthBCCCGEEELkhXzPts6eu0T5sqV1iTZAzWqV0Gq1nL9w+YXHX7l2kxOnztPpo/fyMswCQ6PREpWg5ujNRAIs21Aowg3beBUf+9XFzuLJRGiOVjIpmhBCCCGEEELklXyfOjww6CEN6tbU22ZpaYGdrQ2BQQ+fe6xGo2Hp8rW0ebcZ9nYvbgF/G0QnakhJ07LiSDQaIxdMUrvhamrEB5WetGpbmSn11tgWQgghhBBCCJG78j3Zjk9IwMLCPMN2S0sL4uITMjniiV17D5GckkLL5g2zfb3U1FRS09J0r7VaDZCeuGs0mmyf53Wk1miJTkjlwJV4HkQ+ukeFks7VbTE1BrQaUICduVGBv9fX3ePnK89Z5CapVyKvSN0SeUHqlcgrUrdEXtBoNCiVudsgme/Jdk5Fx8Ty54a/6P15N4NmLd+0dTfrNm3XvTY1NWHSmAFEhgXlRZivnJFay/rdV0DhDUBhW3i3VCzG6jhdmcjQ/Iru7RMeEpDfIYg3kNQrkVekbom8IPVK5BWpWyK3uRYybLLvF8n3ZNvSwoKExKQM2+PjE7CytMjyuLXr/6Jo0cKUKVWS+Ect4BqNBrVaQ3x8AmZmphgZZRyX/H7rprRs8aQlXKvVEBcVhr2Tu0FJ++tGrdHyICKVH6fP5ILjLdBWxyaxFh1rupFsYkEyYKSEwvYqlEqZgTyvaTQawkMCcHQpnOvfkIm3l9QrkVekbom8IPVK5BWpWyIv5EVPiVzJLk+cPMXNm7eoVrUKnp4eBh1byN01w9jshIREoqJjKOTumuVxgUEPuXL1Bj36fplhX4++X/LlkN5UrFA2wz6VSoVK9WT5K41GQxygVCoL9C9rZEIad+4HclVzlygrLWiPUcqiEnVKWcKj5b3srIwwNpaJ0V6lgl6vxOtJ6pXIK1K3RF6QeiXyitQt8bozONkeOHg4JiYmTJn0IwArV/3O6O9+AMDExITFC+dR27/m806hx69CWTZt3UV8fAKWj1qyj/53GoVCQflyPlke93HndiQkJOptW7ZqHSYmKjp++B7FihYy9NYKLLVGS1R8Ggtn/4SbUknpACMC7SryRcOSKB8l2iojBTZm8mEkhBBCCCGEEK+CwdnXiZOnaFC/ru71/AWL6NC+HedOH+OdFs2YNXueQedr0rAOZmamTJu1iHMXLnPg0FFW/b6RJg1r662xPW7SbAaN+F73ukTxIpT18db7Z2FhjpWVJWV9vPWWEnvTRSdqOHb8JADGGgWWSU2pbd2SKiXMdGUcrIxQKKT7uBBCCCGEEEK8CgYn2xERkTg7OwNw7foNgoKC6f5JNywtLWn7wftcvXbdoPNZWVrwzYj+GBkpmTZrEavXbqZhPX+6dWqrV+7xeGyhT6PREp2QxpZ1a3TbIqz8+NjfVpdcm6oUWJpKq7YQQgghhBBCvCoGdyO3t7MjMDAQqMLff/+Di4szpby9ANBo1DkaWF64kBujRvR/bpnRIwe+8DzZKfOmiUnScOnKbaIsS2MXf5UAu0ZUKmaGb+GnWrUtZZy2EEIIIYQQQrxKBifb9evXYdKU6Vy+fJV16zfSps27un3Xrt2gaJHCuRqgyJpWqyU6Qc2hhxYc9bpBkXANaoU3o2o96X5vYaLE3ERatYUQQgghhBDiVTI42R751XDUag0HD/1DgwZ1GTSgn27fzt17qFevTq4GKLIWk6QhIk7NHzf+Jc0ygbvOJrxnaYa3q4mujL2lJNpCCCGEEEII8aoZnGzbWFszeeK4TPetXbPipQMS2ReToGHj6VjM4yrhoFaiVWj5pLGbbr+lmRJTlSTbQgghhBBCCPGq5co62+LVi0vSEBabxl9n41CiwiaxKtU9zCjp8qhVWwEOFjJWWwghhBBCCCHyQ7aS7c5du+stG7Vy+ZI8C0hkT1SCmg0no7GMOk2yZXlQGNGxxpOx2jZmSlTGstSXEEIIIYQQQuSHbCXbNWpUz+s4hAESUjSExqRxZtcCCiWHUSjqAOa1h+u1attJq7YQQgghhBBC5JtsJdsD+/fJ6ziEAaIT1Cw7cIl/SwfgFqXEO8iMz2o56PbbmCsxNpJWbSGEEEIIIYTILwbPnrVy1e/ExsblRSwiG5JTNQRHpfHvpcWkqCDQXoNVpf/pWrUVCrCXVm0hhBBCCCGEyFcGJ9s/TphMzdoNGDp8JMeO/ZcXMYnniErU8Oc/d/EIMaLCHWOKhHvyST0P3X5bCyOMlNKqLYQQQgghhBD5yeBk++i/+/lyxBCu37hJ526f0rBxC+bOX0hw8MO8iE88JVWtJSQ6jSsHV6JEQbFwI6oW76rXqm1rLkt9CSGEEEIIIUR+Mzgzs7Gx4eOundm84Q+2bl5Hw4b1Wbp0OfUaNuOzHr3ZvmMXqampeRHrWy86Qc32c3EkGjsBkKq0oH31JzOQ20mrthBCCCGEEEK8Fl6qGdSnTGlGfzOSrZvXUaVyJQ7+fYh+A4bgX6cRP82cQ1JSUm7F+dZTa7SExqax8UwE9x1bc921G6oqvSjtbgqAUimt2kIIIYQQQgjxusjWbOSZ0Wq1HPz7H/5Yu459+w9iY2PN5z0+pXnTJhw4+DfLlq/k2rXrzJ87MzfjfWvFJGrYfiGayxaLMFUVwj6uPh/VcNTttzU3Qimt2kIIIYQQQgjxWjC4KfTu3XtMnT6TOvWa0KNnHxISEvhp2iQOH9rHVyOGUqmSH4MH9WfKpB85+Pc/eRHzW0er1RIWl8ZvZ0+QZhxNvNlVPJxN8CsqrdpCCCGEECJntFotZ9eMY8dXDdkyoArRD67md0ivpfvHNrP9y/ovLLdlQBWCzu3P9nnDrp9gy4AqpCbE5ji2w7N6cmHd1BwfD9m/P2E4g1u2GzVtiZurKx+2a0P7j9pRuHChTMt5enpS0a/8SwcoIDZJw/5L8RS+eZXYwu+RahRFl4ZFUCjSW7LtLKRVWwghhBBCGCbk8mHuH9uCf/+FWDgVxsTSLl/i2DKgClV7TMW9QsN8ub6hrm5bQPD5A9T/crXe9qbjdqIyt8mfoAqQlPgoTv32DTGB10mNj8bE2gG38vUp07ovKnMrXbmw6ye4uGE6cUG3MLN3pVTz/1G0xnu6/Q/+28blLXNQJydQtMZ7+LYdotuXEB7I0Xl9qTtsud45XzWDk+2FP8+hYYN6KJXPb0n19CjBqhVLcxqXeEpEnJoNRwOwS7pN4QhvrArXoYanOZDeqm1jJq3aQgghhBAinSYtFaWx6oXlEsIeYGbjhIOn3yuI6s1nZuOU3yEUDArlo+S6DyZW9sSH3uf82omkJkRT+ZPxACSEB3B8wUCK125H5Y9/JOzacc6uHoepjRMuPv4kx0Vyds04Knb5DkvHwhxbMBCnUlVxLVcPgPNrJ+LzXv98TbQhB8l240YN8iAMkZWEZA3/XE/A/OZaAIpE7qJJmzoopVVbCCGEEEKQ3pXY2r0kCqURD05sw8bdC/8BC4kJvMGlTTOJuHkaI1NznEvXxLftEEyt7Dm9YgwPjm8F0luWzR3cafLdVvZ81xrP+p3xbNhZd/6DkzrhVr4BpVt+gVar5dr2hdw7tomUmAhUlrYUqtiYch+OAECdmsKVv+YSeHInqYmxWLuXxOe9ATh5V8009j3ftQbgxOJhALo4AO4cWsvNfctJjHqIhWNhvJv9j6LVW2X5HE6vGENqYiz2xctx6+BqNGkplGzQBa9mn3FlyxzuHd2EkcqM0q16U6xmegtp2PUTHJn9BS0mHkBlYQ1A9IOr/D25M43HbMHCUb8X7/1jm7m2Y6HuuQFU7DKGojXe02uhTwgPZO/371L5k/HcPriG6AdXsHQuSrkPv8TJu0qG2NOSE9n9TXP8Oo+mUKUmuu1B5/Zzetk3NBu3C2Mzy0zvW6tRc37tJB789xcKI2NK1PmQ0i1763rBGvqevOjZX9z4E3EP71Dji/S5uW7tX8XFDdOo0Ws2LmX9Adg79n28mnSnuP8HGc5tYmFDibof6V5bOLhTou5H3Ny7/Mn1/1mHhWNhfD9Ib622dvMg4tYZbu1fhYuPPwnhAajMrChcuRkATt5ViX14B9dy9Qg4uQOF0hh3v0ZZ3t+rkqMJ0mJiYti+Yxe3b98hOTlFb59CoWD0tyNzJTgBUQlp/PFfBJZpkbptdUtZAGAkY7WFEEIIIQRw//hWStT5kDqDfgUgNSGWI3N6UaxWG3zbDkGTksylzbM4ueQr/PsvoFy7YVg6FeHu4Q3UHbYMhcIoW9cJOrOXWwdWUrn7BKzdPEmODScm4Jpu/4U/JxEbfJvK3SdgZuNE8Ln9HJvfn/pf/Y6VS7EM56s7dDm7RjWhYpcxOPv46+IIOruPC+unUq7tUJxK1+DhhUOcXfU95nYuOHhlTFYfC792AnM7V2oPWETErbOcXT2WiNvncPSqTJ0hvxF4ehfnfv8R59I1MLd3NeQRA1CoUjNigm4SevkINfvOA0BllnXr6aVNM/FtOxRrN09u7V/BfwsH0/i7zRm67BubmlOoSjPuH9usl2zfP7oF94qNs0y0If29L1bzfeoOXUbUvUuc+/1HzO3dKO7fFjD8PXnes3cqVQ3HklW4d2QjWo0ahdKI8JsnMbG0I/zGCVzK+pMYFUJC2IPnJvNPS4oOJfjsfhy9Kuu2Rd45h1Op6nrlnMvU4uL69PHpls7FUKckEX3/CuYO7kTdu0TRmu+RkhDDlb/m499/QbaundcMTrZv37nLR+27kJKSQkJiIg4O9kRHR5OWpsbW1gZrK2tJtnNJSpqW47eSOJ0wE9OSqZQONKJ+kwEYGz1p1X78jZUQQgghhHh7WToXpez7A3Wvr+1cjG2R0vi820+3rWLnMewZ05K4kLtYuRTH2NQShVJpUPfnxMhgTG2ccC5dHaWRCgsHd+yLlwMgISKI+8e20OT7vzCzdQagZOOPCbl8hPvHNuvF8piptT0AxubWenHc3LeCojXepUTd9gBYNSpO5J3z3Ny3/LnJtsrShnLthqNQKrFyLcGNvb+hTk3Cu9lnAHg3/ZQbu5cScesMhas0z/Z9P2ZkYoaxiQUKpVG2nptH3fYUqtgYgPLtRxJy+Qj3jmzCq8knGcoWq9WGf3/6jKToUMxsnUmOjSDk0j/U7Dv/udcwt3PFt+1QFAoFVq4liA26wa39qyju3zZH78nznn16sl2JtKQEoh9cxbaoD+E3TlOycTeCzx0AIPz6CcxsXbB0LvrcuE8u/Zrg8wfQpCbjWq4efp2+1e1LjgnH1NpBr7yptQNpSfGoU5IwsbChYtfvOL1iNOrUZIpUa4WLjz9nVo3Fo24HEsIDOb5wCFpNGqVa9NT7AuNVMjjZ/nH8JPz8KjB39nR8K1Tl18U/41OmNFv/2s7U6TOZM3t6XsT5VopOVLP0+GWirBIB8HxoxLu1igPprdo20qothBBCCCEAu6I+eq9jAq4Rdv0E24bVyVA2PuwBVi7Fc3SdQpWacOvgKvZ+/z4uPrVwKVsb13L1UBoZExt0A61Gzb4f9LsOa9JSMLG0Neg6cQ9vZ+iC7ODpx+2Da557nLWbJ4qn5pYytXbExr2k7rVCaYSJpS3JcREGxZNT9h4VdD8rjYyxK+ZD3MPbmZctXg5rN0/uH9+Kd9NPefDfNswd3PVafDM9rkR5vQY4+xIVuLlvBVqNOkfvyYuevcrCGpvC3oRfP4HSSIXS2Jji/m25tm0BackJhN889cKYAXzbDqHUO58TH3KPy1vmcHHDdCq0z36jrbtfI72u4mHXTxITeJ1yHw5n39g2VP5kPGY2jhya9jGOXpUzJO+vgsHJ9tlzF5g4fiwmJiYApKakYmRkxPvvtSYyMoqxP4xn7e8rcz3Qt41ao+VSYDIpZ/+lmokxIbYaqlTqjqlx+oeHtGoLIYQQQojHjEzM9V6nJSfiWq4eZd8bkKGs6XNaZBUKJVq0ets06jTdz+b2bjQatZ7Qq8cJu3qU82sncnPvcvwHLiQtORGF0oh6w1egUOp3Szcy1Y8vryiM9NMbhUKRYRsKBWi0j35M/9v66XvWPnW/r1qxWm24fegPvJt+yv1jmyla472X+ps/r94TJ6+qhN04idLYBEevKphY2mLl5kHEzTOE3zhFyYZdX3gOMxsnzGycsHb1QGVhw+GZPSjVvAdmts6Y2jiSHKv/hUhybATGZpYYmZhlOJc6NYXzaydSqdtYEkIfoNWodWPjLV2KE3nnAm7l6+X4fnPK4KbRlJQUrK2sUCqV2NnZ8jAkRLevVCkvLl2W9flyQ0yihk2nYgm0b0aycWuKh5fjo8bpXXSkVVsIIYQQQjyPbdEyxAbdxNzBHUvnonr/jJ+TZJlY2ZEcE6Z7nZoYR0J4gF4ZIxMz3MrXo9yHI/Dvv5DIO+eICbyBbZHSaDVqkuMiM1zzeV2uFUbGaDUavW1Wrh5E3D6rty3i1lms3DwMeQwvZGKV3o09OfrJPUcHPD+fURqr0GrU2Tp/5J3zup816jSi7l/GyjXreyhcrSWJkcHcOria2ODbFK3e+sXXuHshwzUtnYuhUBrl6D3JzrN39KpMxK0zhF47juOjbv2OXlUIOLWD+JC7OGYyCdxzadO/7NCkpQLprfNh147rFQm7ehT7EhUyHApwfddiXHxqYVfUB61Grff+aNVpaLXZe79ym8EZm4dHcQICAwEo6+PDilVriIuLJykpidVr1uLi4pzrQb5ttFotd8NSOHAlHrWROTEWpSlT90PsLNK/jZJWbSGEEEII8TwedduTmhDDqd++JuruReJD7xNy+TBnVn733ETRybsaD/77i/Cbp4kJvM6ZlWP0WkTvH9vMvSMbiQm8QXzYAx6c2IZSZYqFgztWLsUpXPUdTi8fTdDZfSSEBxB59wLXd/3Kw4uHsrymhUMhwq4dJykmjJSEGABKNu7G/WNbuHNoLXEh97i5bwXB5/ZTslG33HtIgKVzEczsXbm6fQFxIfd4ePEQN/eteO4x5g7uJIQHEv3gKslxkahTU7Ise+fQHwSd3Ufsw9ucXzuJ1IRYitZ8L8vyJhY2uFdoyOWNM3EuUzNbk7glRgZzcf104h7eIeDkDm7//TseDToB5Og9yc6zd/CqTFpSAiEX/9El1o7eVQg4sQNTG6fnDlN4ePEf7h3dTEzgDRLCA3l48RDnfh+Pvaefbvb3EnXakRAewKVNM4l9eJs7h/4g8PQevVnyH4sNukXgqd2Ubtk7/Z5dS4BCwb0jG3l48RBxD+9gV8z3hc8xLxjcjbx1q3e4dPkKH7R5jyGD+vPJZz2pVLUWCoUCrVbLlEnj8yLOt0p8spbNZ2JJfepz8P1K6UsRKJVgLetqCyGEEEKI5zCzdab2oF+5vHkWR+f1RZ2WgoWDO84+/qDI+m9Jr6afkhARyPEFAzE2t6JMy94khAfq9hubW3Nj91IubpiOVqPBppAX1XvO0M2uXbHLGK7v/IWLG34iKToEE0s77EuUx7Vc3SyvWbbNYC5tnM69wxsws3OhyXdbca/QkHJth3Fz33IurJ+KhWNh/DqPwcm7KppnWsFfhtJIRZVPxnPujwkcnNQRu2JlKdOqDyeXfJnlMe5+jQk+u58js78gNTFWt/RXZnze68+NPUuJeXANC+eiVP98OqaPWtOzUqzW+wSc3KFbnuxFilZrhTo1iUPTPkahNMKjQSfdTORg+HvyvGf/mImFDTaFvEiODcf6UUu9Y8nKaLUaXUt3VoxUptw7vIGLG6ahSUvF3M4Vd7+GeDX5VFfGwrEw1b+YycX107l9YDVmdi74dfoGFx9/vXNptVrO/j4O3w8G63psGJmYUanLd5xfOwlNWirlPhyBuZ1Ltp5lblOo05K1Ly6WtcCgIA4ePERycgq1atWgdCnv3IrtldBoNIQG38fRpTDGxjlaCS3X3QpJpvnS9UQpb2EbX4t6hUvzzXvpPQbsLY2wt8ze0gwi/zyuV85uRVEq5csRkTukXom8InVL5AWpVyKvFIS69Xid7XojVmFbpLRBx94//hcXN0yj2Q87URqr8ihC8SyNRpPr9emls8tC7u506tg+N2IRQHKqhi3noglTHSHNOAnTNHfaVE7/FkmhkLHaQgghhBBCvInSUhJJjg7jxp6lFK/dThLtN0C2ku3jx08YdNLq1bO3gLnIKCpRzV9nEqh6R80tVyMq3DuFb+EuAFibKzFSylhtIYQQQggh3jQ39yzj+q5fcPSqjHfTT198gHjtZSvZ7tS1u25MNqA3OZdWq80wWdeNq+cRhtNotOy7lEBwZDK+cUqc4pSUqtY0/fkqwM5cuo8LIYQQQgjxurNwLMS7s04adEzpll9QuuUXeRSRyA/ZSra3bvpT93NYeARfff0ttWpW553mzXByciQsLJxtO3Zx9NhxJo7/Ic+CfdPFJmvYeiYWi+Qg3bbOHzQDwMpUibGRtGoLIYQQQgghREGQrWTbx6eM7uc+/Qbxbqt3+OrLYXplGjduyISJU1m56nfq1vF/9hQiG87fT+Lcg2QcU0PTNyiUqIzTx2g/XvZLCCGEEEIIIcTrz+DZtg798y+1a2eeTNep48+/hw+/dFBvo6RUDStO3CfIfhVaxX60aClUuAgAFiZKTIylVVsIIYQQQgghCgqDZyO3sLDg8JGjmbZe//vvYSwsLHIlsFft358+QZPFgvS+HwzOsKbbY/vHf/Tc82b32NQ0LZWjovBTJqIwAktLBamPFoO3tcj4nUhuXVeOzbtjzy8eiJGxMZD5FyWvY8xy7Ot/rNQrOTYvjj04sT3qtLQs69brGLMc+/ofK/VKjs27Y7VZ1q3XN2Y59vU/VkvDr//MdH9OGZxsd+vSiZ9mziEsLJxmTRvj6OhAeHgEu3btYcOmLQwa0DdXA3xbxCRpUGhNUWpSMdIogDSqVK2OqUqBuYks9yWEEEIIIYQQBYlCnZasNfSgZctX8vOCX3gYEqKbpdzFxZlePXvwycdd8iLOPKPRaAgNvo+jS2GMjV962fEciYhLo8O8BwTHqAHwdUigV11TihUvgYuNMVZmkmwXNI/rlbNbUZRKef9E7pB6JfKK1C2RF6ReibwidUvkBY1Gk+v1KUfZ5cfdutC1SyeCgoMJDQnD2cUJdzc3qew5tO9Kgi7RBmhVoyjFiltgbKTA0lTGagshhBBCCCFEQZPjplylUknhQoUoXKhQbsbz1klK1bDpVKzutaOVETVLmgNgY67MsIa5EEIIIYQQQojXnzRF57OLAcnsD99FnNkFrBMu8k55K4yUChQKsJHu40IIIYQQQghRIOXPIGUBgEajZfGRu0RbHgWFlhKhxtQvWRcAa3MlSqW0agshhBBCCCFEQSRNp/noYUwah2/EY5lUFqtEBa7RSjauWQKArblRPkcnhBBCCCGEECKnpGU7H209G4cm2Y4icbUo+fA6ChS80/o9LEyVqIykVVsIIYQQQgghCiqDW7ZHfPUN9+8/yHRfQEAgI7765qWDehukpmn560z6xGjeD5ejJD25LlqsGLbm0uFACCGEEEIIIQoyg7O6des3EhERkem+iMhI1m/Y9NJBvQ2O307gZmgqaJ8sc+7lXQoTYwXmJpJsCyGEEEIIIURBlrNu5FksR3Xnzl3s7exeIpy3x9Ljd4k1u4JblLluW7dP/4ethSTaQgghhBBCCFHQZSvZXrFyDStX/w6AQqFg0JARmJmZ6ZVJTk4mICCAd1o0z/0o3zAhMWnsCjhMhO1xkkytKBltjiY1EVMTY6xMJdkWQgghhBBCiIIuW8m2q6sL5XzLAnDt2nU8PTxwcLDXK2OiUlGypCftP2qX+1G+YTadjiVNAwqtCpPUarTrXhtXSw3W5kYosug1IIQQQgghhBCi4MhWst20SSOaNmmkez2gX2+KFi2SZ0G9ydLUGracjsUhriF28f6UdDGhorcbKMDGTFq1hRBCCCGEEOJNYPCY7SmTfsyLON4a/91O4nZYKgBKrSmty6X3ELA0VWIsy30JIYQQQgghxBshRxOkxcTEsH3HLm7fvkNycorePoVCwehvR+ZKcG+iP/+LAaBo+Fbi7KtSp1RhAFnuSwghhBBCCCHeIAYn27fv3OWj9l1ISUkhITERBwd7oqOjSUtTY2trg7WVtSTbWYiIV7P95jUsUlKwTbxB0UIuaFN9MTG3xEwlybYQQgghhBBCvCkMzvB+HD8JP78KHD/6N1qtll8X/8yl86eYPnUilpaWzJk9PS/ifCOs/S+cAOu13Cj0B2HWGmJuHsbcwgIbadUWQgghhBBCiDeKwVne2XMX6NypPSYmJgCkpqRiZGTE+++15n+ffsLYH8bnepBvit/P3UKLBq0CbBIUlPDwRKkEa5kYTQghhBBCCCHeKAZneSkpKVhbWaFUKrGzs+VhSIhuX6lSXly6fDVXA3xTXHiQRGiIM773mlH9ugoTtYJP/vc51mZKWe5LCCGEEEIIId4wBifbHh7FCQgMBKCsjw8rVq0hLi6epKQkVq9Zi4uLc64H+SZYfzJ9YjTXmLM4xaU/dhMTE2zMjfIzLCGEEEIIIYQQecDgZLt1q3e4dPkKAEMG9efcuQtUqlqLCpVqsGPnbgb275vrQRZ0KWkadl+MB8A8NRSAsuXKY26iQCXLfQkhhBBCCCHEG8fg2ch7fNZd93OlSn7s2LaRgwcPkZycQq1aNShdyjs343sj7LkYR0RiIkZaI+JN3LFMCaJBoybSqi2EEEIIIYQQb6gcrbP9tELu7nTq2D43YnljLfrvIvedl2ORVBpH6/b0q5tEsWJFsDSVidGEEEIIIYQQ4k2Uo2RbrVZz5uw5goMfkpycnGF/2w/ef+nA3hRhsWkcjziF1jyNNKNoGpexxcu7GNbSqi2EEEIIIYQQbyyDk+0LFy/Ru+9AgoKC0Wq1GfYrFApJtp+y4VQsdjGNMEkpglJjQWMfSwBsZLkvIYQQQgghhHhjGZxsfzt6LNbWVkyd9CteXiVRmajyIq43glarZcvpWJSYYJVUjnKFTHC1NcbCVImxTIwmhBBCCCGEEG8sg5Pt6zduMGfWT9SoUS0v4nmjXAhI5nZYKgpNKr6Bcyld9F0S4q1ws7XN79CEEEIIIYQQQuQhw9fZLlGCuLi4vIjljbPuRPra2sXCtwJw9cgWFGiwkInRhBBCCCGEEOKNZnDW983XXzL/50XcvHkrL+J5Y6SmaVlxfQ+hNptJNb6t2+7uYpd/QQkhhBBCCCGEeCWy1Y28Ras2KBRPxhiHhoTSolUbXFycsbGx0SurUMC2LRtyN8oC6MCVOMKNz5BmHEWojREO8Urate+ItUyMJoQQQgghhBBvvGwl2+V9y6Zn0Y/55lU4b44tZ2Owi6+NVrGTwhHpz66Ofw1UMjGaEEIIIYQQQrzxspVsT5k8Pq/jeKPEJ2s4fD2ZQjGmFItIX0+7eAkPrM2lVVsIIYQQQggh3gYGz0Y+4qtv6N+3F0WLFsmwLyAgkJmz5zF54jiDzhkQGMzSlX9y/fptzMxMqVu7Oh3atcbYOOvwIqOi2bZzP+cvXOFhaBgW5uaUKV2Sjh++h7OTg6G3lat2X4wjOU2LVuVIuFVFHOPO0KlrFyxMpFVbCCGEEEIIId4GBje1rlu/kYiIiEz3RURGsn7DJoPOFxefwLjJs0lLUzO4fw86fPgu+w4cZvnq9c897vad+/x38iw1q1di6IDP6drpA+4/COSbsVOJiYk1KIbctuVM+vVTVA5YOJVg2MhRFHZz1Bv3LoQQQgghhBDizWVwyzagP377KXfu3MXezs6gU+3Z/w+JiUkM7d8DKytLADQaDb8u+4P3WzfHwT7zNalLe3sybcI3GBkZ6baV8vKg/9Ax/P3vcVq/09igOHJLWFwa+4KOYWRig1lKcfyr+2Fvb4ONmdGLDxZCCCGEEEII8UbIVrK9YuUaVq7+HQCFQsGgISMwMzPTK5OcnExAQADvtGhuUABnz12ifNnSukQboGa1Svzy2++cv3CZ+nVrZnqcpaVFhm2ODvZYW1sRGRVtUAy5afOZSMKt9qFVpuAU3Yp6pZpiqlJgYiyt2kIIIYQQQgjxtshWsu3q6kI537IAXLt2HU8PDxwc7PXKmKhUlCzpSfuP2hkUQGDQQxo8k1BbWlpgZ2tDYNBDg84VFBxCTEwshQu5GXRcbtp47iGmqW4kqwKpZF8WFxtjrGS5LyGEEEIIIYR4q2Qr2W7apBFNmzTSvR7Qr3emE6TlRHxCAhYW5hm2W1paEBefkO3zaLValq74E3s7W/xrVsmyXGpqKqlpaU8dpwHSu65rNBoDIs/oQUQqtwPNqf1Qi0JjRkmfU0BRLFVGL31uUbA8fr/lfRe5SeqVyCtSt0RekHol8orULZEXNBoNSmXuNpIaPGZ7yqQfczWA3PLnxm1cvHyVr4b2wczUNMtym7buZt2m7brXpqYmTBozgMiwoJeP4UT6L7x5aggAgZcOY5lWk/CQlz61KKDCQwLyOwTxBpJ6JfKK1C2RF6ReibwidUvkNtdCxXP1fDmbIC0XWVpYkJCYlGF7fHwCVpmMy87M3gP/sn7TDnp+1plyZUs/t+z7rZvSskVD3WutVkNcVBj2Tu7PXWrsRbRaLftvPcAu/qJu2yef/g8rh8KYm0g38reNRqMhPCQAR5fCuf4NmXh7Sb0SeUXqlsgLUq9EXpG6JfJCXvSUyPdku5C7a4ax2QkJiURFx1DI3fWFx/938iy/LvuDjz5oRcN6tV5YXqVSoVKpdK81Gg1xgFKpfKlf1itByVyNuk+J5Pu6bWXK+mBplu+PWOSjl61XQmRG6pXIK1K3RF6QeiXyitQt8brL99rpV6EsFy5dJf6p8dlH/zuNQqGgfDmf5x576fJ1Zs9fSqP6/rR9v0Veh/pca04GEOSwjFMlTxNvoqVUmbJYmsoM5EIIIYQQQgjxNspWsr30txWEhYcDEBAYSGpqaq4F0KRhHczMTJk2axHnLlzmwKGjrPp9I00a1tZbY3vcpNkMGvG97nVAYDDTZi3Czc2Zuv7VuH7jtu7fw5DQXIsvO7RaLVuvXwUFaBVgkQJVq1WTtbWFEEIIIYQQ4i2VrWR73PhJBAQEAlC/YXMuXbqcawFYWVrwzYj+GBkpmTZrEavXbqZhPX+6dWqrV06j0aBWP+lHf+PmHRISE7n/IIgxP/7E6HHTdf/Wb9qRa/Flx5WgFJLDS+ER1I3Kt1QoUFC5sh8qWVtbCCGEEM/xy5LfqF2vMV6ly/NF7/4cPXYcT29fzp2/oCszY9ZcTp46neHYrLa/DL/KNZkxa+5LnaNug6aM+X5cLkX09jl67Dhz5y/M1XP+umQZnt6+2So7YeJU+vYfrLft1OkztO/UDZ9ylalWsx7fjf2RxMREvTIzZs3F09s3w7+Vq37XKzd7znyq1axH7XqN+XPdhgzXH/7lKMb+MCFbsY74ahQtWr6frbJ5adfuvXh6+/LgQfqEbQ8eBOBboarutXh7ZWtAsb2dHffu3cevQnm0Wi0ocjeJLFzIjVEj+j+3zOiRA/Ve169bk/rPrM+dX3ZfjANAqyxEmllLvvmyoaytLYQQQojnun3nLuMnTOGLnv+jcaMG2Nvb4+zkxLo/VuFV0lNXbtbseVhaWFClciW947PaLgq2o8f+Y/EvS+jbu+crv/bDhyEsX7ma31cv020LCAik68f/o3q1KsybM4OQkBAmTZlOSEgo8+bM0DvezMyMlct+1dtWtNiT5YIPHfqXxb/+xvhx33H33n1GjhpD5UoV8fT0AODs2XMcOHiIvbu25t1NvgJFihTmnRZNmTFrDlMnZ++LA/Fmylay3aBBPYaNGMnkqT+hUCjo1WcAJiqTTMsqFAoO7Hu1Lcv5SavVsvNCvO516XKVMDczxspUkm0hhBBCZO3WrdtotVo6tv+QYsWK6rZXquSXj1G93pKTU1CpjHN1UqykpCTMzMxy7Xw59fje8tOqNX9QokQxypd70go+7+dF2NrasGD+HExN0//+t7GxoW//wVy8eBlf3ydzLCmViufW33/+PcL777WiVcv0uZY2bNjMv0eO4unpgVar5bux4xk6uD82NjZ5dIevTvsP29Gt+/8Y+eVwHB0d8jsckU+y9Uk1ftx3jP1+NO+0aIZWq6Va1So0adIw03+NGzfI45BfLzdCUrgbnqJ77e9lgYVKiZFSupALIYQQInPDR3zN51/0BaBB4xZ4evvy57oNGbqRP+76O2HSVF233MdlMtsO6Q0BixYvoVHTlpQpW5H6DZvzy5LfMsSwe88+mjRvTRnfSrRp14Gz585nK/Y9e/fx3gftKedXFb/KNXnvg/bsP/B3hnLLV67mg/afULFKLb7o3Z/w8AjdvoSEBMZ8P47GzVpRtnwV6jZoyqhvvycmNlbvHI+7pC9Y+At16jehbPnKREVFA/Dnug280/oDyvhWoladhkydPhO1Wv3c2GfMmks5v6qcPXuOdh91poxvJZavWA3AjRs36dmrHxUq1cC3QlU+69Gbu3fv6R3v6e3L/AWLmDhpKlWr16F8xWoMH/E1cXHxeuUCAgLp02+Q7lyffPo5V65ee+G9ff/DeGbNnkdCQqLufe3UpbvumOzEGBsbx9DhIylfsRpVq9dh4qSpL3wuj23YuJl3mjfT23bp0mWqVa2iS7QB6tWtDcDeffuzdd7HUlJS9L7YMDc3IyUl/e/oP9dtRK3R0P6jdgadE+DAwUO0aPk+ZXwr8V6bjzh9+qze/vUbNvFRx65UqlqLilVq0alLd86ePadX5nHduHL1Gh917ErZ8lVo0fJ9/j70j1651NRUxo6bQKWqtahQqQZfjvyWhIQEnlW1amXs7OzYvOUvg+9HvDmy9fWZSqWiw6OKv2Pnbnp/0QMfnzJ5GlhBsfFMCAGOC7BMKoFTckOqlDCTLuRCCCHEK5acquF+RFp+h0FRB2NMVS/+O6Bf3154eZVk0pTpzJ87ExdnZ4oVK8r1Gzf0yq37YxXt2nfmk25deO/dVgB4eZXMcjvA2B8m8PvadfTt3RM/vwqcOnWayVOmY2ZqRpfOHYD0BKpPv0HUr1eHUSNHcP/BA/oPHKpLfLJy9+49+vYfzLutWzJ86CC0Wg2XL18lOjpGr9yevfspeecuQwf1QYMxP06Ykp5IzpgKQGJiEmq1hqGDB+Do4EBQcDBz5y2kV+/+rFqxVO9cO3bupkTx4oz+ZiRGSiUWFuYs/nUpkyZP57PuH/P1V8O5efMW035KT7a/HD7kufeQmprKoCEj+PTTjxk2dCB2j4ZLftihK6VKeTFl0o8olUrmzl9A10/+x56df+klmsuWr8K3rA9TJo/n/oMHTJn6E8kpKbp7i4uLp1PX7igVSsaNHY2pqSlz5y+kY+dP2LZ1PYXc3bO8N5+yZUhOTmHzlr903bGtrCwBsh3jlyO/4dA//zJ82GCKFinCipVr2Lxl23OfCcCdu3d58CCAKlX0hyUkpyRjaqLfo9XYWIVCoeDGzVt625OSkqlavQ7RMTF4lCjOZ59+TMcOH+n2V6hQjhkz59KtSyfu3X/ApctX+Pabr4iNjWPq9BnMmzPD4F4LIaGhjP7uBwb274OtrQ0/L/iFTz7ryb4923BydATSx1C3bfMexYoVIzU1lS1bt9Gh8yds27oBT48SunOlpaYxeOiXdP+4C/379mLBwl/o028Qhw7swd7eDoAp02awctUaBg7oRzlfH7Zs2cbkKT9liEupVFKpoh///HuET7t3M+iexJvD4L4qf+/flRdxFFhrr54mzTiaePPzNFJ4YWHqJUt+CSGEEK/Y/Yg0Pph9P7/DYEP/oni5Zj7U7mnFixfD49Ef+b5lfShSpDAA1/VzbV2X3EKF3PW652a1/e7deyxbsYpxY0fTqWN7AOrUrkViUhKz5syjU8ePUCqV/LzwFwq5u7Ng/myMjNJXTzEzNeOrr799btwXL10mNTWN70Z/o0sC69Wtk7GgFhbMn01MxEOc3YoSEBjE/J8XotFoUCqVODo6MG7saF3xtLQ0ihQpTPuO3bh1+45+ApSWxpJffsbCwgJIT2ZnzppLz88/Y/jQQQDUreOPSqXixwmT6dnjM11ilJnU1DSGDhlI61bv6LYNGzESO1tbli9djKmpKQCVK1WkQePm/PHnOrp16aQra6IyyfDcRo4azcD+fShZ0pM/120gICCQnds26b4AqVG9KnXqN2HJ0uWMGjkiy3sDcHNzzbQ79qw5814Y4/XrN9i5aw8TfhxL+4/aPnp/atOoScssn8dj586l96YoU7q03vYSxYtz7vwFtFotikfzNp09dx6tVkt0dPRT5Yrx5fAhlC3rQ3JyMpu3/MXX33xHbGwcn/f4FIB3W7fkr207qN+oOQDdunaiWtUq/DhhMrX9a+Vo/oGoqGjmzJqOf630uZyqV69GnbqN+XXJMkYMS5/obUD/PrryGo2GOrVrcfbcedat36irQwApqamMGDaYhg3qAeDp4UG9hs04+Pch2rz/LlFRUaxYuYZePXvQp9fnQHr979j5E4IfPswQW5kypVnxqOeEeDvlqAk2OPghEyZOpd1HnWncrBXtPurMxElTCQ7OWMneZLdCU4gLdcfngRElgxU4pd7Hykyp+yASQgghhHiV/j18FIAWzZuRlpam+1fbvxahoWEEBQUDcObsORo3aqBLGAHeadEs03M+rUyZ0hgZGTFoyHD27t2fodv3Y9WrV9VrDfX2Kklqahrhj5aShfQuy63fa0c5v6qU8vGjfcf01r/bt+/onatG9Wp6yeip06eJj0+gZYvm+vdYuyZJSUlcu3b9hffxOJl67NA/h2ncOP15PD6fra0NZX18dEnoY5k9N61Wq+uG/9+Jk5Qq5a1LtAHs7OyoU9ufEydOPffenic7MT5Oips3a6w7zsjIiKZNG73w/CGhoSiVygxfVHTt0pHrN24yeepPhIdHcPnyFcZ8Pw4jIyO9v3nbvP8un/f4lNr+NWnUsD4zpk/mnRbNmDNvgW7ZYGNjYxYvnMehg7s5duQg34/5hps3b7Fu/Ua+GjGU0NBQPv+iL5Wq1uL9tu31ZuXPirW1tS7RBrCxtsbfv5ZeN/EbN27Sq88AqtWsh1fp8pTy8ePWrdsZ6ppSqaS2/5NzFSlSGDMzM4KC039vrl69TlJSEs2aNtE7rkXzppnG5mBvR0RkZK4umywKFoNbtq9eu07Hzh8/+uD2x8enDGFh4axa/Qd//LmeNauWUcrbKy9ife3suRhP2YB1qDTpj7FSm1JYysRoQgghhMgnkZGRaLVaqlSvnen+wKAgChcuREhIaIZJm6ytrXQtplnx9CjB4oVzmTd/Eb36DkSpVFCvbh2+GzOKwoUK6crZ2FjrHadSqYD0ScAAdu7aw9DhI+nU4SOGDhmIvZ0dIaGh9OozgOTkZL1jnZwc9V5HREQB8G6bDzON8XFilBVzc3MsLS31tkVGRrFk6XKWLF2eobzKRKX3OqvnFhISCkB0TIyu+7L+cY5cfeaLgGfv7XmyE2NIaCgqlTG2trb618kknmclJ6dgbGycodHIv1ZNvhw+hJmz57Fg4S8olUo6d2qPSqXC2dn5ueds1bIF23fs4u7de3pfPjxdV374cSI9e3yGi4sz/QYMwcjImH//3suy5avo028Q+3Zvw8Qk694iDg72GbY5OTly81EX97i4eD75tCcODvaMGjmCwoXdMTU1ZeTXo3X18TEzM9MM11KpVLpyIaHp7/GzdSCr9/HxudInv1NlWka82QxOtidMnEKxYkVZtmSR3i9ydHQ0H3/6ORMmTmHJLwtyNcjX1e7jN1BpnkyIUbd2DSxMJNkWQgghXrWiDsZs6F/0xQVfQRz5ydbWFoVCwR9rlmf6x72nR/oSSy4uznoTlkH6xFrPJrqZqV+vLvXr1SU2No6/D/3DuPGTGPHVNxmWfHqebdt3UtanDD+O+0637dix/zIt+2zyZ2eX/vfn/LkzcXd3y1C+aJEiGbbpny/jNltbWxo2qEfXLh0z7LN6JjHP6rm5uKQnnna2tty6fTvDecLDw7F7Jgk2pDdkdmJ0cXYmNTWN6Ohovb/Tw57qUZAVO1tbUlJSSE5OzvClyxc9/0e3runjrJ2dnLC1taFK9Tp0bJ/5Fx7ZtXvPPu7df8DCn+cA8O/hI0ydPAELCwu6de3E5Kk/cfvOXUqX8s7yHBERkRm2hYWF4+LiBMDp02cICg5m8cK5enNOxcbFkbH2PJ/Loy8XwsMjcHNz1bteZmJiYjFRqXRDLsTbx+D/EU6cPMVP0yZn+MbM1taWvr2/YOjwr3ItuNfZ/fBUUq/v0D3ABh+Pwcrc6LnHCCGEECJvmKqU2RorXRCpVMaZJsGZbfd/1AU2KjKKxo0bZnlOvwrl2bvvAKO+HqHrEr19h2Hz8lhbW9GqZQvOnD3Hlq0vnoDracnJyRm+DNi0JXtrK1eu5Ie5uTnBwQ9p3qzJiw/Ihtr+Nbl2/Tq+ZX30uohnJrPnplAoqFC+HABVq1Rm+45d3Lp1W7d+dHR0NP8ePqI3WVhWVCoVKSkZux1nJ8bHMezctVc3ZlutVrN7974XXtfTswQA9+8/0GuFfszCwoIypUsB8Mfa9Wi1Wlq1bP7cc27Zug0bGxuKFy+WYV9ycgo/jp/M6G9H6rUmJyUlAZCQmAikz67/PLGxsRw+clTXlTwmNpbDh4/QtWv6OPukR78jT9e3k6dO8+BBAN5ehvXGLV3aGzMzM3bt3qO35NmOnbszLf8gIIAST80/IN4+BifbRkbGWc5UmZKSglL5diScW86Hcdc5EPdIJWYpttT0ssRaupALIYQQIpeVLOnJ7r37qFa1CuYW5nh6eGBlZZnpdk+PEnTr2okhw0fSs8en+PlVIC0tjdu373D02HEWzJ8NQK+ePWjTrgNf9O5P1y6duHf/Pot/WfrCbuSrVv/B6dNnqFevDi7Oztx/8ICNm7ZSt46/QfdUu3Ytxnw3jtlz5lOpUkUOHPybw4ePZetYGxsbBg/sx8TJ0wgKDqZmjeoYKZXcu/+APXv3MW/ODMzNzQ2KZ9DAvrRp25FPPutJpw4f4eToSGhYGMeO/0e1qlV0M74DpKSm6J7b/QcPmDxlOu+0aKZLUD9s9wG/Ll3G/z7vzZDBAx7NRr4AIyPjbM1K7VXSM33itKXLqVy5ItZWVnh6emQrRm9vL5o1bcK4HyeSnJJMkcKFWbFyTbbGDPtVKI+xsTEXLl7SS7bv33/Aug2bqOhXAYAjR46x5LflTJ44Tq/x7b02H9H2g/cp6elBUnIymzZvZeeuPXw76qtMe1ks/mUJnp4eNGpYX7etVs0a/LxgMdbWVqzfsBl3Nze9yfIyY2dny1dfj2bQgL7Y2Fjz84Jf0KLVPeuKFStgaWnBmO/H0atnD4IfhjBj1hzcXF2fe97Mr2VH507t+XnhYkzNzHSzkd+7l/nkjOfPX6Ra1coGX0e8OQxOtmv712Tq9Jn4+JTRq/y379xl+ozZ1KldKzfje22tuXiaG+5qbruoaZH6Dg6W2VvqQwghhBDCEN+P+YYfxk3k0x69SEpKYtWKJdSsUT3L7WO+/RpPjxKsWrOW2XPmY2FpgaeHB++882QCNF9fH+bMms7kqT/Rq88ASpXyZtZPU/nks57PjaVMmVLs3XeAHydMJioyCidnJ95t/Q5DBg0w6J46d2zP/fsPWLZ8FYsWL6Fu3drMmD6Zth91evHBQI//dcfV1YVflvzGsuWrMDY2pnixojRqWD9HY2NLFC/OhnVrmP7TLEZ/9wPx8Qm4uDhTvVoVypTRn537426diYiIZMiwL0lNTaVZ0yZ8N3qUbr+VlSWrVyxl3PhJjPrmO9QaDVUqV2LNqnF6y35lpXGjBnTt0pH5CxYRHh5B9WpVWb1yabZjnDzxB8Z8/yOTJk/D1NSUth+8T80a1Zgwaepzr2thYUH9enU4cDB95u3HjI2NOXbsP5YsXU5qaio+ZUozf+5MGjdqoHd88eLF+HXpMkJDw1AoFJQu5c30qZNo837rDNcKCgpm8a9LWf/nGr3tY779mq9GjaZv/8EULVqEObOmP3e8NqR37R4xfAgTJ03l3r37eHt7sfTXhTg7pXcjd3ZyYs6s6UyYOJWevfvj4VGcH3/4jgULf3nuebMyYthg1GlqFi76FY1GQ/OmjRkxfDBDhun37g0LD+fCxUsMHzYoR9cRbwaFOi35+X0znhEQGEinLt0JCgqmlLcXTk6OhIdHcPXadQq5u7N61dJsfZC8LjQaDaHB93F0KYyxcfa+ewiOTsN/1g6iLI9gmmrNmGo96VDDFnvLt6NVX7zY43rl7FbU4PUihciK1CuRV6RuibzwJtYrT29fRn45TLeU1Ztm7979DBo6guNH/ja4d8CrVBDq1rIVq1iyZBn79myXlYoKiMdLE+Ymg1u2CxcqxPatG1j75wZOnDxFdHQMHh4l+OjDtnzYrk2G2R3fRH9fjccy2QfLZB9ATXVPc6ykC7kQQgghhCjAGjVqgEeJEvz+xzq6f9I1v8MpsDQaDUt/W0H/fr0l0X7L5WjKTEtLS7p/0vWt/SU8cCVB93NpN3NcbIxRGcsvkhBCCCGEKLgUCgU/jB3NlStX8zuUAu3hwxDatW2j1x1fvJ3yd32KAighRcPpK/dB4QAKBdU8zLE2k1ZtIYQQQog33a3rF/M7hDznV6E8fhXK53cYBZq7uxt9ez9//gPxdpAs0UCHLkXiFLaCIpE7UWpSqOZhhoV0IRdCCCGEEEII8RTJEg207K91HCifxGmPc9jaaCjtZoLKSLqQCyGEEEIIIYR4QpJtA2g0Wm5GnQFAq9BS08MRK3OZgVwIIYQQQgghhD4Zs22AS4HJlAhV4hirItjOh+r1ZRZyIYQQQgghhBAZ5TjZPvj3Ic6du0BQcDB9+3xB4UKFOH78BMWLF8PV1SU3Y3xt7D0XjgIFNkkKwlX1qFrCHCOldCEXQgghhBBCCKHP4GQ7PDyCXn36c/rMOdzd3QgKCqZzpw4ULlSItX+ux9zCnLHffZsXsea7w/v+0vW7L+/hiL2VdCEXQgghhBBCCJGRwX2gf/hxIuERkez4ayP792xHq9Xq9tWuXYvDR47laoCvi+DoNFKiAnSvq3maY2kiXciFEEIIkTMzZs2lnF/VDNt/nDCZkqXK8fvadbl+zU5duvO/z/vk+nmfp36j5nh6+zJpyvQM+27fuYunty+e3r4cPXbcoPP+uW4Dnt6+RERE5laob6Vdu/fi6e3LgwcBLy78lD/XbWDT5q15FNXrIT9+X0TOfdG7P526dM/vMPQYnC3uP/A3w4YMxMurJAqFfhdqdzc3goODcy2418m+yzEcKaXhbDEn7jv4U9vbAmOZhVwIIYQQuWji5Gn8umQZ48aOpsNH7fI7nFxjaWnBX9t2ZNi+ZctfWFpa5OicDRvUZ90fq7CxsX7Z8EQOrFu/ic1btuV3GEK81gxOttXqNMzNzTPdFx0Tg0qleumgXkd/XrpCvPlD7jsF4ODhRxH7N/M+hRBCCJE/pk2fycJFv/L9mG/o1LF9foeTqxo2qE9w8ENOnT6jt33L1u00bdI4R+d0dHSgUiU/jI1fbr5frVZLcnLKS51DCCEyY3Cy7edXgbV/rs9039a/tlOlcqWXDup1k5Ci4foDU2zia2CZ5EttTxcsZBZyIYQQQuSSGbPmMnf+Qr4b/TVdu3TU27d+wyY+6tiVSlVrUbFKLTp16c7Zs+cyHF/Orypnz52nTbsOlPGtRNPm77J334HnXvfmzVsMGDSM2nUbU7Z8FZq1eJfFvyxFo9Hoyjx4EICnty8bN21hzPfjqFilFjX86zN+4hTS0tKydX/29nbU9q/JlqdaQi9evMztO3do3eqdTI/5c90G3mn9AWV8K1GrTkOmTp+JWq3W2/9sN/Lk5BSmTJtBnfpNKFO2Ik2bv5uhq/PwEV/TouX77D/wNy3f/YAyvhXZu29/tu4DICgomH4DhlCtZj3K+FaiXsNm/PDjRN1+Q57pho2bGfXt9/hVrkm1GnVZ/OtSALZs3UbjZq2oULE6vfoMJCYmRi+GmJgYvh0zlhr+9SlTtiLvtfmIQ4f+fWHsqampjB03gUpVa1GhUg2+HPktCQkJGcpNmjKdFq3aUM6vKrXqNGTAoGGEhITq9nfq0p1jx/9j/4GDumEAM2bNBWDf/oN0+6QH1WrUpULF6nzQriMH/z6U7ef7tMfv1dFjx2n9Xjt8K1SlTbsOXLhwUa9ccnIy48ZPombtBpTxrUSrd9uyc9eeF57/xMlTdOj0MRUq1aB8xWq0aNWGdes3Zii3bftOGjdrRTm/qnTp9il3797T2/+i5wVPuqSv37CJBo1a4FOuMp26dOfWrdt65bRaLYsWL6FR05aUKVuR+g2b88uS3154L6dOn+HzL/pSs3YDfCtUpdW7bdmwcbNemdTUVCZMnErteo0pU7YiNfzr06NnH2JiY7O1P6uhG63ebcvwEV/rXj9+3/759wjvtP4An3KV6dj5Ex48CCAqKop+A4ZQoWJ1GjRqwda/tme4l337D/JBu474lKtM1ep1+Gb02Az19MaNm3Ts/AllfCvRoFGLTN+314HBXwUOHTyAzl0/pUOnj3mnRTMUCgW7d+9l/s+L2H/gb/5Yszwv4sxXx24mQoodDikNAPD3MkclXciFEEIIkQvmzlvArNnz+ObrL/m4W5cM+x88CKBtm/coVqwYqampbNm6jQ6dP2Hb1g14epTQlUtLTWPAwGH873+fULRIEVauWkPvvgPZvHEtZUqXyvTawQ9D8PAowfvvtsLSypJLl64wY9Zc4hMSGNhff6zq1Okzadq4EbNnTuPU6TPMnDWX4sWK0aVzh2zd57utWzJpynS+/eYrlEolW7b+RbWqVXDLZBWbxb8uZdLk6XzW/WO+/mo4N2/eYtpP6cn2l8OHZHmN/gOHcOLkKQb060PJkp4cOPg3Q4Z9ha2tLQ3q19WVexgSwtgfxtO3zxcUKuROoULu2boHgGEjRvIwJJQx347EycmRwMAgzp9/kvwZ+kxbNG/KnFnT2b1nL+MnTCEiIpJjx/7jqxFDiYuL47ux45k4eRrjx30PQEpKCt269yAsLJyhgwfi5ubCxk1b+V/PPs99rwGmTJvBylVrGDigH+V8fdiyZRuTp/yUoVx4eAR9en2Oq4sL4RER/PLrb3Ts8gm7tm/G2NiYsd99w5BhX2FmbsbXXw4HwM3NFYAHDx7QuFEDPu/RHYVCycG/D/FZj96sXP4rNWtUz/Zzfiw0LIzvf5hAr57/w9raminTZtC73yD+WLFQV2bQkBH8fehfhg4eQMmSHqzfsJk+/QaxYP4smjRulOl5Y2Pj+N/nfahapTIzf5qCiYkJN27cJCYmVq/c5ctXWLQ4ghHDBqNWq/lxwmSGDPuKdWtXZft5PXbx0iXu3bvPiOGDAZj20yw++awne3b+hampCQBjf5jA72vX0bd3T/z8KnDq1GkmT5mOmanZc3/XAgICqVK5Ep07dcDU1ISTJ0/z1dffotFoaNe2DQDzf17EqjW/M2L4EEp5eRERGck//xwmJSUlW/sNERoWxviJk+nbu2d6nflhAoOHjsDMzJzq1arQof2H/P7HnwwZ9hWVKvpRuHAhIP2LjQGDhvFhuw8YNLAvISFhTJ76EzExMcyaMRVI/3Llk097Ym5hzvSp6V90/TRjDnHxcZQoXtzgWPOSwcl25UoVWbViCZOn/MT4iVPQarXMnb+QSpX8WLHsF8r5ls2LOPPV31effJPibG2Eb2HTfIxGCCGEEFlJSk5+7n6VsTFGRhlXE0lLU5Omfn4rrZlp5v//Jyen6P5QNlRCQiLTfppFh/bt+OzTjzMtM+CpBE2j0VCndi3OnjvPuvUbGT50kG5fSmoqfft8QfuP2gJQr25tGjVpybz5C3V/pD6rtn9NavvXBNJb1KpWqUxiUhLLV6zKkBhW9KvAmNHprVd16/hz9Ohxtu/cle1ku2mTxoz69nuOHD2Of60abPlrO/36fJGhXFxcPDNnzaXn55/p7q9uHX9UKhU/TphMzx6fYW9vl+G4I0ePsWfvfn77dSF169bWHRcaGsaMWXP0ku3o6BiWLF5AxYoVshX7086eO8/woYP1WuTbfvC+7mdDnmnlShX5dtRXAPjXqsGOnbtZtnwlhw7s0d3j5StX+WPtel2yvWnzX1y+fJW/Nq/D29sLgHp163Dnzl3mzP2ZObMyTkQHEBUVxYqVa+jVswd9en2uO65j508IfvhQr+zkieN0P6vVaipXqoh/3UYcOXKMunVr4+3thZWVFRYWFlSq5Kd37NNfGGk0GmrVrM716zdYvWZtjpLtqKhoVq/8jVKP7tXCwpzOXT/l4qWruBfx4PKVq+zctYdxY8fQuVP68Iv69eoSEBDIrNnzs0y2b9+5Q2xsLMOHDdJ9QfH4fXtaTGwMWzb9iaOjAwAJCQmM+OobgoKCcXd3y9bzeiwsLJzVK3/Do0R6Qli2bBmaNGvNuvUb6dypPXfv3mPZilWMGztaN5SkTu1aJCYlMWvOPDp1/AilMvPete+2bqn7WavVUr1aVYKDH7J6zVpdsn323AXq1PanW5dOurLvtGim+/lF+w3x7PsWEhLCd2PH80XP/9G/X28A/CqUY+euPezavZdPu3dDq9UyYdJUWrVswcTxY3XncnFx4rMevenXtxelvL34c91GHoaEsHvn1gzPskAn2ykpKezbf5CyPmVYs+o3kpKSiI6OwcbGOstx3AWdVqvl/M7FlE0M5qFtHaqWr4uVmSz5JYQQQryOPv1i2HP3D+zzGTWrZxzytu/AvyxZsTbL46ytrVg4e0Km+0Z9P4Wp40cZFugjZmZmVChfjs1bttGubRuqVqmcocyNGzeZOn0mJ0+dITw8XLf99u07Gco2b/Zk/LORkRFNmzZi9+59WV4/OTmZeT8vYvPmrQQGBZGa+uQLh/j4eCwtLXWv69bx1zvWy8uTI0efzCD+dJdyhUKRYSJda2srGjaox5at2zAxUREWFsY7LZoRGBikV+7U6dPExyfQskVzvXPWrl2TpKQkrl27To0a1TLcy6F/DmNnZ0utWjWeOa4WO0bvRq1W675osbe3y1GiDeBbtiyLflmCkZERderUyvDHvSHPtE7tWrqfjYyMKFa0KAqlQu/LBI8SxYmJidEde+iffyldyhsPjxJ691mndi02Pmd28KtXr5OUlESzpk30trdo3pTj/53Q23bg4CHmzP2Za9dvEBcXp9t++84dveQxM0FBwUz7aSb/Hj5KSEiobuWicuV8n3tcVlxdXHQJG4CXV0kAQkLDAPjvxEkAWr6jnxS2atmCceMnkZCQgIVFxkn4ihcripWVFd+OGUv3j7tSs0Z1XUL9NB+fMnrbvR9dPzj4oS7Zzu7zKlXKS5ccApQoXhyfMqU5c/YcnTu159/DRwFo0byZfh32r8WChb8QFBSsawF+VnR0NDNmzmX33n08fBiiG3LxdF0q5+vDwsVLmDFrLg0b1KN8OV+95P1F+w3x7PvmUaKE7l4es7GxwdHRgaBHE2zfvn2HgIBAvh31ld7916heDaVSyfnzFyjl7cWZs+eyfJavG4OSbRMTEwYNGcHSXxdQrFhRzMzMMDMzy6vYXgu3w1KJUN3HSq3EKfYEVT2aYWIsXciFEEII8fKUSgULF8yhc5fu9OjZlzWrftPrBhwXF88nn/bEwcGeUSNHULiwO6ampoz8enSGSb1UKmNsbW31tjk5OhISqj929GmTJk9nzR9/MqBfb8qV88XGxpo9e/YxZ94CkpNT9BJDa2v9Wb9VKhXJj3oSPHgQQL2GT5KdwoULcXDfzgzXe7d1S77+ZgwAdevUxs7OLkOyHRERlV62zYeZxhyUxco3kZGRREVFU8rHL9P9ISGhuuTIydEx0zLZMXvmNKZOn8m0n2Yy+rsf8PT0YNiQgbRo3hR4+Wf6bGL4ePLhx8dGRkZx8dLlTO8zs14bjz2uB88mlE5O+s/i7Lnz9OzVjyaNG9Lri//h6OCIQqGg7UedXjiRnEajoWevfsTGxjFoYD9KFCuGuYU5P82cQ9Az73N2PTvbvMmj5/G4a3NMdAwqlTF2dnYZ7kur1RITE5tpsm1ra8vypYuYMWsuQ4d9RZpaTbWqVRgz+mu930Ebaxu943TvR0p63TfkeTk6ZKx3Tk6OhD56byIjI9FqtVSpnvkXGoFBQVkm28O/HMWpU2fo36833t4lsbKyYuWq3/lr25Mx0X16f4FCqWT9hk3Mmj0PRwcHunbtxIB+vVEoFC/cb4hn3zeViSrz7U99jkREpo8F79VnQKbnDApK/90PDQ3N8lkmJT2/d9OrZnA38pKeHhk+FN9k+y6EccozDbUReD6sSM2Sb2YLvhBCCPEmWLIg8+7Sj6mymLm6UYPa1KtbI0fX/HHM8Bwd95iNtTVLlyzkow5d6f5ZT9auWUHRokUAOH36DEHBwSxeOBcfnzK6Y2Lj4nB75jypqWlER0frJdxh4eG4ODtnee1tO3bSqeNH9Pqih27b/gN/G3wPLi7ObFz/u+61iUnm3eobNayPOk3Nn+s2MG1K5j0F7OzS458/d6YuOX5a0SJFMj3O1tYWRwcHflk8P9P9TyeZhiYOT3NxcWbyxHFoNGM5f+Eic+ctYMCgYezZuZVixYrm2jPNip2dLWXKlGLi+B8Mi/tRPQgPj9CNr4b0rs1P27V7L9bWVsyZNV3XqhkQEJita9y5e4+Lly6zYP5smjZ50n07OSnJoFgNYWtnm3ndDwtHoVA8d2k4P78KLPllAUlJSRw5epwJE6fQq/cADuzLuExdVgx5XuER4Rm2hYWF6363bW1tUSgU/LFmeaYrPHl6eGR63uTkZPbtP8iokSP45OMn3fiXa1brlTM1NWHQgL4MGtCXO3fvsvbPDcycNZdiRYvwQZv3srE/fShNamqq3nmjn5nAL6fsHr1/340ZRUW/jD1PXF3S53dwdnbm4qVLGfaHhYVjZWWVK7HkFoOT7eHDBjN23AS8vEpSoXy5vIjptbLjyD6MTUGthJKutXGxebnlJYQQQgiRd7IaV/0ixsZGGBvnbJhYTsdrP83J0ZHlSxfzUceufNy9B3+sWY6zs7NuDPrTf3ifPHWaBw8C8PbyynCenbv26sZsq9Vqdu/eh59f+Syvm5SUrHdutVrN1q0ZZwd+ERMTkwx/Fz49+/Zjpqam9O79OWfPXtBLxp5WuZIf5ubmBAc/pHmzJpmWyUxt/1osXPQrKpXqlXQnVSqV+FUoz5DBA9izdz93796jWLGiufZMs1LbvyYHDv6Nq4sLrplMLpeV0qW9MTMzY9fuPfj6+ui279i5W69cUlISxsbGel9IZNY9XaVS6Vp3H3ucVD99/wEBgZw8dVrXjTi3PR56sW37Tr0l87bt2IlvWZ9MW7WfZWZmRsMG9bh37z5jx00gOTlZl1i+SHafF8C1aze4c/eubujBnbt3uXzlqi5u/0djxqMio2jcuGG2rg+QnJKCRqPRe+5xcfHPnWW/RPHiDB86iNVr/uDGzVvZ2v/4S5obN2/p6t6NGzd1Lc4vq2RJT9zd3Lh//wEfd+2cZTm/CuXZsHFzps+yWtUquRJLbjE4c5w4eRpRUVG0/bAT9nZ2ODo56lUuhQK2bdmQq0Hml+RUDaprx2iCCTHmZtR61xkzlSz5JYQQQojcV6RIYZb+upCOnT+m+2dfsHrVb1SsWAFLSwvGfD+OXj17EPwwhBmz5uDm6prheBOVirnzFpCckkzRIkVYsXI1QcHB/Dx/VpbXrFO7Fr//8SfeXiWxt7dnxco1OZp52BC9v/j8ufttbGwYPLAfEydPIyg4mJo1qmOkVHLv/gP27N3HvDkzMp0rqG4dfxo3asCn//uCnj0+o0yZUiQmJHLtxg3u3r2vN+FSVuo2aEqRwoVZvXJppvtjYmPp/mlP2rR5F08PD1JTU1m2fCU2Njb4PpokOK+f6Qdt3mfVmrV06tqdz//XHQ+PEsTExHLx0mVSU1MZMWxwpsfZ2dnRuVN7fl64GFMzM91s5Pfu3dcrV6e2P0uWLue7sT/SrGkTTp8+w4ZNWzKcz6ukJ+s2bGLv3v04uzjj6uKCp2d6sjR56nTUajUJCQnMmDUX10zqq6e3L+0+eJ8pk8e/1PPwKVOa5s2a8OOEySQlJePpWYKNm7Zy6tQZFs6fneVx+/Yf5I+162nerDGF3N0JDQvjt+UrqVKlUrYTbcj+84L0bs6f9+zH4EH9AJg+Yzauri66Ccw8PUrQrWsnhgwfSc8en+LnV4G0tDRu377D0WPHWZDF/dhYW1OhQjl+XrgYBwd7jI2N+XnBYqytrPVa07/o3Z9yvmUpW9YHCwtz9u47QHR0DLVq1sjW/op+FXB3d2Pc+EkMHzqIuLg4fl64GPtnuvDnlEKhYNTXIxg0ZAQJCYk0alAPcwtzAgKC2H/gIMOGDsLTowQftmvD3PkL6NGzL0MG9QfSZyN3dnbKlThyk8HJdnnfspDDCQ4Kmt+3HwNAgYIUo3L4e0sXciGEEELkndKlvPll0Xy6fdKDHj37sGzJIubMms6EiVPp2bs/Hh7F+fGH71iw8JcMxxqrjJn50xRGfz+Oa1evUaRoEebN+em5rbxjRn/Nt6PH8v3Y8ZiZm9GubRuaN2vMyFFj8vI2X6jH/7rj6urCL0t+Y9nyVRgbG1O8WFEaNayfaffax+bO/omfFyxmxao1BAYEYm1tTalSXnzY7oNsXTcxIfG5f7CbmphSurQ3y5avIjAwCDMzU8qX8+W3JQtxcLAH8v6ZmpqasHLZr8ycnb42e2hoKPb29pT18cmwRvuzRgwbjDpNzcJFv6LRaGjetDEjhg9myLCvdGUaNqjHl8OHsGz5Kv5ct5EqlSuxeOE8GjdtqXeunp9/xp279xg64mtiYmIY0L8Pgwb0Zd7cGYz5fhz9BgzB3d2Nfn2+4MiRY5x/am3sx2smO+VScvTTtElMmTaDnxcuJjoqGs+SHsyd/dNzW4dLFC+GUqlg6vSZhIdHYG9vR53a/owYNsiga2f3eUH65Hotmjdl4uRphISEUtGvAuPGjtbrHTPm26/x9CjBqjVrmT1nPhaWFnh6ePDOO8+fFXzG9MmM+vZ7hn85Cjs7W7p/3JX4hAQW/7JEV6ZK5Ur8tX0Hi3/9DbU6DU8PD36aNkk3Ud+L9qtUKn6eN4tvx4yl34AhFC9ejG++/pLxEyYb9Myep+U7zbG2tmbe/IVs2pz+pUWRwoWpV6+Obn4BMzMzfluyiG9Hj2XI0C9xdXWlX99e7Nm7L8PSbflNoU5L1uZ3EPlJo9EQGnwfR5fCeuvgAYz+9Sgnjx/FJukm970Gs2NYcZmJXGTL43rl7FY0x7M4CvEsqVcir0jdKvhmzJrL4l+WcOHsiRcXfkUKWr26e/ceDZu8w4Y/V+OXyXhRkXv+PXyU//Xoxf69OzIdl/8iBa1uAXTq0h0LCwt+WTQvv0MRWdBoNLlen2QA8nNcjC9Egqk7cWbFqFjcHEvTgvHLLIQQQgghDHPy1Gnq1PaXRPsVOHnyFG3bvp+jRFuIgiRHyXZMTAzbd+zi9u07mS4BMGb01y8dWH4Lj0vjSMJWUoqFYp1Ymf95mr/UzJVCCCGEEOL11faD92n7wfv5HcZbYUD/PvkdghCvhMHJ9u07d/mofRdSUlJISEzEwcGe6Oho0tLU2NraYG1l/UYk2//dTiLR5DZpxtFYJpXB30vGawshhBDi9fR4uR4hxOspq0n3xJvN4H7RP46fhJ9fBY4f/RutVsuvi3/m0vlTTJ86EUtLS+bMnp4Xcb5yR28k4BjbHJv4GhQ386aE08sv6yGEEEIIIYQQ4u1gcLJ99twFOndqj4lJevKZmpKKkZER77/Xmv99+gljf3i56ftfFycv38M8xQOHuAb4FyuMsZF0IRdCCCGEEEIIkT0GJ9spKSlYW1mhVCqxs7PlYUiIbl+pUl5cunw1VwPMD+euB2J17TdKPlwJWi01PKULuRBCCCGEEEKI7DM42fbwKE5AYCAAZX18WLFqDXFx8SQlJbF6zVpcXJxzPchXbe3mfQCYp4aCQiHjtYUQQgghhBBCGMTgZLt1q3e4dPkKAEMG9efcuQtUqlqLCpVqsGPnbgb2L/iTc5y7epS7TmriTbR4OBvjaqvK75CEEEIIIYQQQhQgBs9G3uOz7rqfK1XyY8e2jRw8eIjk5BRq1apB6VLeuRnfKxcVHUOIrYYLxdIwSTWnfzGz/A5JCCGEEEIIIUQBk6N1tp9WyN2dTh3b50Ysr4UdB0+hUWgxTQFVWkn8vS3zOyQhhBBCCCGEEAWMwd3IAdRqNSdPneavbTtYv2FThn8F2d69B/AMMabhBSuc4prJ5GhCCCGEyBO7du9l+crVGbYPH/E1LVq+/8qvm1fnHfHVqDy9n9xw9NhxPL19OXf+Qq6cr26Dpoz5fpzu9bPv6aVLl5kxay6JiYkvPNeMWXMp51c1V+ISQrxaBrdsX7h4id59BxIUFIxWq82wX6FQ0PaD1/sD9XniosMBiDX3wsvZAlsLo3yOSAghhBBvot2793L+wkW6den0Rlw3v+7ndfTz3FnY2tpkuf/S5SvMmj2Pj7t2xtz8+Q07HT5qR8MG9XI7RCHEK2Bwsv3t6LFYW1sxddKveHmVRGXyZk0eFlviQxRBxwi1rs77xaVVWwghhBBvJ61WS0pKKqamJvkdSoHj6+uTa+dyd3fD3d0t184nhHh1DO5Gfv3GDUYMG0KNGtVwdHTAxto6w7+CKjgqjTtphbjj3I4UlT3VPGRyNCGEEELkvuEjvmbdhk1cu34DT29fPL19GT7ia70yR48dp/V77fCtUJU27Tpw/sJFvf1arZZFi5fQqGlLypStSP2GzfllyW85vu7jrs77D/xNy3c/oIxvRfbu28+f6zbg6e1LRESk3rlavdtW79hnzzviq1EG3c+zunz8GUOGfaV7ff7CRTy9fenbf7Bu261bt/H09uXsufO6badOn6FLt0/xrVCVCpVqMHDwcMLCw597rcfCwyPo1WcgvhWqUsO/PnPnL8zw/J7tEh8TE4Onty9/rtug2/ZsN/Kn/bluAyO++gaAqjXq4OntS90GTbOM6dlu5I+7vB/65zADBw+nfMVq1KnfhAULf8nWPQohXh2DW7Y9SpQgLi4uL2LJdyfvJhLo+CtKjSn2cQ2pUbJEfockhBBCiDdQv769CI+I5Nat2/w0bRIADg72uv2hYWF8/8MEevX8H9bW1kyZNoNefQZwYO8OVKr0XoVjf5jA72vX0bd3T/z8KnDq1GkmT5mOmakZXTp3yNF1H4aEMPaH8fTt8wWFCrlTqJA7J06cytH92NnbGXQ/z6perSpr/1yve33s+H+Ympry34mTetssLS0o51sWSE+0O3fpToMG9Zg1YyqJiYlM+2kWX/Tqz7q1q154H6O++Y53W7ek89wZ/PvvUaZNn4mdrW2WzzMnGjaoT78+XzBn3gKW/roAaytrTEwM7z3w7eixtGnzLvM/nMXuPXuZNGU6ZcqUon69urkWqxDi5RicbH/z9ZeMHTcBnzKlKVnSMy9iyjcHb4aQapz+zWchazMcLGW8thBCCFGQBMSlt746mVthavQkiUtITSYyOQEjhQI3Szu9Y8IT40hSp2KpMsXO1OKlz5cdxYsXw9HBnsDAQCpV8suwPyoqmtUrf6OUtxcAFhbmdO76KWfOnqNa1SrcvXuPZStWMW7saN2qMHVq1yIxKYlZc+bRqeNHKJUZOzC+6LrR0TEsWbyAihUr6LZlJ9nO7LwajYbQ4PvZup/MVK9WhZmz5hIQEEjhwoX477+TfPThB6xa/Qc3b96iZElPjv93ksqVKmJklP432+QpP1G+vC/z585EoVAAULqUNy1atWH/gb9fOPa5Vq0ajPxqGAD16tYhLDycufMWZPk8c8LR0YFixYoCUM7XV+/LDkO0aN6EQQP6AlDbvyb7D/zN9h27JdkW4jWSrWS7Ras2ug8sgNCQUFq0aoOLizM2NvqTPygUsG3LhmdPUSBcfqDANeojklQB+HsVy+9whBBCCGGgMku+BGDPh19Sw72kbvvOO+f5eMdCClnacfWzyXrHDDqwgo03TtG3YhMm1m3/0ufLDa4uLrrEFMDLK/3awcEPAfj38FEAWjRvRlpamq5cbf9aLFj4C0FBwRQuXMjg69rb2+kl2rnlRfeTmUoV/TBRqTj+3wnaFHqXEydPMWPaZE6dOsOx4ycoWdKT/06cpFOHjwBITEzk5KnTjPxyGGq1WnceD48SuLu7ce78hRcm282aNtZ7/U6LpmzYuJmg4GAKFzL8eealOnVq635WKBSULOlJcHBwPkYkhHhWtpLt8r5l07Pox3zzKpz8s3XXIZQXDuNu6UeUZV1qlJT1tYUQQgiRP2xs9OfAMXnU1To5ORmAyMhItFotVarXznAsQGBQUI6SbSdHR4OPyY4X3U9mzMzMKF+hHMf/O0HZsj7ExsZRuXJFqleryvH/TlCvbm0CA4OoVi19PHN0dAxqtZpx4ycxbvykDOcLCnpxIuro6KD32snRCYDQkLDXLtnO7JnGxMTmUzRCiMxkK9meMnl8XseR706dv4FFSjAWKcFEWfpQ3UNmIhdCCCEKmiufpidZTuZWetublyjPlU8nYfR048EjMxp0ZWLdDliqTHPlfK+Cra0tCoWCP9Ysz3TMs6eHR47Oq8jkfkxN059Lamqq3vbomJgcXcMQ1atVZcfO3ZQt60PZsmWwtLSkWrUq/DBuIsf/O4GJiQl+FcoD6cmnQqGgT6/PafpMCzWAg/2Lu2uHh0fovQ4LDwPA2SU96TYxNSXl2ecQnffPQQhRMBk8ZjsziYmJPAwJoXixYpl+SBcE925dASDMqiLO1kYUsn+zljQTQggh3gaFrTJPqCxUplhkkkwDOD6TSL/s+bJLZaJ6bsvu8/j71wQgKjKKxo0b5ul13dxcAbhx8xauri7pP9+4maGl+GXuJyvVq1Vh/s+L2LZtJzWqVXu0rSrBDx+y9s8N+FUor1uazMLCgsqV/Lhx8xZDhwzM0fV27d5L82ZNdK+379iNq6sL7m7pS2+5u7kSHPyQ+Ph4LC3Te0Ee+uewwddRZaNlXwhR8Bk808OixUuYOXue7vXx/05Sq04jmjRrTaMm73D37r1cDfBViTfRcM9RTYSlBeUKv9x/nkIIIYQQL1KypCcPAgLZvOUvzp2/wIMHAdk+1tOjBN26dmLI8JHMnbeAf/49woGDh1iydDlf9O6fq9et6FcBd3c3xo2fxL79B9m85S8GDB6GvZ1drt1PVqpUroyRkRHHjv9H9WrpE6k5OjrgVdJTb9tjX305jP0H/qb/wKHs3LWHo8eOs3HTFoaNGMnRY8dfeL0jR44xYeJUDh36lwkTp7Jh42b69O6pmxytebMmJCcn8+XIbzn0z2F+XbKMZctXGnxfXiXTx6wvX7maM2fOceXqNYPPIYR4/RmcbP++dp3uG06AHydMwtu7JAvmz8be3p6p02fmaoCvSqithnMl0rjjega/YrK+thBCCCHyVvsP2/FOi2Z8/8N42rTtwMxZcw06fsy3XzN0cH+2/LWdHp/3ZsiwL/lr2w6qV6/63OMMva5KpeLnebMwNTWh34AhzF+wmFEjv9S1cmd13llz5mVxxuyzsrKkbNkyKJVKqlatrNtevfqTVu6nValciT/WLCc+PoERX33DZz16M2vOfMzMzCle7MWT3477YQy379yhV9+BbNi0mSGD+tOtSyfdfm9vL6ZOHs/FS5f5ond/Dhw8pFvqzBC+vj4MHNCXTZu28mGHLnz+RV+DzyGEeP0p1GnJWkMO8K1QlV8XzadGjWoEBz+kdr3GrF75G9WrVWHX7r18O2Ysxw4fzKt4c93jZSk6zJvORftEjNN82dy+J9VkzLZ4CY/rlbNb0VxbKkQIqVcir0jdEnlB6pXIK1K3RF7QaDS5Xp8MHrNtZmZKXFwcAIePHMXSwoIqlSsC6RNTxMbG5WqAr0qJUGPcH5hysWgLfAtJN3IhhBBCCCGEEDlncLJdoUJ5fl64GIVSycLFS6hfvy5GRkYA3L13P0OXooIkQeVKSWdzLEzlGzIhhBBCCCGEEDlncFb59ZfDCAkJ4/Mv+pKQkMDQIQN0+/76a4eulbuguenSgTDrKpQvKuO1hRBCCCGEEEK8HINbtr29vTi4fyeRkVHY29vp7ft65HCcnZxyK7ZXKsXYkQQLRyoUkWRbCCGEEEIIIcTLyfE6288m2gBlSpd6mVjyVZD9aoyjK1G+aJH8DkUIIYQQQgghRAEng5MfSVUFY2ysoaSzSX6HIoQQQgghhBCigMtxy/abxja+FuUdS2JspMjvUIQQQgghhBBCFHDSsv2IbXxNahYruDOpCyGEEEIIIYR4fUiy/Ujl6BW0KG+V32EIIYQQQgghhHgDSLL9SNlSxfEtLDORCyGEEEIIIYR4ea/FmO2AwGCWrvyT69dvY2ZmSt3a1enQrjXGxs8PT6vVsvmv3eze9w8xsXGUKFaYbp3a4u3lYXAMnsWL5jR8IYQQQgghhBBCT763bMfFJzBu8mzS0tQM7t+DDh++y74Dh1m+ev0Lj938127+3Lidls0bMGLwF9jZ2TBh6jwehoQZHEexooVyEr4QQgghhBBCCJFBvifbe/b/Q2JiEkP798CvvA8N69Wic4f32bP/XyIio7M8LiUllU1/7aZVi4a0bN6IcmVLM6D3p1haWrB1x16D4yhdquTL3IYQQgghhBBCCKGT78n22XOXKF+2NFZWlrptNatVQqvVcv7C5SyPu3bjFomJSdSsVlm3zdjYmOpV/Thz9pLBcSiV+f4ohBBCCCGEEEK8IfI9wwwMekghd1e9bZaWFtjZ2hAY9PC5xwEZji3k7kp4RCQpKSm5H6wQQgghhBBCCJEN+T5BWnxCAhYW5hm2W1paEBefkPVx8YmojI0xMVHpbbeytECr1RIXn4iDiUmG41JTU0lNS9O91mjUAKif2ibEy9JoNACkpaVJrwmRa6ReibwidUvkBalXIq9I3RJ5QaPRYGxsjEKhQKFQ5Mo58z3ZftU2bd3Nuk3bda9trK0YO7IXURFZt6ILkVORYUH5HYJ4A0m9EnlF6pbIC1KvRF6RuiXygr2jGyamprlyrnxPti0tLEhITMqwPT4+AStLi6yPszQnNS2NlJRUvdbtuPgEFAoFVpYZW8sB3m/dlJYtGupd55vvpzJt4igszLO+nhCGSExKpO/g0cz9aSzmZpnXRSEMJfVK5BWpWyIvSL0SeUXqlsgLT+rVD2TsH50z+Z5sF3J3zTA2OyEhkajomAzjsZ89DiAo+CHFixXRbQ8Meoijgz0mmXQhB1CpVKhU+l3PY2LjUCqNpBuKyDUKhZLk5BQUCqXUK5FrpF6JvCJ1S+QFqVcir0jdEnnhSb3KnS7k8BpMkOZXoSwXLl0l/qnx2Uf/O41CoaB8OZ8sjyvl5Ym5uRlH/zuj25aWpua/k2ep6Fc2L0MWQgghhBBCCCGeK99btps0rMPOPQeZNmsRbd5tRkRkNKt+30iThrVxsLfVlRs3aTZh4RHMmDwGABMTFe+3asq6jduxsbaiaBF3du/7h7i4BFq3aJxftyOEEEIIIYQQQuR/sm1lacE3I/qzdMVaps1ahJmZGQ3r+dPhw9Z65TQaDWq1Rm/be62aogW27thLbEwcxYsV4athfXB1ccr29VXGxrR7/x1Uxvn+KMQbROqVyAtSr0Rekbol8oLUK5FXpG6JvJAX9UqhTkvW5trZhBBCCCGEEEIIkf9jtoUQQgghhBBCiDeNJNtCCCGEEEIIIUQue2sHOgQEBrN05Z9cv34bMzNT6tauTod2rTGWsR8ihw4eOsrPv6zMsP29lk3o1P79fIhIFFTBD0PZun0vN27e4X5AEIXcXZny49cZyu0/eITN23YTHh6Ju7sLHdq9S+WK5fIhYlEQZKdejZ0wk8tXb2Q4dur4URQu5PaqQhUFyNHjp/nnyH/cvnOf+PgE3Fydad60Pg3q1tRbPkc+r4QhslOv5PNK5MTpsxfZvG0PAYHBJCYm4WBvS9XKFWj3/jtYWDxZs/3k6fP8sX4rQUEhODra837rZjSoW9Pg672VmWVcfALjJs/GzdWFwf17EBEZxYrVG0hJSeHTbu3zOzxRwH01tA8W5ma61w72dvkXjCiQHgQEcfrcRbw8S6DRatFqM06tcfjoSRYtXU2b1s3wLVuKI8dOMX32IsaMHIS3l0c+RC1ed9mpVwClvT3p0qGN3jZnJ8dXEKEoiLbt3IeTkwNdO7bB2tqK8xevsmjJasIjIvmwTUtAPq+E4bJTr0A+r4Th4uIT8PIsTosm9bG2suR+QCB/btzO/QdBfD28LwBXrt1k+uzFNKxXi487t+PipWss/HUV5mam1KhWyaDrvZXJ9p79/5CYmMTQ/j2wsrIE0mc7/3XZH7zfurnekmNCGMqjRFFsrK3yOwxRgFWuWI6qlSsAMH/Rcm7duZ+hzJ8bt1GrRmXat0tfucHXpxT3HgSyfvMOvhzS+5XGKwqG7NQrAAsLc0mARLYNG/SF3v955cqWJi4unm0799P2vRYolUr5vBIGy069Avm8Eoar618NqKZ7XdbHG2NjYxYvXUNEZDQO9rZs2LwDL8/i9OjeEUj/zHoYGsbaDdsMTrbfyjHbZ89donzZ0rpEG6BmtUpotVrOX7icj5EJIQS6PyKy8jAkjKDgEGo+84HvX6MyFy5dIzU1NS/DEwXUi+qVEDmR2ZfLJYoXITExieTkFPm8EjnyonolRG6yfpQTqtPSSE1N5eLl6xmS6lo1qhAQGExoaLhB534rW7YDgx5m6HNvaWmBna0NgUEP8ykq8aYYPmo8sbFxODs50LC+P++1bCJ/5Ipc9fhzqpC7q972Qu5upKWlERIaLuPVRI5dvnqD7j2HotFoKFmyBO3btsKntFd+hyUKkKvXbuFgb4e5uRlXrt0E5PNKvLyn69Vj8nklckqj0ZCWpiYgMJj1m3ZQpVJ5nJ0deRAQhFqtzvCZVfjR64Cghzg7Z3+owluZbMcnJOgNgH/M0tKCuPiEfIhIvAns7Gz58IOWeHmWQKF4NLHCuq1ERkbJXAAiV8U/+pyytLDQ225paaG3XwhD+ZTxom7t6ri7OhMZFc3WHfv4cfIcRo8cSCnpqimy4cq1mxw+dpKuHT8A5PNK5I5n6xXI55V4Of2HjiEiMgoAv/I+9Ov1CQDxCYkAWD6TK+b0M+utTLaFyAt+5X3wK++je12hnA8mKhO27dpPm3ebY28ncwEIIV5vH33QSu91pYrlGP71eDbI2FqRDeERkcyatwRfH29aNK2f3+GIN0RW9Uo+r8TLGDGkF8nJKTwICGLD5p1MmbGAUcP75fp13sq+rZYWFiQkJmXYHh+fgJWlRSZHCJEzNatXQqPRcPfeg/wORbxBHn+7mpCYqLdd14Ikn2Mil5iZmlLJzzfLydSEeCw+PoFJ0+djZWXJ4H49dMOn5PNKvIys6lVm5PNKGKJ40cKU8vKgUX1/hg38nEuXr/PfybO6Fu1nc8Wcfma9lcl2IXfXDGOzExISiYqOydA/XwghXjePP6ee/RwLDHqIsbExri5O+RGWEOItlZKSwpQZC0hISOKrob31hurJ55XIqefVKyFyU7GihTEyMiL4YSiuLk4YGRll+pkFT8ZuZ9dbmWz7VSjLhUtX9frcH/3vNAqFgvLlfJ5zpBCGOXzsJEqlkhLFiuZ3KOIN4urihLubC0f/O623/ejxU5QrWwpjsBec2wAADShJREFUYxkhJHJHUnIyp89eoKRHsfwORbym1Go1M+cuISAwmK+G9sbB3k5vv3xeiZx4Ub3KjHxeiZy6cfMOarUaFxcnVCoVvj7eHDuh/5l15PgpChdyM2hyNHhLx2w3aViHnXsOMm3WItq824yIyGhW/b6RJg1ryxrbIscmTJ2Lr08pihYpBKRPkLbv4GFaNK2PnZ1NPkcnCpLk5BTOnLsIQFh4JImJSRx79IeqT2kvbGysadfmHeYuWIarixO+ZUpx5Pgpbty6w+iRg/IxcvE6e1G9Cgx6yJbte6lWxQ9nJwciI6P5a+c+oqJjGdi3RX6GLl5jvy77g1NnL9C14wckJiZx/cZt3b4SxYugUqnk80oY7EX16uatu/J5JXJk+uxFeJYoRrGihTExUXH3XgBbt++lWNFCVKtcAYAP3mvBDxNn8cuy36lVrTIXr1zj8NGTDOj9qcHXU6jTkrW5fRMFQUBgMEtXrOXajduYmZlRz786HT5sLd+wihz7beWfnDl3iYjIKLQaLW5uLjSqX4vmTeqjUCjyOzxRgISGhjNg+HeZ7vv2/+3deVSVdR7H8Tcq4DKorGmWyhVkuYAgqGlqLmCu5QJDZpaakea+YKOZSKKSC6MoiFg5ajUVimLa4lRW2jjNmdHUCGualAJZFCikHGVx/nDmjlf2GzNM+nmdc8/hPs/vfr/f5+Gee873/n7Pc5+ejbeXOwCHPzrG/rf+wMWCItq3c+Gh0FF09/f5H1YqvyS1va8cHNqyfVcK336XzaWSH2lua4O7myvjRg/DzdD5f1qr/HLMWhDFxYLCKvfFr11umgXS55XUR23vq/KKCn1eiUXSDhzi2J9PkJ9/kYprFTg7OdIzsBsjhg2iZYv/XKrwl3/9qlBObj5OjvY8MCKEgf171zvfbdtsi4iIiIiIiPy33JbXbIuIiIiIiIj8N6nZFhEREREREWlgarZFREREREREGpiabREREREREZEGpmZbREREREREpIGp2RYRERERERFpYGq2RURERERERBqYmm0RERERERGRBqZmW0REbjvFxcUY3I3s3rO3sUupk8hFSxg6/MHGLsPMS9t3cvjDjytt7zcghKjomEao6Lrde/ZicDdSWFgEQFZWNgZ3I2+9/a7FMURERCzRrLELEBERkZrNnDGNny5fbuwyzGzfsYtBA+9j4ID+jV1KjZydndnzxqu4unZq7FJEROQ2o2ZbRETk/1ynTh0bu4RfLFtbGwICujV2GSIichvSMnIREbnlvfZ6Cv0GhODtG8iER6dwLvPbSmNS96YR9tAjBAT1xj+wN+MnTOLkyVOm/We+/AqDu5EjR/9o9rry8nJ69x1I7PPrAMjJyWXm7Pn0uKc/nsYA+g8cwoqVsTXW99Xfvmby1Gl079EHb99ABg8ZwdbkF037b15G/u9lzunpGUx+/EmMfkEMDB5G6t60SrE/OPwRYeET8PYNNB1XenqGaX9xcTHPRj1Hrz734entzwOjwzhy5JMa6+03IITs7PPsevn3GNyNVS7J3/nyq/S9Lxi/gF48OX0WBQWFZvstyftvqXvTGPnAODyNAQT2vJfJU6eRnX2+yrHVLSOvTwyAlN178fDuxuspe+pUo4iIiGa2RUTklvb+Bx+yZOlyxo0dzaiRwzj9+RfMnD2/0risrGzGjn6Ajh07UlpaypsH3iL84cd468BeDK6d8fToin83P1J2p9Kvbx/T6z76+Ch5efmEhY4FYOGixeTlXyDq2cU4OTly/nwOp0+n11jjExEzcHJyJHbVCuzsfkVm5rfk5ObVemzzFiwiPDyUKZMf4/U3dhP59DP4+frg5tYFgAMH32bOvEhCggexIW4N1tbW/PX4CXLz8jAavbh69SoTJ03l4sUCFsybQ7t2LuxLO8DjEU+xf18Knh5dq8yblBDPlCemERTYnalTJgHQsePdpv3vvX+Yc+cyiY5aSlFRETGr1hC9YhXxG65/IWFpXoDkbS8Ru2Y9vw4bx4L5cygrK+PYsU8pKCykQ4c7az1nlsTYsfMVVseuZd2a1YwaObxOOURERNRsi4jILS0hcSs9ggJZ+/xKAPr368vVK1fYlJBkNm72rKdMf1dUVND33t6cPHWaPan7iFwwF4Dw8FCWR6/khx9+oE2bNgCk7E4lsLs/XboYADh56jSRC+YxcsQwU7yxY6q/uVlhYRHfZWWxbOlvGDx4IAC97+lVp2ObOPFhJk4YD0Bgd38Of/gx77z7B2a6deHatWusjl1Hv759SEqMN73mxmus0/YfJCPjSw7u34O7u5vp/Jw7l8nmhCQ2x8dVmddo9MLGxgYnJ8eql2hfg+SkBGxtbQDIyj7PlqRkKioqaNKkicV5iy9dYuOmRMaHh7EyZrlpe0jwoDqdL0tiJCZtI35TIps3xRE8uO55REREtIxcRERuWeXl5Xye/gVDhgw22z506JBKY7/++u9Me2o2Pe7pj5uHL129uvHNN2c5e/acacyoEcOwbtaMtDcPAtcb5Q8Of0hY6DjTGKO3N9te3M7Lr7zGuczMWmu0t29Lhw53snb9Bvak7iMnJ7fOx3fjDHvLli3pcGd704z4N9+cJSc31zTjXpUjRz/Bo6s7rq6dKSsrMz363tubU6c/r3MdN+vZM8jUaAO4u3WhtLSMgoKCn5X3xInPuHz5Mr8Oq/6YalOfGOvjNpK4ZSsvJCeq0RYRkXrTzLaIiNyyCgsLKSsrw9HB0Wy7k5P585KSH3lscgQODvY8s3gRHTq0x9bWlsVLlnHlylXTuJYtWzJq5HBSUlJ59JGH2Zf2JjY2NowYfr9pzKaN61kXt5H1v93IsuUrMBhcWTh/DkPvD6myRisrK3Zs38b6uI1ERcfw00+X8fExsnTxInr2DKrx+FrbtTZ7bm1tzdUrVwAo+v57AFxcnKt9fVHR96R/kUFXr8qz002bNq0xd411tbarVBdgOpeW5i0q+gEAFxcXi2urT4y33z2ER9eu9AjqbnE+ERG5fanZFhGRW5aDgwPNmjWjoLDAbPvFi+bPT5z4jJzcXF5ITsDLy9O0/VJJCe1uihkeHsrvX08hI+MMu1P3MnzYUFq1amXa7+LizJrYGCoqnuP05+kkJG5l9tyFvPfuAbPrmm9kcO1MwqbfUlpayvHjn7EubgNTn5zBsaMfmMWuD/u2bQHIz79Q7Zi2bdvg6dmV2FUrLMphKUvz2ttfX7qfn59P+/Y3/2caPkZy0mamz5jL9Blz2bol3vSlgYiISF1oGbmIiNyymjZtitHbi0OH3jfb/s47h8ye/+Nfs8E3NlN/PX6CrKzsSjH9fH3w9vIkesVqzpz5irDQMVXmbtKkCd38fJk/bzZlZWVkVnEH9JtZW1vTq1cPpkVMpaSkhLwaGuXaGAyutG/XrtJdwm90b597+O67LO5wccHP16fSo7Zab5z1rw9L8wYE+NOiRQtS9uyzKG99YxhcXdm14wVOnjrFnHmRlJeXW5xXRERuP5rZFhGRW9qMpyKImDaLyKefMd2NfG/am2Zj/P39aNWqJVHRMUyLmEpuXj4b4jfT7o47qowZHh5K1PIYDAZXggL/s8S4+NIlJk2OYPToURhcXSktLWXnrldo3bo1RqN3lbEyznzJqtVrGTliKB073s2lSyVsSdrGXXd1oFM1M+F1YWVlxeLfLGTOvEimz5jD2DEPYmNjzfETJ/Hz9WHwoAGMGf0gr76WwvhHJvHE45Nwde1McfEl0r/IoLS0lEUL51Ub362LgWPHPuXI0T/Spk1r7r7rLuzt29apNkvztrazY/bM6Ty/No5rFRUEBw+ioqKCP336Z0aNHF7rFwSWxPD06MqO7duYMHEKCxctYf3a1TRporkKERGpnZptERG5pQUPHkTMc1EkbNnKgYNv49/Nj00b1jEmdLxpjLOTE5vj41gdu46I6bNwde3EyhXLzX7r+kb3hwQTtTyGsHHms9q2NrZ4eLizc9ernD+fQ/Pmtvj6GNmxPRkHB/sqYzk7O+Hs7MSWpBfIzcvDzs6OHkHdiVv//M+6bhpg5IhhtGjenIQtycyeuxBbW1uMRi/uD7l+wzhbWxte2fkSGzclkLAlmQsXLmBvb4+3lxePTHioxtgLF8zl2WXPMWPmXEp+/JE1sTGEjqt6lv9mPyfvkxGP4+DgwEu/28me1H20+lUrAvy74ejoULeTYkEMH6M3v3txK49Onsozzy5nVUw0VlZWdc4nIiK3J6vysivXGrsIERGRX5I3UlJZuiyaTz5+D2fn6m9AJiIiIrcvzWyLiIjUUVZWNmfPZbI5MYkRw4eq0RYREZFqqdkWERGpo43xCew/cJDuAQEsWRzZ2OWIiIjI/zEtIxcRERERERFpYLqdpoiIiIiIiEgDU7MtIiIiIiIi0sDUbIuIiIiIiIg0MDXbIiIiIiIiIg1MzbaIiIiIiIhIA1OzLSIiIiIiItLA1GyLiIiIiIiINDA12yIiIiIiIiINTM22iIiIiIiISAP7J0BkKyWO4Sz2AAAAAElFTkSuQmCC", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "grid = np.linspace(0.05, 30, 400)\n", "f_grid = reported_fraction(grid, fit[\"k\"], fit[\"lam\"])\n", "f_lo = np.percentile([reported_fraction(grid, r.k, r.lam) for r in boot.itertuples()], 2.5, axis=0)\n", "f_hi = np.percentile([reported_fraction(grid, r.k, r.lam) for r in boot.itertuples()], 97.5, axis=0)\n", "\n", "# The KM curve, rescaled to converter-conditional terms by its own plateau, is the\n", "# nonparametric check: if the parametric fit has the shape wrong, they separate here.\n", "km_curve_t = km_t[km_t <= 30]\n", "km_conditional = (1 - km_S[:len(km_curve_t)]) / (1 - km_plateau)\n", "\n", "fig, ax = plt.subplots(figsize=(10, 5))\n", "ax.fill_between(grid, f_lo, f_hi, color=BLUE, alpha=0.16, lw=0)\n", "ax.plot(grid, f_grid, color=BLUE, lw=2.4, label=\"fitted share reported (95% band)\")\n", "ax.step(km_curve_t, km_conditional, where=\"post\", color=MUTED, lw=1.5, ls=(0, (4, 3)),\n", " label=\"Kaplan-Meier, same data, no shape assumed\")\n", "ax.plot(grid, reported_fraction(grid, TRUE_K, TRUE_LAMBDA), color=GREEN, lw=1.6,\n", " ls=(0, (1, 2)), label=\"the truth we built in\")\n", "ax.axhline(MIN_TRUST_FRACTION, color=WARM, lw=1.2, ls=(0, (5, 4)))\n", "ax.text(30, MIN_TRUST_FRACTION + 0.02, \"refuse to multiply below 30% \", color=WARM,\n", " ha=\"right\", fontsize=10)\n", "ax.set_xlabel(\"days since the click\")\n", "ax.set_ylabel(\"share of that day's conversions already reported\")\n", "ax.set_title(\"How much of a day's result you can see, by how old the day is\")\n", "ax.set_ylim(0, 1.02); ax.set_xlim(0, 30)\n", "ax.legend(frameon=False, loc=\"lower right\")\n", "fig.tight_layout()\n", "fig.savefig(f\"{FIGDIR}/fig_reported_fraction.png\", dpi=170)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "0d76fb20", "metadata": {}, "source": [ "## Readout two: how long a reporting window has to be\n", "\n", "The same fitted curve answers the other half of the question. Read it at 99% instead\n", "of at today's age and you get the day by which essentially every conversion is in:\n", "the point past which waiting longer changes nothing.\n", "\n", "That number is the honest length of a reporting window, and it is usually far longer\n", "than the seven days most reports assume." ] }, { "cell_type": "code", "execution_count": 10, "id": "d20c08e8", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:48:11.246846Z", "iopub.status.busy": "2026-10-24T17:48:11.246432Z", "iopub.status.idle": "2026-10-24T17:48:11.947056Z", "shell.execute_reply": "2026-10-24T17:48:11.946060Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "half of all conversions are in by day 2.6\n", "99% are in by day 23.7 (95% band 21.5 to 26.0; truth 24.1)\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "dropping the unconverted clicks -> window of 20.7 days (too short)\n", "keeping them, no cure share -> window of 2,468,291 days. Not a long answer but a runaway one: forced to send every click to a conversion sooner or later, the only way the curve can hold 94% of them back is to push the timescale out without limit.\n", "the fit that separates the two -> window of 23.7 days\n" ] } ], "source": [ "look = float(lookback_days(fit[\"k\"], fit[\"lam\"]))\n", "look_lo, look_hi = band(lambda k, lam, c: float(lookback_days(k, lam)))\n", "look_truth = float(lookback_days(TRUE_K, TRUE_LAMBDA))\n", "half = float(fit[\"lam\"] * (-np.log(0.5)) ** (1 / fit[\"k\"]))\n", "\n", "print(f\"half of all conversions are in by day {half:.1f}\")\n", "print(f\"99% are in by day {look:.1f} (95% band {look_lo:.1f} to {look_hi:.1f}; \"\n", " f\"truth {look_truth:.1f})\")\n", "\n", "# The two ways of getting this wrong, both of which are the default somewhere.\n", "conv_only = cohort[cohort.observed]\n", "naive_k, naive_lam = None, None\n", "res_naive = minimize(\n", " lambda p: -np.sum(np.log(np.exp(p[0]) / np.exp(p[1]))\n", " + (np.exp(p[0]) - 1) * np.log(conv_only.elapsed.values / np.exp(p[1]))\n", " - (conv_only.elapsed.values / np.exp(p[1])) ** np.exp(p[0])),\n", " [np.log(0.9), np.log(4.0)], method=\"Nelder-Mead\")\n", "naive_k, naive_lam = float(np.exp(res_naive.x[0])), float(np.exp(res_naive.x[1]))\n", "look_naive = float(lookback_days(naive_k, naive_lam))\n", "\n", "def weibull_censored_negloglik(p, elapsed, observed):\n", " k, lam = np.exp(p[0]), np.exp(p[1])\n", " t = np.maximum(elapsed, 1e-9)\n", " z = (t / lam) ** k\n", " log_f = np.log(k / lam) + (k - 1) * np.log(t / lam) - z\n", " return -(log_f[observed].sum() + (-z[~observed]).sum())\n", "\n", "res_nocure = minimize(weibull_censored_negloglik, [np.log(0.9), np.log(4.0)],\n", " args=(cohort.elapsed.values, cohort.observed.values.astype(bool)),\n", " method=\"Nelder-Mead\", options={\"maxiter\": 3000})\n", "nocure_k, nocure_lam = float(np.exp(res_nocure.x[0])), float(np.exp(res_nocure.x[1]))\n", "look_nocure = float(lookback_days(nocure_k, nocure_lam))\n", "\n", "NOCURE_RUNAWAY = bool(look_nocure > 10 * look)\n", "print(f\"\\ndropping the unconverted clicks -> window of {look_naive:.1f} days (too short)\")\n", "if NOCURE_RUNAWAY:\n", " print(f\"keeping them, no cure share -> window of {look_nocure:,.0f} days. Not a long \"\n", " \"answer but a runaway one: forced to send every click to a conversion sooner or later, \"\n", " \"the only way the curve can hold 94% of them back is to push the timescale out without \"\n", " \"limit.\")\n", "else:\n", " print(f\"keeping them, no cure share -> window of {look_nocure:.1f} days (too long)\")\n", "print(f\"the fit that separates the two -> window of {look:.1f} days\")" ] }, { "cell_type": "markdown", "id": "ba6b045e", "metadata": {}, "source": [ "### What the platform itself reads\n", "\n", "The fit above is the teaching route: it names the shape and the share who never convert.\n", "The platform reads both answers another way. It takes the Kaplan-Meier curve only as far\n", "as the account's history still holds at least 30 clicks, and 5% of the cohort, at risk\n", "(the **support edge**), and reads everything as a share of the incidence it reaches\n", "there, so the never-convert share cancels and no fitted tail is extended past the data:\n", "\n", "- the reported share at age N is $F_{KM}(N) / F_{KM}(\\text{edge})$;\n", "- the window is the first time $F_{KM}$ reaches 99% of $F_{KM}(\\text{edge})$." ] }, { "cell_type": "code", "execution_count": 11, "id": "ef61b4f0", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:48:11.959804Z", "iopub.status.busy": "2026-10-24T17:48:11.955970Z", "iopub.status.idle": "2026-10-24T17:48:11.995785Z", "shell.execute_reply": "2026-10-24T17:48:11.994438Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "support edge at day 53.0\n", "platform window: 22.6 days (the fit's 23.7; truth 24.1)\n", "platform reported share: 25.6% after one day, 79.4% after a week (the fit's 26.4% and 80.2%)\n" ] } ], "source": [ "SUPPORT_AT_RISK_FLOOR = 30 # the platform's at-risk floor at the support edge\n", "SUPPORT_AT_RISK_SHARE = 0.05 # and its share-of-cohort leg; the larger binds\n", "\n", "\n", "def support_curve(elapsed, observed):\n", " \"\"\"Kaplan-Meier incidence up to the support edge: the last time with at least the larger\n", " of the floor and the share of the cohort still at risk.\"\"\"\n", " t = np.asarray(elapsed, float); o = np.asarray(observed, bool)\n", " times, inv = np.unique(t, return_inverse=True)\n", " leaving = np.bincount(inv, minlength=times.size).astype(float)\n", " converting = np.bincount(inv, weights=o.astype(float), minlength=times.size)\n", " at_risk = np.cumsum(leaving[::-1])[::-1]\n", " floor = max(SUPPORT_AT_RISK_FLOOR, SUPPORT_AT_RISK_SHARE * t.size)\n", " edge = int(np.flatnonzero(at_risk >= floor)[-1])\n", " incidence = 1.0 - np.cumprod(1.0 - converting[:edge + 1] / at_risk[:edge + 1])\n", " return times[:edge + 1], incidence\n", "\n", "\n", "sup_t, sup_F = support_curve(cohort.elapsed.values, cohort.observed.values)\n", "sup_window = float(sup_t[np.searchsorted(sup_F, LOOKBACK_QUANTILE * sup_F[-1] * (1 - 1e-12))])\n", "sup_share = {N: float(sup_F[np.searchsorted(sup_t, N, side=\"right\") - 1] / sup_F[-1])\n", " for N in (1, 7)}\n", "print(f\"support edge at day {sup_t[-1]:.1f}\")\n", "print(f\"platform window: {sup_window:.1f} days (the fit's {look:.1f}; truth {look_truth:.1f})\")\n", "print(f\"platform reported share: {sup_share[1]:.1%} after one day, {sup_share[7]:.1%} after a week \"\n", " f\"(the fit's {float(reported_fraction(1, fit['k'], fit['lam'])):.1%} and \"\n", " f\"{float(reported_fraction(7, fit['k'], fit['lam'])):.1%})\")" ] }, { "cell_type": "code", "execution_count": 12, "id": "1207c88e", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:48:12.003710Z", "iopub.status.busy": "2026-10-24T17:48:12.003356Z", "iopub.status.idle": "2026-10-24T17:48:13.076095Z", "shell.execute_reply": "2026-10-24T17:48:13.075004Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots(figsize=(10, 4.4))\n", "labels = [\"drop the clicks that\\nhaven't converted\",\n", " \"keep them, assume\\neveryone converts eventually\",\n", " \"separate the two\\n(the fit this post uses)\"]\n", "xmax = max(look_naive, look, look_truth) * 2.0\n", "# A runaway fit gets a hatched bar running off the axis, never a length a reader can\n", "# measure: drawing it to any finite width invites exactly the reading it deserves least.