add win_prob documentation/notebook

docs/examples/win_prob.ipynb

@@ -0,0 +1,162 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Win Probability Calculation\n",
+    "\n",
+    "We can use the win_probability function to calculate win probabilities at specific ticks in the match.\n",
+    "\n",
+    "First, you need to download the trained model file by running `awpy get winprob`. This will download a `.joblib` file to your current directory.\n",
+    "\n",
+    "Next, you need to import the win_probability function from the awpy library and parse the demo. This can be done by adding the following lines to the top of your code:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from awpy import Demo\n",
+    "from awpy.stats.win_prob import win_probability\n",
+    "dem = Demo('g2-vs-faze-m2-ancient.dem')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's choose some interesting ticks to analyze. For this example, we'll use ticks from different points in the match."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ticks_to_analyze = [110716, 111110, 112935]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Now, let's calculate the win probabilities"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "probabilities = win_probability(dem, ticks_to_analyze)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The final step is to display the results"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(\"Win Probabilities:\")\n",
+    "print(probabilities)"
+   ]
+  },
+  {
+   "cell_type": "raw",
+   "metadata": {},
+   "source": [
+    "Win Probabilities:\n",
+    "     tick  CT_win_probability  T_win_probability\n",
+    "0  110716                0.40               0.60\n",
+    "1  111110                0.37               0.63\n",
+    "2  112935                0.51               0.49"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "vscode": {
+     "languageId": "raw"
+    }
+   },
+   "source": [
+    "You can also analyze a single tick if you're interested in a specific moment"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from awpy import Demo\n",
+    "from awpy.stats.win_prob import win_probability\n",
+    "\n",
+    "# Simply call `Demo(\"...\")` to parse a demo\n",
+    "dem = Demo('g2-vs-faze-m2-ancient.dem')\n",
+    "\n",
+    "single_tick = 153063 \n",
+    "single_tick_probability = win_probability(dem, single_tick)\n",
+    "\n",
+    "print(\"Win Probability for a single tick:\")\n",
+    "print(single_tick_probability)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Training a Custom Model with Awpy\n",
+    "\n",
+    "This guide demonstrates how to use the `awpy train` CLI command to train your own model on a folder of Counter-Strike 2 demo files. This feature allows you to create custom win probability models based on your specific dataset.\n",
+    "\n",
+    "Before running the train model command, ensure you have a folder containing multiple .dem files. For this example, we'll assume you have a folder named `my_demos` with numerous demo files inside.\n",
+    "\n",
+    "### Basic Usage\n",
+    "\n",
+    "To train a model on your demo folder, use the following command:\n",
+    "\n",
+    "`awpy train my_demos/`\n",
+    "\n",
+    "This command will process all the demo files in the `my_demos` folder, extract relevant features, and train a win probability model. The trained model will be saved as `wpa_model_rf.joblib` in the current directory by default.\n",
+    "\n",
+    "### Advanced Usage\n",
+    "\n",
+    "The awpy train model command supports several options to customize the training process:\n",
+    "\n",
+    "`awpy train my_demos/ --output custom_model.joblib --batch-size 20`\n",
+    "\n",
+    "In this example:\n",
+    "\n",
+    "- `output custom_model.joblib` specifies a custom name and location for the output model file.\n",
+    "- `batch-size 20` sets the number of demos to process in each batch, which can help manage memory usage for large datasets.\n",
+    "\n",
+    "\n",
+    "\n",
+    "\n",
+    "\n",
+    "\n",
+    "\n"
+   ]
+  }
+ ],
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}