test_lsh_mips.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0d839b9f-a2ab-40c0-b04c-04daa47685c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import necessary modules\n",
    "import numpy as np\n",
    "from algorithms.lsh_mips import LSH_MIPS\n",
    "from utils.constants import SCALING_TOPK, SCALING_NUM_HFUNC, SCALING_NUM_TABLES, ASYMMETRIC, SYMMETRIC\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c761a9eb-9a47-4852-a1ee-6dc10c9d4832",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 5 candidates: [47 81 30 31 93]\n",
      "Budget (number of computations): 5777\n"
     ]
    }
   ],
   "source": [
    "# Generate synthetic test data\n",
    "def generate_test_data(num_atoms, len_signal):\n",
    "    atoms = np.random.randn(num_atoms, len_signal)\n",
    "    signals = np.random.randn(1, len_signal)  # Single signal\n",
    "    return atoms, signals\n",
    "\n",
    "# Parameters for testing\n",
    "num_atoms = 100\n",
    "len_signal = 50\n",
    "atoms, signals = generate_test_data(num_atoms, len_signal)\n",
    "\n",
    "# Initialize LSH_MIPS object and test\n",
    "lsh_mips = LSH_MIPS(\n",
    "    m=3, \n",
    "    num_hfunc=5, \n",
    "    num_tables=10, \n",
    "    type=ASYMMETRIC\n",
    ")\n",
    "\n",
    "# Preprocess and run the LSH_MIPS algorithm\n",
    "lsh_mips.preprocess_data(atoms)\n",
    "lsh_mips.preprocess_query(signals[0])\n",
    "candidates, budget = lsh_mips.run(top_k=5)\n",
    "\n",
    "# Output results\n",
    "print(f\"Top 5 candidates: {candidates}\")\n",
    "print(f\"Budget (number of computations): {budget}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "2754d541-7695-49a2-aaf1-dd939b847212",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of LSH_MIPS: 80.0%\n"
     ]
    }
   ],
   "source": [
    "# Naive method: Calculate the top k atoms based on the highest inner product with the signal\n",
    "inner_products = np.dot(atoms, signals[0])\n",
    "top_k_naive = np.argsort(inner_products)[-5:]  # Get the indices of top 5 atoms\n",
    "\n",
    "# Evaluate accuracy\n",
    "accuracy = len(np.intersect1d(candidates, top_k_naive)) / 5\n",
    "print(f\"Accuracy of LSH_MIPS: {accuracy * 100}%\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8c33617f-e4b0-48f6-9caf-11bd00334a3f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "f49fce0c",
   "metadata": {},
   "source": [
    "Movie Lens 100k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c0cc0527-9c68-4df1-983c-93554d1ce6f6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 candidates: [ 12 161 321 103 104 163 106  81 345 453]\n",
      "Budget (number of computations): 86920\n",
      "Accuracy compared to naive method: 80.0%\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "from algorithms.lsh_mips import LSH_MIPS\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from surprise import Dataset, Reader\n",
    "from surprise.model_selection import train_test_split\n",
    "\n",
    "# Load the MovieLens dataset (100k version for simplicity)\n",
    "reader = Reader(line_format='user item rating timestamp', sep='\\t')\n",
    "data = Dataset.load_builtin('ml-100k')\n",
    "trainset, testset = train_test_split(data, test_size=0.2)\n",
    "\n",
    "# Build a user-item matrix\n",
    "user_item_matrix = np.zeros((trainset.n_users, trainset.n_items))\n",
    "for uid, iid, rating in trainset.all_ratings():\n",
    "    user_item_matrix[int(uid), int(iid)] = rating\n",
    "\n",
    "# Use SVD to reduce dimensionality\n",
    "svd = TruncatedSVD(n_components=50)\n",
    "atoms = svd.fit_transform(user_item_matrix.T)  # Transpose to get item embeddings\n",
    "\n",
    "# Create a signal (simulate a user's preference vector)\n",
    "user_id = 0  # Choose a user ID from the dataset\n",
    "user_ratings = user_item_matrix[user_id]\n",
    "signal = np.dot(user_ratings, atoms)  # Weighted average of rated item embeddings\n",
    "\n",
    "# Initialize and run LSH_MIPS\n",
    "lsh_mips = LSH_MIPS(m=3, num_hfunc=5, num_tables=10, type=\"ASYMMETRIC\")\n",
    "lsh_mips.preprocess_data(atoms)\n",
    "lsh_mips.preprocess_query(signal)\n",
    "\n",
    "# Get top 5 candidates\n",
    "top_k = 10\n",
    "candidates, budget = lsh_mips.run(top_k=top_k)\n",
    "\n",
    "# Compare with naive method for evaluation\n",
    "inner_products = np.dot(atoms, signal)\n",
    "top_k_naive = np.argsort(inner_products)[-top_k:]\n",
    "\n",
