Upload 3行要約_Scraping.ipynb

3行要約_Scraping.ipynb

@@ -0,0 +1,607 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "f5308d26",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "  8%|▊         | 264/3200 [16:53<2:51:12,  3.50s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 11%|█         | 350/3200 [22:08<2:46:27,  3.50s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 14%|█▎        | 435/3200 [27:25<2:55:49,  3.82s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 15%|█▍        | 479/3200 [30:17<2:46:24,  3.67s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 29%|██▊       | 919/3200 [59:14<2:24:16,  3.79s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 50%|████▉     | 1598/3200 [1:43:30<1:35:05,  3.56s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 52%|█████▏    | 1667/3200 [1:47:59<1:31:01,  3.56s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 52%|█████▎    | 1680/3200 [1:48:47<1:29:22,  3.53s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\r",
+      " 53%|█████▎    | 1681/3200 [1:48:50<1:25:53,  3.39s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 53%|█████▎    | 1696/3200 [1:49:46<1:30:02,  3.59s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 53%|█████▎    | 1703/3200 [1:50:12<1:27:52,  3.52s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 69%|██████▊   | 2193/3200 [2:20:32<59:49,  3.56s/it]  "
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 70%|███████   | 2240/3200 [2:23:29<57:22,  3.59s/it]  "
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 72%|███████▏  | 2309/3200 [2:27:49<52:10,  3.51s/it]  "
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 75%|███████▌  | 2406/3200 [2:33:56<47:49,  3.61s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 76%|███████▌  | 2421/3200 [2:34:53<46:57,  3.62s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 78%|███████▊  | 2483/3200 [2:38:41<41:46,  3.50s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 83%|███████��▎ | 2657/3200 [2:49:21<31:59,  3.53s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 83%|████████▎ | 2667/3200 [2:49:57<30:42,  3.46s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 84%|████████▍ | 2690/3200 [2:51:21<29:56,  3.52s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 85%|████████▍ | 2708/3200 [2:52:27<28:48,  3.51s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 85%|████████▍ | 2710/3200 [2:52:34<28:06,  3.44s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 85%|████████▍ | 2713/3200 [2:52:44<27:39,  3.41s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 85%|████████▌ | 2725/3200 [2:53:28<27:43,  3.50s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 86%|████████▌ | 2749/3200 [2:54:56<26:28,  3.52s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      " 86%|████████▋ | 2761/3200 [2:55:41<26:05,  3.57s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Exception\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|██████████| 3200/3200 [3:22:49<00:00,  3.80s/it]\n"
+     ]
+    }
+   ],
+   "source": [
+    "from urllib.request import urlopen\n",
+    "from bs4 import BeautifulSoup\n",
+    "from bs4.element import NavigableString\n",
+    "from pprint import pprint\n",
+    "import time\n",
+    "import tqdm\n",
+    "\n",
+    "# 収集するニュース記事のインデックス\n",
+    "start_index = 2800 # 開始インデックス\n",
+    "end_index = 6000 # 終了インデックス\n",
+    "\n",
+    "\n",
+    "\n",
+    "\n",
+    "# コンテンツの取得\n",
+    "def get_content(id):\n",
+    "    # サーバに負荷をかけないように3秒スリープ\n",
+    "    time.sleep(3)\n",
+    "\n",
+    "    URL = 'https://news.livedoor.com/article/detail/'+id+'/'\n",
+    "#     print(URL)\n",
+    "    try:\n",
+    "        with urlopen(URL) as res:\n",
+    "            # 本文の抽出\n",
+    "            output1 = ''\n",
+    "            html = res.read().decode('euc_jp', 'ignore')\n",
+    "            soup = BeautifulSoup(html, 'html.parser')\n",
+    "            line_list = soup.select('.articleBody p')\n",
+    "            for line in line_list:\n",
+    "                if len(line.contents) > 0 and type(line.contents[0]) == NavigableString:\n",
+    "                    output1 += line.contents[0].strip()\n",
+    "            if output1 == '': # 記事がない\n",
+    "                return\n",
+    "   \n",