\n", "values = [look_naive, xmax if NOCURE_RUNAWAY else min(look_nocure, xmax), look]\n", "bars = ax.barh(labels, values, color=[WARM, WARM, BLUE], height=0.55)\n", "if NOCURE_RUNAWAY:\n", " bars[1].set_hatch(\"///\")\n", " bars[1].set_alpha(0.55)\n", " ax.annotate(\"\", xy=(xmax * 1.03, bars[1].get_y() + bars[1].get_height() / 2),\n", " xytext=(xmax * 0.93, bars[1].get_y() + bars[1].get_height() / 2),\n", " arrowprops=dict(arrowstyle=\"-|>\", color=WARM, lw=2))\n", "ax.axvline(look_truth, color=GREEN, lw=1.8, ls=(0, (1, 2)))\n", "ax.text(look_truth + 0.5, -0.62, f\"the truth: {look_truth:.0f} days\", color=GREEN, fontsize=10)\n", "captions = [f\"{look_naive:.0f} days\",\n", " \"runs away - no finite answer\" if NOCURE_RUNAWAY else f\"{look_nocure:.0f} days\",\n", " f\"{look:.0f} days\"]\n", "for b, cap, v in zip(bars, captions, values):\n", " inside = NOCURE_RUNAWAY and b is bars[1]\n", " ax.text(v * 0.5 if inside else v + 0.5, b.get_y() + b.get_height() / 2, cap,\n", " va=\"center\", ha=\"center\" if inside else \"left\",\n", " color=PAPER if inside else INK, fontsize=11,\n", " fontweight=\"bold\" if inside else \"normal\")\n", "ax.set_xlabel(\"days until 99% of conversions have arrived\")\n", "ax.set_title(\"Two ways to get the window wrong, in opposite directions\", pad=14)\n", "ax.set_xlim(0, xmax * 1.06)\n", "ax.set_ylim(-0.8, 2.6)\n", "ax.grid(axis=\"y\", visible=False)\n", "fig.tight_layout()\n", "fig.savefig(f\"{FIGDIR}/fig_two_wrong_windows.png\", dpi=170)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "e04a3768", "metadata": {}, "source": [ "## Putting the multiplier back on the dashboard\n", "\n", "Now apply readout one to the days that are still young. Each day gets the multiplier\n", "for its own age, and the interval comes from the same bootstrap. Days too young to\n", "carry a multiplier are left blank rather than guessed at." ] }, { "cell_type": "code", "execution_count": 13, "id": "4654b3ca", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:48:13.091351Z", "iopub.status.busy": "2026-10-24T17:48:13.087621Z", "iopub.status.idle": "2026-10-24T17:48:14.593063Z", "shell.execute_reply": "2026-10-24T17:48:14.592046Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " day_index age_days reported adjusted adjusted_lo adjusted_hi eventual trustworthy\n", " 42 13.5 17 18.1 14.9 19.2 21 True\n", " 43 12.5 30 32.3 28.7 34.7 32 True\n", " 44 11.5 23 25.1 21.7 27.5 27 True\n", " 45 10.5 25 27.8 24.1 30.9 28 True\n", " 46 9.5 27 30.7 27.1 34.5 30 True\n", " 47 8.5 26 30.5 25.7 34.5 26 True\n", " 48 7.5 24 29.2 23.3 34.1 27 True\n", " 49 6.5 20 25.6 20.1 30.9 25 True\n", " 50 5.5 15 20.5 16.0 25.6 23 True\n", " 51 4.5 15 22.4 14.9 28.7 25 True\n", " 52 3.5 18 30.4 21.8 40.7 27 True\n", " 53 2.5 11 22.5 14.0 32.4 19 True\n", " 54 1.5 6 17.1 5.8 29.1 23 True\n", " 55 0.5 3 NaN NaN NaN 16 False\n", "\n", "of the last 13 days the model would report on, 12 have the eventual total inside the interval\n" ] } ], "source": [ "pred_rng = np.random.default_rng(SEED + 7)\n", "adj_rows = []\n", "for r in by_day.itertuples():\n", " age = TODAY - r.day_index - 0.5\n", " f_hat = float(reported_fraction(age, fit[\"k\"], fit[\"lam\"]))\n", " trustworthy = f_hat >= MIN_TRUST_FRACTION\n", " draws = np.array([predict_total(r.reported, [age], b.k, b.lam, pred_rng)\n", " for b in boot.itertuples()])\n", " lo, hi = np.percentile(draws, [2.5, 97.5])\n", " adj_rows.append({\n", " \"day_index\": r.day_index, \"age_days\": round(age, 2),\n", " \"reported\": int(r.reported), \"eventual\": int(r.eventual),\n", " \"reported_fraction\": round(f_hat, 4),\n", " \"adjusted\": round(r.reported / f_hat, 1) if trustworthy else np.nan,\n", " \"adjusted_lo\": round(float(lo), 1) if trustworthy else np.nan,\n", " \"adjusted_hi\": round(float(hi), 1) if trustworthy else np.nan,\n", " \"trustworthy\": bool(trustworthy),\n", " })\n", "adjusted = pd.DataFrame(adj_rows)\n", "recent = adjusted.tail(14)\n", "print(recent[[\"day_index\", \"age_days\", \"reported\", \"adjusted\", \"adjusted_lo\",\n", " \"adjusted_hi\", \"eventual\", \"trustworthy\"]].to_string(index=False))\n", "\n", "covered = recent.dropna(subset=[\"adjusted\"])\n", "inside = int(((covered.eventual >= covered.adjusted_lo) &\n", " (covered.eventual <= covered.adjusted_hi)).sum())\n", "print(f\"\\nof the last {len(covered)} days the model would report on, \"\n", " f\"{inside} have the eventual total inside the interval\")" ] }, { "cell_type": "code", "execution_count": 14, "id": "652359f8", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:48:14.607271Z", "iopub.status.busy": "2026-10-24T17:48:14.606338Z", "iopub.status.idle": "2026-10-24T17:48:15.729405Z", "shell.execute_reply": "2026-10-24T17:48:15.728298Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "tail = adjusted.tail(21)\n", "fig, ax = plt.subplots(figsize=(10, 4.8))\n", "ax.bar(tail.day_index, tail.eventual, color=LINE, width=0.8,\n", " label=\"what the day will eventually show\")\n", "ax.bar(tail.day_index, tail.reported, color=BLUE, width=0.8, label=\"reported today\")\n", "ok = tail.dropna(subset=[\"adjusted\"])\n", "ax.errorbar(ok.day_index, ok.adjusted,\n", " yerr=[ok.adjusted - ok.adjusted_lo, ok.adjusted_hi - ok.adjusted],\n", " fmt=\"o\", color=WARM, ms=5, lw=1.4, capsize=3, label=\"lag-adjusted estimate\")\n", "blank = tail[~tail.trustworthy]\n", "if len(blank):\n", " ax.text(blank.day_index.mean(), tail.eventual.max() * 0.55,\n", " \"too young\\nto estimate\", color=MUTED, ha=\"center\", fontsize=10)\n", "days_ago_axis(ax, [(21, \"3 weeks ago\"), (14, \"2 weeks ago\"), (7, \"a week ago\"), (1, \"yesterday\")])\n", "ax.set_ylabel(\"conversions\")\n", "ax.set_title(\"The last three weeks, corrected for how much of each day is still in flight\")\n", "# the legend sits in a band cleared above the bars and the error bars' tops\n", "ax.set_ylim(0, max(tail.eventual.max(), ok.adjusted_hi.max()) * 1.24)\n", "ax.legend(frameon=False, loc=\"upper center\", ncol=3)\n", "fig.tight_layout()\n", "fig.savefig(f\"{FIGDIR}/fig_adjusted_dashboard.png\", dpi=170)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "483edb9a", "metadata": {}, "source": [ "## Does the interval mean what it says?\n", "\n", "An interval that is right 95% of the time is useful. An interval that is right 60% of\n", "the time is worse than no interval at all, because it is trusted.\n", "\n", "The check replays the whole exercise on fresh accounts drawn from the same process:\n", "fit on what is visible at day 56, estimate the last week's total, and see whether the\n", "eventual truth lands inside. The number that comes out is the coverage, and it should\n", "sit near 95%." ] }, { "cell_type": "code", "execution_count": 15, "id": "07cc3688", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T17:48:15.744445Z", "iopub.status.busy": "2026-10-24T17:48:15.738953Z", "iopub.status.idle": "2026-10-24T18:16:45.161843Z", "shell.execute_reply": "2026-10-24T18:16:45.158246Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "36 of 40 replays contained the truth -> coverage 90%\n", "average interval width: 26.2% of the estimate\n" ] } ], "source": [ "def coverage_replay(n_replicates=40, seed=SEED + 1, horizon_days=7):\n", " \"\"\"Fresh accounts, same process. Each one: fit on today's view, estimate the last\n", " week, compare against what that week eventually delivers.\"\"\"\n", " g = np.random.default_rng(seed)\n", " hits, widths = 0, []\n", " for i in range(n_replicates):\n", " df = simulate_clicks(seed=int(g.integers(1, 2**31)))\n", " cut = observe(df)\n", " f = fit_cure(cut.elapsed.values, cut.observed.values,\n", " start=[np.log(fit[\"cure\"] / (1 - fit[\"cure\"])),\n", " np.log(fit[\"k\"]), np.log(fit[\"lam\"])])\n", " bt = bootstrap_fit(cut.elapsed.values, cut.observed.values, B=60,\n", " seed=int(g.integers(1, 2**31)))\n", " recent_days = df[df.click_day >= TODAY - horizon_days]\n", " reported = int((recent_days.convert_day <= TODAY).sum())\n", " eventual = int(recent_days.will_convert.sum())\n", " ages = TODAY - recent_days.click_day.values\n", " def total(k, lam):\n", " return float(np.sum((recent_days.convert_day.values <= TODAY)\n", " / np.maximum(reported_fraction(ages, k, lam), 1e-6)))\n", " point = total(f[\"k\"], f[\"lam\"])\n", " draws = np.array([predict_total(reported, ages, r.k, r.lam, g)\n", " for r in bt.itertuples()])\n", " lo, hi = np.percentile(draws, [2.5, 97.5])\n", " hits += int(lo <= eventual <= hi)\n", " widths.append((hi - lo) / max(point, 1e-9))\n", " return n_replicates, hits, hits / n_replicates, float(np.mean(widths))\n", "\n", "\n", "n_rep, n_hit, coverage, mean_width = coverage_replay()\n", "print(f\"{n_hit} of {n_rep} replays contained the truth -> coverage {coverage:.0%}\")\n", "print(f\"average interval width: {mean_width:.1%} of the estimate\")" ] }, { "cell_type": "markdown", "id": "43228ee4", "metadata": {}, "source": [ "## The numbers this post quotes\n", "\n", "Everything the article, the widget and the video say is written here, so a re-run of\n", "this notebook regenerates every figure quoted anywhere in the bundle." ] }, { "cell_type": "code", "execution_count": 16, "id": "83dc76e8", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T18:16:45.226172Z", "iopub.status.busy": "2026-10-24T18:16:45.221822Z", "iopub.status.idle": "2026-10-24T18:16:45.306375Z", "shell.execute_reply": "2026-10-24T18:16:45.305126Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{\n", " \"fit\": {\n", " \"cure\": 0.0617,\n", " \"k\": 0.856,\n", " \"lam\": 3.982,\n", " \"cure_profile_lo\": 0.059,\n", " \"cure_profile_hi\": 0.065,\n", " \"cure_identified\": true,\n", " \"km_plateau\": 0.9383,\n", " \"km_implied_cure\": 0.0617\n", " },\n", " \"lookback\": {\n", " \"days\": 23.7,\n", " \"lo\": 21.5,\n", " \"hi\": 26.0,\n", " \"truth\": 24.1,\n", " \"half_life_days\": 2.6,\n", " \"drop_censored_days\": 20.7,\n", " \"no_cure_days\": 2468291.4,\n", " \"no_cure_runaway\": true\n", " },\n", " \"last_week\": {\n", " \"reported\": 88,\n", " \"eventual\": 158,\n", " \"reported_share\": 0.557\n", " },\n", " \"yesterday\": {\n", " \"reported\": 3,\n", " \"eventual\": 16,\n", " \"age_days\": 0.5,\n", " \"reported_fraction\": 0.1558\n", " }\n", "}\n" ] } ], "source": [ "worked = {\n", " \"label\": \"synthetic\",\n", " \"seed\": SEED,\n", " \"account\": {\n", " \"window_days\": WINDOW_DAYS, \"clicks_per_day\": CLICKS_PER_DAY,\n", " \"total_clicks\": int(len(clicks)),\n", " \"eventual_converters\": eventual_converters,\n", " \"reported_today\": int(cohort.observed.sum()),\n", " \"true_cure\": TRUE_CURE, \"true_k\": TRUE_K, \"true_lambda\": TRUE_LAMBDA,\n", " },\n", " \"fit\": {\n", " \"cure\": round(fit[\"cure\"], 4), \"k\": round(fit[\"k\"], 3),\n", " \"lam\": round(fit[\"lam\"], 3),\n", " \"cure_profile_lo\": round(cure_lo, 3), \"cure_profile_hi\": round(cure_hi, 3),\n", " \"cure_identified\": cure_identified,\n", " \"km_plateau\": round(km_plateau, 4),\n", " \"km_implied_cure\": round(1 - km_plateau, 4),\n", " },\n", " \"horizons\": horizons.to_dict(orient=\"records\"),\n", " \"lookback\": {\n", " \"days\": round(look, 1), \"lo\": round(look_lo, 1), \"hi\": round(look_hi, 1),\n", " \"truth\": round(look_truth, 1), \"half_life_days\": round(half, 1),\n", " \"drop_censored_days\": round(look_naive, 1),\n", " \"no_cure_days\": round(look_nocure, 1), \"no_cure_runaway\": NOCURE_RUNAWAY,\n", " },\n", " \"platform_support\": {\n", " \"edge_days\": round(float(sup_t[-1]), 2), \"window_days\": round(sup_window, 1),\n", " \"reported_share_day1\": round(sup_share[1], 4), \"reported_share_day7\": round(sup_share[7], 4),\n", " },\n", " \"last_week\": {\n", " \"reported\": int(by_day.tail(7).reported.sum()),\n", " \"eventual\": int(by_day.tail(7).eventual.sum()),\n", " \"reported_share\": round(float(by_day.tail(7).reported.sum()\n", " / by_day.tail(7).eventual.sum()), 4),\n", " },\n", " \"yesterday\": {\n", " \"reported\": int(by_day.iloc[-1].reported),\n", " \"eventual\": int(by_day.iloc[-1].eventual),\n", " \"age_days\": round(float(by_day.iloc[-1].age_days), 2),\n", " \"reported_fraction\": round(float(reported_fraction(by_day.iloc[-1].age_days,\n", " fit[\"k\"], fit[\"lam\"])), 4),\n", " },\n", " # The last three weeks, day by day. The video's opening scene is this series and nothing\n", " # else; without it the deck would have to hand-type the bars it draws.\n", " \"daily\": [\n", " {\"day_index\": int(r.day_index), \"age_days\": round(float(TODAY - r.day_index - 0.5), 2),\n", " \"reported\": int(r.reported), \"eventual\": int(r.eventual)}\n", " for r in by_day.tail(21).itertuples()\n", " ],\n", " \"coverage_check\": {\n", " \"replicates\": n_rep, \"inside\": n_hit, \"coverage\": round(coverage, 3),\n", " \"mean_relative_width\": round(mean_width, 4),\n", " },\n", " \"min_trust_fraction\": MIN_TRUST_FRACTION,\n", "}\n", "Path(\"data/worked_example.json\").write_text(json.dumps(worked, indent=2) + \"\\n\",\n", " encoding=\"utf-8\")\n", "print(json.dumps({k: worked[k] for k in (\"fit\", \"lookback\", \"last_week\", \"yesterday\")},\n", " indent=2))" ] }, { "cell_type": "code", "execution_count": 17, "id": "a51c1a29", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T18:16:45.311874Z", "iopub.status.busy": "2026-10-24T18:16:45.311483Z", "iopub.status.idle": "2026-10-24T18:16:45.335092Z", "shell.execute_reply": "2026-10-24T18:16:45.333861Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "wrote 4 pinned sets\n", " this account (fitted) day-1 0.264 | day-7 0.802 | window 23.7d | first trustworthy day 2\n", " fast: impulse buys day-1 0.499 | day-7 0.997 | window 5.6d | first trustworthy day 1\n", " slow: considered purchase day-1 0.144 | day-7 0.487 | window 91.9d | first trustworthy day 4\n", " very front-loaded day-1 0.438 | day-7 0.844 | window 31.9d | first trustworthy day 1\n" ] } ], "source": [ "# The widget recomputes the curve from three parameters; these are the sets it must\n", "# reproduce. They are computed here so the widget is checked against the notebook\n", "# rather than against itself.