02_sentiment_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import openai\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from openai.embeddings_utils import cosine_similarity, get_embedding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 获取访问open ai的密钥\n",
    "openai.api_key = 'OPENAI_API_KEY'\n",
    "\n",
    "# 选择模型\n",
    "EMBEDDING_MODEL = \"text-embedding-ada-002\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 获取好评和差评\n",
    "positive_review = get_embedding('好评')\n",
    "negative_review = get_embedding('差评')\n",
    "\n",
    "positive_example = get_embedding('买的银色版真的很好看，一天就到了，晚上就开始拿起来完系统很丝滑流畅，做工扎实，手感细腻，很精致哦苹果一如既往的好品质')\n",
    "negative_example = get_embedding(\"降价厉害，保价不合理，不推荐\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "好评例子的评分： 0.070963\n",
      "差评例子的评分： -0.072895\n"
     ]
    }
   ],
   "source": [
    "def get_score(sample_embedding):\n",
    "    return cosine_similarity(sample_embedding, positive_review) - cosine_similarity(sample_embedding, negative_review)\n",
    "\n",
    "positive_score = get_score(positive_example)\n",
    "negative_score = get_score(negative_example)\n",
    "\n",
    "print(\"好评例子的评分： %f\" % positive_score)\n",
    "print(\"差评例子的评分： %f\" % negative_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "好评餐馆的评分 : 0.062719\n",
      "差评餐馆的评分 : -0.074591\n"
     ]
    }
   ],
   "source": [
    "\n",
    "good_restraurant = get_embedding(\"这家餐馆太好吃了，一点都不糟糕\")\n",
    "bad_restraurant = get_embedding(\"这家餐馆太糟糕了，一点都不好吃\")\n",
    "\n",
    "good_score = get_score(good_restraurant)\n",
    "bad_score = get_score(bad_restraurant)\n",
    "print(\"好评餐馆的评分 : %f\" % (good_score))\n",
    "print(\"差评餐馆的评分 : %f\" % (bad_score))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.metrics import classification_report\n",
    "\n",
    "datafile_path = './data/fine_food_reviews_with_embeddings_1k.csv'\n",
    "\n",
    "df = pd.read_csv(datafile_path)\n",
    "df['embedding'] = df.embedding.apply(eval).apply(np.array)\n",
    "\n",
    "# 将五星评级转换为二元情绪\n",
    "df = df[df.Score != 3]\n",
    "df[\"sentiment\"] = df.Score.replace({1: \"negative\", 2: \"negative\", 4:\"positive\", 5:\"positive\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "    negative       0.98      0.73      0.84       136\n",
      "    positive       0.96      1.00      0.98       789\n",
      "\n",
      "    accuracy                           0.96       925\n",
      "   macro avg       0.97      0.86      0.91       925\n",
      "weighted avg       0.96      0.96      0.96       925\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import PrecisionRecallDisplay\n",
    "\n",
    "def evaluate_embeddings_approach(\n",
    "    labels = ['negative', 'positive'], \n",
    "    model = EMBEDDING_MODEL,\n",
    "):\n",
    "    label_embeddings = [get_embedding(label, engine=model) for label in labels]\n",
    "\n",
    "    def label_score(review_embedding, label_embeddings):\n",
    "        return cosine_similarity(review_embedding, label_embeddings[1]) - cosine_similarity(review_embedding, label_embeddings[0])\n",
    "\n",
    "    probas = df[\"embedding\"].apply(lambda x: label_score(x, label_embeddings))\n",
    "    preds = probas.apply(lambda x: 'positive' if x>0 else 'negative')\n",
    "\n",
    "    report = classification_report(df.sentiment, preds)\n",
    "    print(report)\n",
    "\n",
    "    display = PrecisionRecallDisplay.from_predictions(df.sentiment, probas, pos_label='positive')\n",
    "    _ = display.ax_.set_title(\"2-class Precision-Recall curve\")\n",
    "\n",
    "evaluate_embeddings_approach(labels=['An Amazon review with a negative sentiment.', 'An Amazon review with a positive sentiment.'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "gpt",
   "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": "3.10.10"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}