+    "\n",
+    "            # 要約の抽出\n",
+    "            output0 = ''\n",
+    "            summary_list = soup.select('.summaryList li')\n",
+    "            for summary in summary_list:\n",
+    "                output0 += summary.contents[0].strip()+'\\n'\n",
+    "            output0 = output0.strip(\"\\n\")\n",
+    "            if output0 == '': # 記事がない\n",
+    "                return\n",
+    "\n",
+    "            # 出力\n",
+    "#             print(output0+output1)\n",
+    "            \n",
+    "            prompt =  f\"\"\"Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n",
+    "### Instruction:\n",
+    "以下の文を３行で要約しなさい\n",
+    "### Input:\n",
+    "{output1}\n",
+    "### Response:\n",
+    "{output0}\n",
+    "\"\"\"\n",
+    "            return prompt\n",
+    "    except Exception:\n",
+    "        print('Exception')\n",
+    "\n",
+    "# IDリストの生成の取得\n",
+    "id_list = []\n",
+    "with open('downloaded_train.csv', mode='r') as f:\n",
+    "    lines = f.readlines()\n",
+    "    for line in lines:\n",
+    "        id = line.strip().split(',')[3].split('.')[0]\n",
+    "        id_list.append(id)\n",
+    "\n",
+    "# コンテンツの取得\n",
+    "output = []\n",
+    "for i in tqdm.tqdm(range(start_index, end_index)):\n",
+    "#     print('index:', i)\n",
+    "    gen = get_content(id_list[i])\n",
+    "    if gen!= \"\":\n",
+    "        output.append(gen)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "309fe137",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "CSVファイル 'three_line_summarize_2800_6000_.csv' が作成されました。\n"
+     ]
+    }
+   ],
+   "source": [
+    "import csv\n",
+    "\n",
+    "def create_csv_from_list(data_list, file_path):\n",
+    "    # CSVファイルのカラム名\n",
+    "    fieldnames = [\"text\"]\n",
+    "\n",
+    "    # CSVファイルを書き込む\n",
+    "    with open(file_path, 'w', newline='', encoding='utf-8') as csvfile:\n",
+    "        writer = csv.writer(csvfile)\n",
+    "\n",
+    "        # カラム名を書き込む\n",
+    "        writer.writerow(fieldnames)\n",
+    "\n",
+    "        # データをリストとしてCSVファイルに書き込む\n",
+    "        for item in data_list:\n",
+    "            writer.writerow([item])\n",
+    "\n",
+    "\n",
+    "\n",
+    "# CSVファイルを作成する\n",
+    "csv_file_path = \"three_line_summarize_2800_6000_.csv\"\n",
+    "out = [i for i in output if i!=None]\n",
+    "create_csv_from_list(out, csv_file_path)\n",
+    "\n",
+    "print(f\"CSVファイル '{csv_file_path}' が作成されました。\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "7ce6ef7e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "3200"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(output)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "7246a989",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "out = [i for i in output if i!=None]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "71015093",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "1454"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(out)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dfa94b6b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import re\n",
+    "\n",
+    "def new_format(text):\n",
+    "    # 正規表現を使用してInputとResponseを抽出\n",
+    "    input_match = re.search(r'Input:(.*?)###', text, re.DOTALL)\n",
+    "    response_match = re.search(r'Response:(.*?)$', text, re.DOTALL)\n",
+    "\n",
+    "    if input_match and response_match:\n",
+    "        input_text = input_match.group(1).strip()\n",
+    "        response_text = response_match.group(1).strip()\n",
+    "    new_format = f\"\"\"<START_Q>{input_text}<END_Q>\n",
+    "<START_A>{response_text}<END_A>\"\"\"\n",
+    "    return new_format\n",
+    "\n",
+    "from datasets import load_dataset, DatasetDict,Dataset\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "import pandas as pd\n",
+    "\n",
+    "# df=pd.read_csv(\"llama2_ft_data/three_line_summarize.csv\", index_col=0)\n",
+    "df=pd.read_csv(\"llama2_ft_data/three_line_summarize_concat.csv\")\n",
+    "df[\"text\"] = df[\"text\"].apply(new_format)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "llama2",
+   "language": "python",
+   "name": "llama2"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.0"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}