\n", "def check_set(name, k, lam, cure):\n", " return {\n", " \"name\": name, \"k\": k, \"lam\": lam, \"cure\": cure,\n", " \"expected\": {\n", " \"reported_fraction\": {str(N): round(float(reported_fraction(N, k, lam)), 6)\n", " for N in (1, 2, 3, 7, 14, 21, 30)},\n", " \"lag_factor\": {str(N): round(float(1 / reported_fraction(N, k, lam)), 6)\n", " for N in (1, 2, 3, 7, 14, 21, 30)},\n", " \"lookback_99\": round(float(lookback_days(k, lam)), 6),\n", " \"half_life\": round(float(lam * (-np.log(0.5)) ** (1 / k)), 6),\n", " \"first_trustworthy_day\": next(\n", " (N for N in range(1, 61)\n", " if float(reported_fraction(N, k, lam)) >= MIN_TRUST_FRACTION), None),\n", " },\n", " }\n", "\n", "checks = {\n", " \"label\": \"synthetic\",\n", " \"min_trust_fraction\": MIN_TRUST_FRACTION,\n", " \"tolerance\": 1e-4,\n", " \"sets\": [\n", " check_set(\"this account (fitted)\", round(fit[\"k\"], 3), round(fit[\"lam\"], 3),\n", " round(fit[\"cure\"], 4)),\n", " check_set(\"fast: impulse buys\", 1.1, 1.4, 0.09),\n", " check_set(\"slow: considered purchase\", 0.75, 12.0, 0.031),\n", " check_set(\"very front-loaded\", 0.6, 2.5, 0.05),\n", " ],\n", "}\n", "Path(\"data/checks.json\").write_text(json.dumps(checks, indent=2) + \"\\n\", encoding=\"utf-8\")\n", "print(f\"wrote {len(checks['sets'])} pinned sets\")\n", "for s in checks[\"sets\"]:\n", " e = s[\"expected\"]\n", " print(f\" {s['name']:26} day-1 {e['reported_fraction']['1']:.3f} | \"\n", " f\"day-7 {e['reported_fraction']['7']:.3f} | \"\n", " f\"window {e['lookback_99']:.1f}d | first trustworthy day {e['first_trustworthy_day']}\")" ] }, { "cell_type": "markdown", "id": "a6a39c27", "metadata": {}, "source": [ "---\n", "\n", "# Part two: run it on your own data\n", "\n", "Everything below runs on your machine. Your file is read from the folder you point it\n", "at, and nothing is uploaded anywhere.\n", "\n", "## What the file needs to look like\n", "\n", "**Entry point A: one row per click.** The direct format. Columns:\n", "\n", "| column | what it is | example |\n", "|---|---|---|\n", "| `click_date` | when the click happened | `2026-08-14` |\n", "| `converted` | 1 if this click has converted, 0 if not (yet) | `0` |\n", "| `conversion_date` | when it converted; blank when `converted` is 0 | `2026-08-17` |\n", "\n", "**Entry point B: a raw event log**, one row per touch, which the consolidator below\n", "turns into entry point A. Columns: `user_id`, `event_time`, `event_name` (with\n", "whatever your conversion event is called). Set `MY_EVENT_LOG` and `CONVERSION_EVENT` in\n", "the cell below instead of `MY_CSV`.\n", "\n", "An optional `campaign` column lets you fit one campaign at a time (`CAMPAIGN` below):\n", "campaigns that convert on different clocks should not share one curve. A `.xlsx` file\n", "works as well as a CSV. Dates written 2026-08-14 work, and so do 14/08/2026 or\n", "08/14/2026 once some day in the file shows which part is the day; when none does, the\n", "notebook stops and asks you to set `DAY_FIRST`. Timestamps with a time of day are read\n", "to the minute.\n", "\n", "**Where to get it.** No ad platform or GA4 screen exports one row per click with the\n", "clicks still waiting in it. Google Ads' *Days to conversion* segment counts only the\n", "conversions, so it feeds the spreadsheet that ships with this post, not this notebook.\n", "The file comes from GA4's BigQuery export, where one query over `events_*` pulls each\n", "user's first paid click and their first conversion after it (the recipe is in the\n", "give-to-your-AI brief that ships with this post), or from a CRM or server log that\n", "records clicks and conversions against a user id.\n", "\n", "The one thing that matters: **the file has to include the clicks that have not\n", "converted.** A file of conversions only cannot measure this, because the clicks still\n", "waiting are the entire signal." ] }, { "cell_type": "code", "execution_count": 18, "id": "5010a22c", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T18:16:45.348362Z", "iopub.status.busy": "2026-10-24T18:16:45.347880Z", "iopub.status.idle": "2026-10-24T18:16:45.462503Z", "shell.execute_reply": "2026-10-24T18:16:45.461150Z" } }, "outputs": [], "source": [ "# ---- export reader (shared across every Stochastic Strata notebook; do not edit here) ----------\n", "# One reader for the file you bring. Ad platforms and analytics tools do not hand over clean tables:\n", "# a Google Ads download opens with a title line and a date-range line and closes with \"Total:\" rows,\n", "# a GA4 download opens with \"#\" lines and writes dates as 20261019, a European download separates\n", "# with semicolons and writes 1.234,56, and Excel turns numbers into \"1,234.56\" or \"€12.30\". This\n", "# reader finds the real header row, sets the title, comment and total rows aside, turns formatted\n", "# numbers back into numbers, and reads dates in the platforms' own forms. Where a file is genuinely\n", "# ambiguous - 06/07/2026 could be the 6th of July or the 7th of June - it stops and asks rather than\n", "# guessing, because a guessed date silently moves a day to the wrong weekday.\n", "#\n", "# table, notes = read_export(\"my_export.csv\", columns={\"date\": [\"day\", \"date\"], ...})\n", "# for note in notes: print(note)\n", "#\n", "# `columns` maps the names this notebook uses to the names your file might use (case and spacing do\n", "# not matter, and a currency suffix such as \"(EUR)\" is ignored). `date_columns` are read as dates,\n", "# `id_columns` are kept as text. Everything else that reads as numbers becomes numbers; a share\n", "# written \"49.41%\" becomes 0.4941. A share the platform shows only as \"< 10%\" or \"> 90%\" is read as\n", "# 0.10 or 0.90 and the direction is kept in a column ending \"_bound\", because it is a limit, not a\n", "# measurement. Every row set aside, and why, comes back in `notes`.\n", "import csv as _csv\n", "import io as _io\n", "import re as _re\n", "from pathlib import Path as _Path\n", "\n", "import numpy as _np\n", "import pandas as _pd\n", "\n", "\n", "class ExportReadError(ValueError):\n", " \"\"\"The file cannot be read without guessing; the message says what to change.\"\"\"\n", "\n", "\n", "_TOTAL_ROW = _re.compile(r\"^\\s*(grand\\s+)?totals?\\b\", _re.I)\n", "_CURRENCY = _re.compile(r\"[€$£¥₹]|\\b(EUR|USD|GBP|CHF|SEK|NOK|DKK|PLN|AUD|CAD)\\b\", _re.I)\n", "_EMPTY = {\"\", \"--\", \"-\", \"—\", \"–\", \"n/a\", \"na\", \"nan\", \"none\", \"null\", \"(not set)\"}\n", "_MONTHS = {m: i for i, m in enumerate(\n", " [\"jan\", \"feb\", \"mar\", \"apr\", \"may\", \"jun\", \"jul\", \"aug\", \"sep\", \"oct\", \"nov\", \"dec\"], 1)}\n", "\n", "\n", "def _norm(name):\n", " name = _re.sub(r\"\\((?:[A-Z]{3}|[€$£])\\)\", \"\", str(name))\n", " return _re.sub(r\"[\\s_\\-./]+\", \" \", name).strip().lower()\n", "\n", "\n", "def _decode(raw):\n", " if raw[:2] in (b\"\\xff\\xfe\", b\"\\xfe\\xff\"):\n", " return raw.decode(\"utf-16\")\n", " for encoding in (\"utf-8-sig\", \"cp1252\", \"latin-1\"):\n", " try:\n", " return raw.decode(encoding)\n", " except UnicodeDecodeError:\n", " continue\n", " return raw.decode(\"utf-8\", errors=\"replace\")\n", "\n", "\n", "def _rows_from(source):\n", " \"\"\"Every row of the file as a list of strings, whatever the container.\"\"\"\n", " if isinstance(source, _pd.DataFrame):\n", " return [list(map(str, source.columns))] + source.astype(str).values.tolist(), \",\"\n", " path = _Path(source)\n", " if not path.exists():\n", " raise ExportReadError(f\"There is no file at {path}. Check the name and the folder.\")\n", " if path.suffix.lower() in (\".xlsx\", \".xlsm\", \".xls\"):\n", " sheet = _pd.read_excel(path, header=None, dtype=str).fillna(\"\")\n", " return sheet.values.tolist(), \",\"\n", " text = _decode(path.read_bytes())\n", " lines = [ln for ln in text.splitlines() if not ln.lstrip().startswith(\"#\")]\n", " counts = {d: sum(ln.count(d) for ln in lines[:60]) for d in (\",\", \";\", \"\\t\")}\n", " delimiter = max(counts, key=counts.get) if any(counts.values()) else \",\"\n", " return [row for row in _csv.reader(_io.StringIO(\"\\n\".join(lines)), delimiter=delimiter)], delimiter\n", "\n", "\n", "def _find_header(rows, wanted):\n", " \"\"\"The first row that names the columns: the one matching most expected names, else the first\n", " row as wide as the data below it that is mostly words rather than numbers.\"\"\"\n", " widths = [len([c for c in r if str(c).strip()]) for r in rows]\n", " if wanted:\n", " best, best_hits = None, 0\n", " for i, row in enumerate(rows[:40]):\n", " names = {_norm(c) for c in row if str(c).strip()}\n", " hits = sum(1 for options in wanted.values() if names & set(options))\n", " if hits > best_hits:\n", " best, best_hits = i, hits\n", " if best is not None:\n", " return best\n", " body = [w for w in widths if w > 1]\n", " width = max(set(body), key=body.count) if body else 1\n", " for i, row in enumerate(rows[:40]):\n", " cells = [str(c).strip() for c in row if str(c).strip()]\n", " if len(cells) >= max(2, width - 1) and sum(_looks_number(c) for c in cells) <= len(cells) / 2:\n", " return i\n", " return 0\n", "\n", "\n", "def _looks_number(text):\n", " if _re.fullmatch(r\"\\d{1,4}[./\\-]\\d{1,2}[./\\-]\\d{1,4}\", str(text).strip()):\n", " return False # a date such as 22.09.2026, not a number with separators\n", " return bool(_re.fullmatch(r\"[<>]?\\s*[-+(]?[€$£]?\\s*[\\d.,\\s']+\\s*%?\\)?\", str(text).strip()))\n", "\n", "\n", "def _decimal_comma(values):\n", " \"\"\"True when these values write 1.234,56 (a comma followed by one or two final digits, or dots\n", " grouping thousands with a comma after them), False when they write 1,234.56, None when they do\n", " not say (whole numbers only, or a lone '1.340').\"\"\"\n", " comma_votes = dot_votes = 0\n", " for v in values:\n", " s = _re.sub(r\"[^\\d.,]\", \"\", v)\n", " if \",\" in s and \".\" in s:\n", " comma_votes += s.rfind(\",\") > s.rfind(\".\")\n", " dot_votes += s.rfind(\".\") > s.rfind(\",\")\n", " elif _re.search(r\",\\d{1,2}$\", s):\n", " comma_votes += 1\n", " elif _re.search(r\"\\.\\d{1,2}$\", s) or _re.search(r\"\\.\\d{4,}$\", s):\n", " dot_votes += 1\n", " if comma_votes == dot_votes:\n", " return None\n", " return comma_votes > dot_votes\n", "\n", "\n", "def _to_number(text, comma_decimal):\n", " s = str(text).strip().replace(\" \", \" \")\n", " if s.lower() in _EMPTY:\n", " return _np.nan, None, True\n", " bound = None\n", " if s[:1] in \"<>\":\n", " bound, s = s[0], s[1:].strip()\n", " negative = s.startswith(\"(\") and s.endswith(\")\")\n", " s = _CURRENCY.sub(\"\", s).strip(\"() \").replace(\" \", \"\").replace(\"'\", \"\")\n", " percent = s.endswith(\"%\")\n", " s = s.rstrip(\"%\")\n", " if comma_decimal:\n", " s = s.replace(\".\", \"\").replace(\",\", \".\")\n", " else:\n", " s = s.replace(\",\", \"\")\n", " try:\n", " value = float(s)\n", " except ValueError:\n", " return _np.nan, None, False\n", " if negative:\n", " value = -value\n", " return (value / 100.0 if percent else value), bound, True\n", "\n", "\n", "def _parse_dates(values, column, dayfirst):\n", " text = [str(v).strip() for v in values]\n", " present = [t for t in text if t.lower() not in _EMPTY]\n", " if not present:\n", " return _pd.Series([_pd.NaT] * len(text)), []\n", " notes = []\n", " if all(_re.fullmatch(r\"\\d{8}\", t) for t in present):\n", " return _pd.to_datetime(_pd.Series(text), format=\"%Y%m%d\", errors=\"coerce\"), notes\n", " if all(_re.fullmatch(r\"\\d{10}(\\.\\d+)?|\\d{13}|\\d{16}\", t) for t in present):\n", " digits = len(present[0].split(\".\")[0])\n", " unit = {10: \"s\", 13: \"ms\", 16: \"us\"}[digits]\n", " notes.append(f\"'{column}' holds epoch times in {dict(s='seconds', ms='milliseconds', us='microseconds')[unit]}; read as such.\")\n", " return _pd.to_datetime(_pd.to_numeric(_pd.Series(text), errors=\"coerce\"), unit=unit), notes\n", " cleaned = [_re.sub(r\"\\s+(UTC|GMT|Z)$\", \"\", t) for t in text]\n", " slash = [t for t in cleaned if _re.fullmatch(r\"\\d{1,2}[/.\\-]\\d{1,2}[/.\\-]\\d{2,4}(\\s.*)?\", t)]\n", " if slash and len(slash) >= len(present) * 0.9:\n", " firsts = [int(_re.split(r\"[/.