02_zero_shot_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3860afdb",
   "metadata": {},
   "source": [
    "## 大语言模型 - 10行代码的情感分析解决方案"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "843b09f3",
   "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 : -0.081472\n"
     ]
    }
   ],
   "source": [
    "import openai\n",
    "import os\n",
    "from openai.embeddings_utils import cosine_similarity, get_embedding\n",
    "\n",
    "openai.api_key = os.environ.get(\"OPENAI_API_KEY\")\n",
    "EMBEDDING_MODEL = \"text-embedding-ada-002\"\n",
    "\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(\"随意降价，不予价保，服务态度差\")\n",
    "\n",
    "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": "markdown",
   "id": "8b19192a",
   "metadata": {},
   "source": [
    "## 语序问题"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "90a515e9",
   "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": [
    "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": "markdown",
   "id": "1b611338",
   "metadata": {},
   "source": [
    "## 在更大的数据集上的真实案例"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "97dac5a2",
   "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",
    "# convert 5-star rating to binary sentiment\n",
    "df = df[df.Score != 3]\n",
    "df[\"sentiment\"] = df.Score.replace({1: \"negative\", 2: \"negative\", 4: \"positive\", 5: \"positive\"})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7a8346aa",
   "metadata": {},
   "source": [
    "## 生成评测结果"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "38513ac7",
   "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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\n",
      "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.'])\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.9"
  },
  "vscode": {
   "interpreter": {
    "hash": "8114e84f04cf14e493992e1b725447accf84073d5ec18e7063d492738bf032cb"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}