    "accuracy = len(np.intersect1d(candidates, top_k_naive)) / top_k\n",
    "print(f\"Top {top_k} candidates: {candidates}\")\n",
    "print(f\"Budget (number of computations): {budget}\")\n",
    "print(f\"Accuracy compared to naive method: {accuracy * 100}%\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a2c22692",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "943"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(user_item_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17dac07e",
   "metadata": {},
   "source": [
    "Movie Lens 1M"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "19d97464",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 candidates: [ 80 256 704  99  86 272 295 276  27  50]\n",
      "Budget (number of computations): 96672\n",
      "Accuracy compared to naive method: 80.0%\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "from algorithms.lsh_mips import LSH_MIPS\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from surprise import Dataset, Reader\n",
    "from surprise.model_selection import train_test_split\n",
    "\n",
    "# Load the MovieLens dataset (100k version for simplicity)\n",
    "reader = Reader(line_format='user item rating timestamp', sep='\\t')\n",
    "data = Dataset.load_builtin('ml-1m')\n",
    "trainset, testset = train_test_split(data, test_size=0.2)\n",
    "\n",
    "# Build a user-item matrix\n",
    "user_item_matrix = np.zeros((trainset.n_users, trainset.n_items))\n",
    "for uid, iid, rating in trainset.all_ratings():\n",
    "    user_item_matrix[int(uid), int(iid)] = rating\n",
    "\n",
    "# Use SVD to reduce dimensionality\n",
    "svd = TruncatedSVD(n_components=50)\n",
    "atoms = svd.fit_transform(user_item_matrix.T)  # Transpose to get item embeddings\n",
    "\n",
    "# Create a signal (simulate a user's preference vector)\n",
    "user_id = 6039  # Choose a user ID from the dataset\n",
    "user_ratings = user_item_matrix[user_id]\n",
    "signal = np.dot(user_ratings, atoms)  # Weighted average of rated item embeddings\n",
    "\n",
    "# Initialize and run LSH_MIPS\n",
    "lsh_mips = LSH_MIPS(m=3, num_hfunc=5, num_tables=10, type=\"ASYMMETRIC\")\n",
    "lsh_mips.preprocess_data(atoms)\n",
    "lsh_mips.preprocess_query(signal)\n",
    "\n",
    "# Get top 5 candidates\n",
    "top_k = 10\n",
    "candidates, budget = lsh_mips.run(top_k=top_k)\n",
    "\n",
    "# Compare with naive method for evaluation\n",
    "inner_products = np.dot(atoms, signal)\n",
    "top_k_naive = np.argsort(inner_products)[-top_k:]\n",
    "\n",
    "accuracy = len(np.intersect1d(candidates, top_k_naive)) / top_k\n",
    "print(f\"Top {top_k} candidates: {candidates}\")\n",
    "print(f\"Budget (number of computations): {budget}\")\n",
    "print(f\"Accuracy compared to naive method: {accuracy * 100}%\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b59563fe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6040"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(user_item_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60c98d70",
   "metadata": {},
   "source": [
    "Netflix Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "041e448c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 candidates: [ 738 1305 1274 1171  559  305  179  314  217   24]\n",
      "Budget (number of computations): 105884\n",
      "Accuracy compared to naive method: 90.0%\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from algorithms.lsh_mips import LSH_MIPS\n",
    "\n",
    "# Load preprocessed factors and biases\n",
    "movie_factors = np.load(\"../BanditMIPS/data/netflix/Movie_factors_15_new.npy\")\n",
    "movie_biases = np.load(\"../BanditMIPS/data/netflix/Movie_biases_15_new.npy\")\n",
    "customer_factors = np.load(\"../BanditMIPS/data/netflix/Customer_factors_15_new.npy\")\n",
    "customer_biases = np.load(\"../BanditMIPS/data/netflix/Customer_biases_15_new.npy\")\n",
    "global_mean = np.load(\"../BanditMIPS/data/netflix/netflix_global_mean.npy\")\n",
    "\n",
    "# Use movie factors as atoms for LSH-MIPS\n",
    "atoms = movie_factors\n",
    "\n",
    "# Choose a specific user\n",
    "user_id = 0  # Replace with the desired user ID (index-based, starting at 0)\n",
    "user_factors = customer_factors[user_id]\n",
    "user_bias = customer_biases[user_id]\n",
    "\n",
    "# Construct the user preference signal\n",
    "signal = user_factors  # Optionally, add user_bias and global_mean if needed for personalization\n",