\\-]\", t)[0]) for t in slash]\n", " seconds = [int(_re.split(r\"[/.\\-]\", t)[1]) for t in slash]\n", " dotted = all(\".\" in t.split(\" \")[0] for t in slash)\n", " if dayfirst is None:\n", " if max(firsts) > 12 and max(seconds) <= 12:\n", " dayfirst = True\n", " elif max(seconds) > 12 and max(firsts) <= 12:\n", " dayfirst = False\n", " elif dotted:\n", " dayfirst = True\n", " else:\n", " raise ExportReadError(\n", " f\"The dates in '{column}' (for example {slash[0]}) could be day-first or month-first, \"\n", " \"and nothing in the file says which. Re-run with dayfirst=True for 6 July written \"\n", " \"06/07, or dayfirst=False for June 7 written 06/07 - or export the dates as \"\n", " \"2026-07-06.\")\n", " return _pd.to_datetime(_pd.Series(cleaned), dayfirst=dayfirst, errors=\"coerce\"), notes\n", " parsed = _pd.to_datetime(_pd.Series(cleaned), errors=\"coerce\", format=\"mixed\")\n", " return parsed, notes\n", "\n", "\n", "def read_export(source, columns=None, date_columns=(), id_columns=(), dayfirst=None,\n", " max_unreadable_share=0.02):\n", " \"\"\"Read a downloaded report into a table with this notebook's column names. Returns\n", " (table, notes); raises ExportReadError with a plain-words fix when it cannot read without\n", " guessing.\"\"\"\n", " columns = columns or {}\n", " # Each name's candidates in the caller's order of preference, the name itself first.\n", " wanted = {name: list(dict.fromkeys([_norm(name)] + [_norm(a) for a in aliases]))\n", " for name, aliases in columns.items()}\n", " rows, delimiter = _rows_from(source)\n", " rows = [r for r in rows if any(str(c).strip() for c in r)]\n", " if not rows:\n", " raise ExportReadError(\"The file is empty.\")\n", " notes = []\n", " h = _find_header(rows, wanted)\n", " if h:\n", " notes.append(f\"set aside {h} line(s) above the header (a report title or date range).\")\n", " header = [str(c).strip() for c in rows[h]]\n", " body, totals = [], 0\n", " for r in rows[h + 1:]:\n", " first = next((str(c) for c in r if str(c).strip()), \"\")\n", " if _TOTAL_ROW.match(first):\n", " totals += 1\n", " continue\n", " if [str(c).strip() for c in r] == header:\n", " continue\n", " body.append([str(c) for c in r] + [\"\"] * (len(header) - len(r)))\n", " if totals:\n", " notes.append(f\"set aside {totals} total row(s).\")\n", " frame = _pd.DataFrame([r[:len(header)] for r in body], columns=header)\n", " rename = {}\n", " for name, options in wanted.items():\n", " by_norm = {}\n", " for c in frame.columns:\n", " by_norm.setdefault(_norm(c), c)\n", " match = next((by_norm[o] for o in options if o in by_norm and by_norm[o] not in rename), None)\n", " if match is not None:\n", " rename[match] = name\n", " frame = frame.rename(columns=rename)\n", " missing = [name for name in columns if name not in frame.columns]\n", " if missing:\n", " seen = \", \".join(f\"'{c}'\" for c in header[:12])\n", " raise ExportReadError(\n", " f\"Could not find the column(s) {', '.join(missing)}. The file's columns are {seen}. \"\n", " \"Rename the column in the file, or add its name to the list this cell accepts.\")\n", " dates = set(date_columns)\n", " # The file's own convention, for a column that cannot tell by itself: a clicks column of whole\n", " # numbers says nothing, but its \"1.340\" means one thousand three hundred and forty when the\n", " # file's other columns write decimals with a comma.\n", " every = [str(v) for c in frame.columns if c not in date_columns for v in frame[c].tolist()\n", " if _looks_number(v)]\n", " file_comma = _decimal_comma(every)\n", " if file_comma is None:\n", " file_comma = delimiter == \";\"\n", " ids = set(id_columns) | {c for c in frame.columns if _re.search(r\"(^|[ _])(id|pseudo id)$\", _norm(c))}\n", " for col in frame.columns:\n", " values = frame[col].tolist()\n", " if col in dates:\n", " parsed, extra = _parse_dates(values, col, dayfirst)\n", " notes += extra\n", " bad = int(parsed.isna().sum() - sum(str(v).strip().lower() in _EMPTY for v in values))\n", " if bad > max(1, max_unreadable_share * len(values)):\n", " example = next(v for v, p in zip(values, parsed) if _pd.isna(p) and str(v).strip())\n", " raise ExportReadError(f\"{bad} value(s) in '{col}' are not dates I can read, for example '{example}'.\")\n", " frame[col] = parsed.values\n", " continue\n", " if col in ids:\n", " continue\n", " present = [v for v in values if str(v).strip().lower() not in _EMPTY]\n", " if not present or sum(_looks_number(v) for v in present) < 0.5 * len(present):\n", " continue\n", " comma = _decimal_comma(present)\n", " if comma is None:\n", " comma = file_comma\n", " parsed = [_to_number(v, comma) for v in values]\n", " unreadable = [v for v, (_, _, ok) in zip(values, parsed) if not ok]\n", " if unreadable and col not in columns:\n", " notes.append(f\"'{col}' mixes numbers with labels such as '{unreadable[0]}'; kept as text.\")\n", " continue\n", " if unreadable:\n", " notes.append(f\"{len(unreadable)} value(s) in '{col}' could not be read as numbers, for example \"\n", " f\"'{unreadable[0]}'; they are left empty.\")\n", " if len(unreadable) > max(1, max_unreadable_share * len(values)):\n", " raise ExportReadError(\n", " f\"{len(unreadable)} of {len(values)} values in '{col}' are not numbers I can read, for \"\n", " f\"example '{unreadable[0]}'. Fix them in the file, or export the column unformatted.\")\n", " frame[col] = [p[0] for p in parsed]\n", " bounds = [p[1] for p in parsed]\n", " if any(bounds):\n", " # An object column, so a day with no limit stays None: a newer pandas would infer a string\n", " # column here and turn None into NaN, which is truthy and reads as a limit.\n", " frame[f\"{col}_bound\"] = _pd.Series(bounds, index=frame.index, dtype=object)\n", " notes.append(f\"'{col}' has values the platform shows only as a limit (for example \"\n", " f\"'{next(v for v, b in zip(values, bounds) if b)}'); kept as the limit, with the \"\n", " f\"direction in '{col}_bound'.\")\n", " if dates:\n", " undated = frame[list(dates)].isna().all(axis=1)\n", " if undated.any():\n", " frame = frame[~undated].reset_index(drop=True)\n", " notes.append(f\"set aside {int(undated.sum())} row(s) with no date (GA4 writes its grand total \"\n", " \"this way, under the header).\")\n", " return frame, notes\n", "# ---- end of export reader -------------------------------------------------------------------------" ] }, { "cell_type": "code", "execution_count": 19, "id": "da153e98", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T18:16:45.636359Z", "iopub.status.busy": "2026-10-24T18:16:45.633955Z", "iopub.status.idle": "2026-10-24T18:16:45.730632Z", "shell.execute_reply": "2026-10-24T18:16:45.729480Z" } }, "outputs": [], "source": [ "# ---------------------------------------------------------------- YOUR FILE HERE\n", "MY_CSV = \"\" # e.g. \"clicks_export.csv\" or \".xlsx\", one row per click; empty runs the demos\n", "MY_EVENT_LOG = \"\" # or a raw per-touch log (user_id, event_time, event_name) instead of MY_CSV\n", "CONVERSION_EVENT = \"\" # the log's name for a conversion, e.g. \"purchase\"; only with MY_EVENT_LOG\n", "CAMPAIGN = \"\" # fit one campaign, by its name in a 'campaign' column; empty fits every row\n", "TODAY_OVERRIDE = \"\" # e.g. \"2026-09-06\", the day you pulled the file; empty uses its latest date\n", "DAY_FIRST = None # True if 06/07/2026 means 6 July, False if 7 June; only when the file cannot tell\n", "# -------------------------------------------------------------------------------\n", "\n", "COLUMN_ALIASES = {\n", " \"click_date\": [\"click date\", \"click_time\", \"click time\", \"click_day\", \"date\"],\n", " \"converted\": [\"conversion\", \"is_converted\", \"converted_flag\"],\n", " \"conversion_date\": [\"conversion date\", \"conv_date\", \"conversion_time\", \"conversion time\",\n", " \"converted_at\"],\n", " \"campaign\": [\"campaign name\", \"campaign_name\", \"utm_campaign\"],\n", "}\n", "OPTIONAL_COLUMNS = (\"conversion_date\", \"campaign\")\n", "\n", "\n", "def _read_clicks(path, aliases, date_columns, dayfirst):\n", " \"\"\"The shared reader, asked for every column; a missing optional one is dropped and asked again.\"\"\"\n", " wanted = dict(aliases)\n", " while True:\n", " try:\n", " return read_export(path, columns=wanted,\n", " date_columns=[c for c in date_columns if c in wanted],\n", " dayfirst=dayfirst)\n", " except ExportReadError as exc:\n", " text = str(exc)\n", " gone = [c for c in OPTIONAL_COLUMNS if c in wanted\n", " and text.startswith(\"Could not find the column(s)\")\n", " and c in text.split(\". The file's\")[0]]\n", " if not gone:\n", " raise\n", " for c in gone:\n", " del wanted[c]\n", "\n", "\n", "def survival_rows(df, today=None, campaign=\"\"):\n", " \"\"\"A table of clicks (click_date, converted, conversion_date, optional campaign) in the\n", " survival shape: how long each click has been watched, and whether its conversion is in.\n", "\n", " `today` is the day the file was pulled. A click after it is set aside, and a conversion\n", " after it is counted as not yet in — the clicks still waiting on that day are the signal.\n", " \"\"\"\n", " df = df.copy()\n", " if \"campaign\" in df.columns:\n", " names = df[\"campaign\"].astype(str).str.strip()\n", " if campaign:\n", " if campaign not in set(names):\n", " raise ValueError(f\"No rows for campaign '{campaign}'. The file's campaigns: \"\n", " f\"{sorted(set(names))[:12]}\")\n", " df = df[names == campaign]\n", " elif names.nunique() > 1:\n", " top = names.value_counts().head(5)\n", " print(f\"! {names.nunique()} campaigns in one file, fitted as one curve: \"\n", " + \", \".join(f\"{n} ({c:,} clicks)\" for n, c in top.items())\n", " + \". Campaigns that convert on different clocks give a curve that fits none of \"\n", " \"them; set CAMPAIGN to fit one at a time.\")\n", " click = pd.to_datetime(df[\"click_date\"], errors=\"coerce\")\n", " conv = (pd.to_datetime(df[\"conversion_date\"], errors=\"coerce\")\n", " if \"conversion_date\" in df.columns else pd.Series(pd.NaT, index=df.index))\n", " flag = pd.to_numeric(df[\"converted\"], errors=\"coerce\").fillna(0) > 0\n", " with_time = bool(((click.dropna() - click.dropna().dt.normalize()) > pd.Timedelta(seconds=0)).any())\n", " unreadable = click.isna()\n", " if unreadable.all():\n", " raise ValueError(\"None of the click dates could be read as dates. Expected something \"\n", " \"like 2026-08-14.\")\n", " no_date = flag & conv.isna()\n", " dated_unflagged = ~flag & conv.notna()\n", " converted = flag | dated_unflagged\n", " if today:\n", " now = pd.to_datetime(today)\n", " if with_time:\n", " now = now.normalize() + pd.Timedelta(days=1) # \"today\" means through the end of that day\n", " else:\n", " now = max(click.max(), conv.max()) if conv.notna().any() else click.max()\n", " after = click > now\n", " late = converted & ~after & (conv > now)\n", " converted = converted & ~late\n", " days = lambda delta: delta.dt.total_seconds().to_numpy() / 86400.0\n", " elapsed = np.where(converted, days(conv - click), days(now - click)).astype(float)\n", " before_click = converted & (elapsed < 0)\n", " keep = ~unreadable & ~after & ~(converted & no_date) & ~before_click\n", " for n, why in [\n", " (int(unreadable.sum()), \"click date that could not be read\"),\n", " (int(after.sum()), f\"click after the day treated as today ({now.date()})\"),\n", " (int(no_date.sum()), \"marked converted with no conversion date, so its delay is unknown\"),\n", " (int(before_click.sum()), \"conversion dated before its click\")]:\n", " if n:\n", " print(f\"note: {n:,} row(s) set aside — {why}.\")\n", " if late.sum():\n", " print(f\"note: {int(late.sum()):,} conversion(s) dated after {now.date()} counted as not yet \"\n", " \"in, as they were on that day.\")\n", " if dated_unflagged.sum():\n", " print(f\"note: {int(dated_unflagged.sum()):,} row(s) with a conversion date but converted = 0 \"\n", " \"read as converted.\")\n", " # Whole-day dates: a same-day conversion has a delay of zero days, read as half a day.\n", " floor = 1 / 1440 if with_time else 0.5\n", " return pd.DataFrame({\n", " \"elapsed\": np.maximum(elapsed[keep.values], floor),\n", " \"observed\": converted.values[keep.values].astype(bool),\n", " }), now\n", "\n", "\n", "def load_clicks(path, today=None, campaign=\"\", dayfirst=None):\n", " \"\"\"Read a click-level CSV or .xlsx into the survival shape, with plain-words errors.\"\"\"\n", " try:\n", " df, notes = _read_clicks(path, COLUMN_ALIASES, [\"click_date\", \"conversion_date\"], dayfirst)\n", " except ExportReadError as exc:\n", " if \"dayfirst=\" not in str(exc):\n", " raise\n", " raise ExportReadError(f\"{exc} Here that is DAY_FIRST, in the YOUR FILE HERE cell.\") from None\n", " for n in notes:\n", " print(\"note:\", n)\n", " return survival_rows(df, today, campaign)\n", "\n", "\n", "def consolidate_event_log(path, conversion_event, click_event=None):\n", " \"\"\"Entry point B: a raw per-touch log becomes one row per click.\n", "\n", " Each user's first touch is the click; their first conversion event after it, if\n", " any, is the conversion. Users with no conversion stay in as censored rows, which\n", " is the whole point.\n", " \"\"\"\n", " df, notes = read_export(path, columns={\n", " \"user_id\": [\"user_pseudo_id\", \"user id\", \"client_id\", \"customer_id\"],\n", " \"event_time\": [\"event_timestamp\", \"event time\", \"timestamp\", \"time\"],\n", " \"event_name\": [\"event name\", \"event\"]}, date_columns=[\"event_time\"], id_columns=[\"user_id\"])\n", " for n in notes:\n", " print(\"note:\", n)\n", " df = df.dropna(subset=[\"event_time\"]).sort_values(\"event_time\")\n", " rows = []\n", " for user, g in df.groupby(\"user_id\"):\n", " clicks_g = g[g.event_name == click_event] if click_event else g\n", " if clicks_g.empty:\n", " continue\n", " first_click = clicks_g.event_time.iloc[0]\n", " convs = g[(g.event_name == conversion_event) & (g.event_time >= first_click)]\n", " rows.append({\"click_date\": first_click,\n", " \"converted\": int(len(convs) > 0),\n", " \"conversion_date\": convs.event_time.iloc[0] if len(convs) else pd.NaT})\n", " return pd.DataFrame(rows)" ] }, { "cell_type": "code", "execution_count": 20, "id": "c5551055", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T18:16:45.781878Z", "iopub.status.busy": "2026-10-24T18:16:45.781019Z", "iopub.status.idle": "2026-10-24T18:16:45.838723Z", "shell.execute_reply": "2026-10-24T18:16:45.837365Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "MY_CSV is empty — running the two demonstrations below instead.\n" ] } ], "source": [ "def sufficiency_report(elapsed, observed, horizon=7):\n", " \"\"\"Refuse loudly rather than answer quietly. Returns (verdict, messages).\"\"\"\n", " msgs, verdict = [], \"OK\"\n", " n_conv = int(np.sum(observed))\n", " if n_conv < 30:\n", " return \"REFUSED\", [\n", " f\"Only {n_conv} converted clicks in this file. Below about 30 the curve \"\n", " \"cannot be fitted at all — the answer would be the starting guess, not \"\n", " \"your data.\"]\n", " if n_conv < 100:\n", " verdict = \"WIDE\"\n", " msgs.append(f\"{n_conv} converted clicks is enough to fit, but the interval \"\n", " \"will be wide. Treat the direction as the finding, not the size.\")\n", " censored_share = float(np.mean(~np.asarray(observed, bool)))\n", " if censored_share > 0.995:\n", " verdict = \"WIDE\"\n", " msgs.append(\"Almost every click in this file is unconverted. The eventual-\"\n", " \"conversion share will be poorly pinned down.\")\n", " return verdict, msgs\n", "\n", "\n", "def run_on_your_data(elapsed, observed, horizons=REFERENCE_HORIZONS):\n", " \"\"\"The same fit as part one, with the same refusals.\"\"\"\n", " verdict, msgs = sufficiency_report(elapsed, observed)\n", " for m in msgs:\n", " print(\"!\", m)\n", " if verdict == \"REFUSED\":\n", " print(\"\\nNo numbers reported — see above.\")\n", " return None\n", " f = fit_cure(np.asarray(elapsed, float), np.asarray(observed, bool))\n", " lo, hi, identified = profile_cure(np.asarray(elapsed, float),\n", " np.asarray(observed, bool), centre=f[\"cure\"])\n", " print(f\"\\neventual-conversion share: {f['cure']:.2%} \"\n", " f\"(profile interval {lo:.1%} to {hi:.1%})\")\n", " if not identified:\n", " print(\"! The share that never converts is not identified by this file — the \"\n", " \"history is too short for the curve to flatten. The window below is an \"\n", " \"over-estimate; it errs long. Do not widen it further.\")\n", " window = lookback_days(f['k'], f['lam'])\n", " span = float(np.max(elapsed))\n", " print(f\"half of conversions in by day \"\n", " f\"{f['lam'] * (-np.log(0.5)) ** (1 / f['k']):.1f}; \"\n", " f\"99% in by day {window:.1f}\")\n", " if window > span:\n", " print(f\"! The window reaches day {window:.0f}, past the {span:.0f} days this file \"\n", " \"spans. Beyond your history the curve is the shape of the fitted tail, not \"\n", " \"conversions anyone has seen: read the window, and every row marked below, \"\n", " \"as a projection.\")\n", " print(\"\\n days since click | share reported | multiply reported by\")\n", " for N in horizons:\n", " frac = float(reported_fraction(N, f[\"k\"], f[\"lam\"]))\n", " beyond = \" (projection)\" if N > span else \"\"\n", " if frac < MIN_TRUST_FRACTION:\n", " print(f\"{N:>17} | {frac:>13.1%} | too little in yet — no estimate{beyond}\")\n", " else:\n", " print(f\"{N:>17} | {frac:>13.1%} | {1 / frac:.2f}{beyond}\")\n", " # The platform's reading of the same file: the Kaplan-Meier curve inside its support.\n", " s_t, s_F = support_curve(elapsed, observed)\n", " if s_F.size and s_F[-1] > 0:\n", " s_win = float(s_t[np.searchsorted(s_F, LOOKBACK_QUANTILE * s_F[-1] * (1 - 1e-12))])\n", " print(f\"\\nthe platform's reading, inside the {s_t[-1]:.0f} days your history supports: \"\n", " f\"window {s_win:.1f} days (the fit's {window:.1f})\")\n", " if s_win >= 0.8 * s_t[-1]:\n", " print(\"! The curve is still rising near the edge of your history: the conversion tail \"\n", " \"likely runs past it, so this window and these shares read short.\")\n", " for N in horizons:\n", " if N >= s_t[-1]:\n", " print(f\"{N:>17} | past your history — no reading\")\n", " else:\n", " share = float(s_F[np.searchsorted(s_t, N, side=\"right\") - 1] / s_F[-1])\n", " print(f\"{N:>17} | {share:>13.1%} reported\")\n", " if np.all(np.isin(np.mod(np.asarray(elapsed, float), 1.0), (0.0, 0.5))):\n", " print(\"\\nYour dates are whole days, so 'day N' here means conversions dated within N days of \"\n", " \"the click's date. The platform's reading measures exactly that; the fit takes each day \"\n", " \"count as an exact delay and reads the first few days low. For the first week act on the platform's \"\n", " \"reading; past it the two agree, and beyond your history only the fit reads at all. \"\n", " \"Yesterday, read this morning, is day 0: its conversions dated the same day.\")\n", " return f\n", "\n", "\n", "if MY_EVENT_LOG:\n", " if not CONVERSION_EVENT:\n", " raise ValueError(\"Set CONVERSION_EVENT to the log's name for a conversion, e.g. 'purchase'.\")\n", " my_clicks = consolidate_event_log(MY_EVENT_LOG, CONVERSION_EVENT)\n", " print(f\"{len(my_clicks):,} users with a first touch in the log\")\n", " my_cohort, my_today = survival_rows(my_clicks, TODAY_OVERRIDE or None)\n", "elif MY_CSV:\n", " my_cohort, my_today = load_clicks(MY_CSV, TODAY_OVERRIDE or None, CAMPAIGN, DAY_FIRST)\n", "if MY_CSV or MY_EVENT_LOG:\n", " print(f\"{len(my_cohort):,} clicks read; treating {my_today.date()} as today\")\n", " my_fit = run_on_your_data(my_cohort.elapsed.values, my_cohort.observed.values)\n", "else:\n", " print(\"MY_CSV is empty — running the two demonstrations below instead.\")" ] }, { "cell_type": "markdown", "id": "1b37f5c3", "metadata": {}, "source": [ "### The two demonstrations\n", "\n", "The first is a second synthetic account with a different truth and a slow tail, to show\n", "what happens when the conversions take longer than the history: the fit answers without\n", "warning and misses the true window by about a third. The platform's reading, printed\n", "below it, misses too, and in the other direction: it reads only as far as the history\n", "reaches, so it comes out short. It warns only when the curve is still rising in the last\n", "fifth of the history, which here it is not, so no warning is not proof the tail has ended. The second is a deliberately thin file, to\n", "show what the refusal looks like, because a method that always answers is a method\n", "that will answer wrongly." ] }, { "cell_type": "code", "execution_count": 21, "id": "2dba3df2", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T18:16:45.845378Z", "iopub.status.busy": "2026-10-24T18:16:45.843518Z", "iopub.status.idle": "2026-10-24T18:17:06.102343Z", "shell.execute_reply": "2026-10-24T18:17:06.100533Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "wrote data/example_clicks.csv — a slow-converting account (truth: 3.1% convert, window 92 days)\n", "\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "/tmp/ipykernel_916080/3649983762.py:64: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n", " with_time = bool(((click.dropna() - click.dropna().dt.normalize()) > pd.Timedelta(seconds=0)).any())\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "eventual-conversion share: 2.94% (profile interval 2.7% to 3.2%)\n", "half of conversions in by day 7.4; 99% in by day 63.1\n", "\n", " days since click | share reported | multiply reported by\n", " 1 | 11.2% | too little in yet — no estimate\n", " 2 | 19.6% | too little in yet — no estimate\n", " 3 | 26.9% | too little in yet — no estimate\n", " 7 | 48.4% | 2.07\n", " 14 | 70.4% | 1.42\n", " 21 | 82.5% | 1.21\n", "\n", "the platform's reading, inside the 67 days your history supports: window 48.0 days (the fit's 63.1)\n", " 1 | 19.4% reported\n", " 2 | 26.5% reported\n", " 3 | 31.2% reported\n", " 7 | 52.0% reported\n", " 14 | 70.7% reported\n", " 21 | 83.9% reported\n", "\n", "Your dates are whole days, so 'day N' here means conversions dated within N days of the click's date. The platform's reading measures exactly that; the fit takes each day count as an exact delay and reads the first few days low. For the first week act on the platform's reading; past it the two agree, and beyond your history only the fit reads at all. Yesterday, read this morning, is day 0: its conversions dated the same day.\n" ] } ], "source": [ "if MY_CSV or MY_EVENT_LOG:\n", " print(\"Your file is set, so the demonstrations are skipped.\")\n", "else:\n", " demo = simulate_clicks(n_days=70, per_day=260, cure=0.031, k=0.75, lam=12.0, seed=99)\n", " demo_cut = observe(demo, today=70.0)\n", " demo_path = Path(\"data/example_clicks.csv\")\n", " pd.DataFrame({\n", " \"click_date\": pd.Timestamp(\"2026-06-01\") + pd.to_timedelta(demo.click_day.astype(int), \"D\"),\n", " \"converted\": (demo.convert_day <= 70.0).astype(int),\n", " \"conversion_date\": [\n", " (pd.Timestamp(\"2026-06-01\") + pd.to_timedelta(int(cd), unit=\"D\")).date() if cd <= 70.0 else None\n", " for cd in demo.convert_day],\n", " }).to_csv(demo_path, index=False)\n", " print(f\"wrote {demo_path} — a slow-converting account (truth: 3.1% convert, window \"\n", " f\"{lookback_days(0.75, 12.0):.0f} days)\\n\")\n", "\n", " demo_cohort, demo_today = load_clicks(demo_path, \"2026-08-10\")\n", " demo_fit = run_on_your_data(demo_cohort.elapsed.values, demo_cohort.observed.values)" ] }, { "cell_type": "code", "execution_count": 22, "id": "18126ee7", "metadata": { "execution": { "iopub.execute_input": "2026-10-24T18:17:06.119485Z", "iopub.status.busy": "2026-10-24T18:17:06.113275Z", "iopub.status.idle": "2026-10-24T18:17:06.924431Z", "shell.execute_reply": "2026-10-24T18:17:06.923176Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "120 clicks, 2 converted\n", "\n", "! Only 2 converted clicks in this file. Below about 30 the curve cannot be fitted at all — the answer would be the starting guess, not your data.\n", "\n", "No numbers reported — see above.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "/tmp/ipykernel_916080/3649983762.py:64: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n", " with_time = bool(((click.dropna() - click.dropna().dt.normalize()) > pd.Timedelta(seconds=0)).any())\n" ] } ], "source": [ "if not (MY_CSV or MY_EVENT_LOG):\n", " # Two conversions among 120 clicks, both early, read on the file's last day.\n", " thin = pd.DataFrame({\n", " \"click_date\": pd.date_range(\"2026-09-01\", periods=40, freq=\"D\").repeat(3),\n", " \"converted\": [1, 1] + [0] * 118,\n", " \"conversion_date\": [\"2026-09-03\", \"2026-09-05\"] + [None] * 118,\n", " })\n", " thin_path = Path(\"data/example_thin.csv\")\n", " thin.to_csv(thin_path, index=False)\n", " thin_cohort, _ = load_clicks(thin_path)\n", " print(f\"{len(thin_cohort)} clicks, {int(thin_cohort.observed.sum())} converted\\n\")\n", " run_on_your_data(thin_cohort.elapsed.values, thin_cohort.observed.values)" ] }, { "cell_type": "markdown", "id": "694fcd4f", "metadata": {}, "source": [ "### What to do with the answer\n", "\n", "Three things, in order of how much they change.\n", "\n", "**Stop reading the last few days as performance.** Whatever the multiplier says for a\n", "day that young, the honest read of the most recent days is that they are not\n", "readable yet.\n", "\n", "**Set the reporting window to the number your history supports**, not to the ad\n", "platform's default. If 99% of conversions are in by day twenty-two, a seven-day window is\n", "reporting roughly four fifths of the result and calling it the result.\n", "\n", "**Re-fit it periodically.** Conversion timing moves: seasonal peaks pull it in,\n", "consent and tracking changes push it out. A curve fitted once and trusted for a year\n", "is a curve that will be wrong at exactly the moment it matters." ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.12.3" } }, "nbformat": 4, "nbformat_minor": 5 }