    "\n",
    "# Initialize and run LSH-MIPS\n",
    "lsh_mips = LSH_MIPS(m=3, num_hfunc=5, num_tables=10, type=\"ASYMMETRIC\")\n",
    "lsh_mips.preprocess_data(atoms)\n",
    "lsh_mips.preprocess_query(signal)\n",
    "\n",
    "# Get top 10 similar movies\n",
    "top_k = 10\n",
    "candidates, budget = lsh_mips.run(top_k=top_k)\n",
    "\n",
    "# Naive method for evaluation (optional)\n",
    "inner_products = np.dot(atoms, signal)\n",
    "top_k_naive = np.argsort(inner_products)[-top_k:]\n",
    "\n",
    "# Evaluate accuracy\n",
    "accuracy = len(np.intersect1d(candidates, top_k_naive)) / top_k\n",
    "print(f\"Top {top_k} candidates: {candidates}\")\n",
    "print(f\"Budget (number of computations): {budget}\")\n",
    "print(f\"Accuracy compared to naive method: {accuracy * 100}%\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "f15e823f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 217,   24,  305, 1305, 1171,  314,  559,   46,  738, 1274],\n",
       "      dtype=int64)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_k_naive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "9bd6a430",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dimensions = [0, 2500, 5000, 7500, 10000, 12500, 15000, 17500, 20000]\n",
    "results = []\n",
    "\n",
    "for d in dimensions:\n",
    "    # Reduce dimensions of movie_factors\n",
    "    atoms_d = movie_factors[:, :d]\n",
    "    signal_d = user_factors[:d]\n",
    "\n",
    "    lsh_mips.preprocess_data(atoms_d)\n",
    "    lsh_mips.preprocess_query(signal_d)\n",
    "    _, budget = lsh_mips.run(top_k=10)\n",
    "\n",
    "    results.append((d, budget))\n",
    "\n",
    "# Plot results\n",
    "import matplotlib.pyplot as plt\n",
    "dimensions, budgets = zip(*results)\n",
    "plt.plot(dimensions, budgets, marker=\"o\", label=\"LSH-MIPS Budget\")\n",
    "plt.xlabel(\"Signal Vector Size (d)\")\n",
    "plt.ylabel(\"Computational Complexity (Budget)\")\n",
    "plt.title(\"Scalability of LSH-MIPS with Signal Vector Size\")\n",
    "plt.legend()\n",
    "plt.grid()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b6ebb18",
   "metadata": {},
   "source": [
    "Crypto-Pairs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c7449766",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top 10 candidates: [2]\n",
      "Budget (number of computations): 2756\n",
      "Accuracy compared to naive method: 10.00%\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "from algorithms.lsh_mips import LSH_MIPS\n",
    "\n",
    "# Load the preprocessed dataset\n",
    "dataset_path = \"../BanditMIPS/data/crypto_pairs/crypto_pairs_1m_dimensions.npy\"  # Path to the saved .npy file\n",
    "crypto_data = np.load(dataset_path, allow_pickle=True)\n",
    "\n",
    "# Step 1: Use Truncated SVD to reduce dimensionality (if necessary)\n",
    "svd = TruncatedSVD(n_components=50)  # Reduce to 50 dimensions\n",
    "atoms = svd.fit_transform(crypto_data)  # The reduced dataset\n",
    "\n",
    "# Step 2: Create a query signal\n",
    "# Example: Use the first crypto pair as the query vector\n",
    "query_index = 0\n",
    "signal = atoms[query_index]  # A single crypto pair's embedding\n",
    "\n",
    "# Step 3: Initialize and run LSH-MIPS\n",
    "lsh_mips = LSH_MIPS(m=3, num_hfunc=5, num_tables=10, type=\"ASYMMETRIC\")\n",
    "lsh_mips.preprocess_data(atoms)  # Preprocess the dataset\n",
    "lsh_mips.preprocess_query(signal)  # Preprocess the query vector\n",
    "\n",
    "# Step 4: Retrieve top-k candidates\n",
    "top_k = 10\n",
    "candidates, budget = lsh_mips.run(top_k=top_k)\n",
    "\n",
    "# Step 5: Evaluate with naive method for comparison\n",
    "inner_products = np.dot(atoms, signal)\n",
    "top_k_naive = np.argsort(inner_products)[-top_k:]\n",
    "\n",
    "# Compare results\n",
    "accuracy = len(np.intersect1d(candidates, top_k_naive)) / top_k\n",
    "print(f\"Top {top_k} candidates: {candidates}\")\n",
    "print(f\"Budget (number of computations): {budget}\")\n",
    "print(f\"Accuracy compared to naive method: {accuracy * 100:.2f}%\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b65465e9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 49,  57, 104,  35,  97,  58,   2,  51,  28,  13], dtype=int64)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_k_naive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "92172a7d",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}