{"id":534,"date":"2026-10-07T14:27:28","date_gmt":"2026-10-07T06:27:28","guid":{"rendered":"https:\/\/www.linerroom.cn\/?p=534"},"modified":"2026-10-07T14:27:29","modified_gmt":"2026-10-07T06:27:29","slug":"langchain-%e6%95%b0%e6%8d%ae%e5%8a%a0%e8%bd%bd%e5%99%a8","status":"publish","type":"post","link":"https:\/\/www.linerroom.cn\/?p=534","title":{"rendered":"LangChain \u6570\u636e\u52a0\u8f7d\u5668"},"content":{"rendered":"\n<div class=\"wp-block-jetpack-markdown\"><h1>LangChain \u6570\u636e\u52a0\u8f7d\u5668\u8be6\u89e3<\/h1>\n<blockquote>\n<p>\u672c\u6587\u53ea\u4ecb\u7ecd LangChain \u7684\u6570\u636e\u52a0\u8f7d\u5668\uff08Document Loaders\uff09\uff1a\u5b83\u4eec\u662f\u4ec0\u4e48\u3001\u6709\u54ea\u4e9b\u7c7b\u578b\u3001\u5982\u4f55\u4f7f\u7528\uff0c\u4ee5\u53ca\u5982\u4f55\u6839\u636e\u6570\u636e\u6e90\u9009\u62e9\u5408\u9002\u7684\u52a0\u8f7d\u5668\u3002<\/p>\n<\/blockquote>\n<h2>\u76ee\u5f55\u524d\u7f6e\uff1a\u6570\u636e\u52a0\u8f7d\u5668\u529f\u80fd\u6c47\u603b\u8868<\/h2>\n<p>LangChain \u7684\u52a0\u8f7d\u5668\u6765\u81ea\u6838\u5fc3\u5305\u3001<code>langchain-community<\/code> \u4ee5\u53ca\u5927\u91cf\u7b2c\u4e09\u65b9\u96c6\u6210\u5305\u3002\u4e0b\u9762\u6c47\u603b\u672c\u6587\u6d89\u53ca\u7684\u5168\u90e8\u4e3b\u8981\u52a0\u8f7d\u5668\u3002\u7531\u4e8e\u7b2c\u4e09\u65b9\u96c6\u6210\u4f1a\u6301\u7eed\u589e\u52a0\uff0c\u8868\u4e2d\u7684\u201c\u529f\u80fd\u201d\u6309\u6570\u636e\u6765\u6e90\u548c\u5b9e\u9645\u7528\u9014\u6982\u62ec\uff1b\u4f7f\u7528\u524d\u4ecd\u5e94\u4ee5\u5f53\u524d\u7248\u672c\u7684\u5b98\u65b9\u96c6\u6210\u6587\u6863\u4e3a\u51c6\u3002<\/p>\n<table>\n<thead>\n<tr>\n<th>\u7c7b\u522b<\/th>\n<th>Loader<\/th>\n<th>\u4e3b\u8981\u529f\u80fd<\/th>\n<th>\u5e38\u89c1\u4f7f\u7528\u573a\u666f<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>\u672c\u5730\u6587\u4ef6<\/td>\n<td><code>TextLoader<\/code><\/td>\n<td>\u8bfb\u53d6 TXT\u3001Markdown \u7b49\u7eaf\u6587\u672c\u6587\u4ef6<\/td>\n<td>\u672c\u5730\u7b14\u8bb0\u3001\u8bf4\u660e\u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>\u672c\u5730\u6587\u4ef6<\/td>\n<td><code>DirectoryLoader<\/code><\/td>\n<td>\u904d\u5386\u76ee\u5f55\uff0c\u5e76\u628a\u6bcf\u4e2a\u6587\u4ef6\u4ea4\u7ed9\u6307\u5b9a Loader \u6279\u91cf\u5904\u7406<\/td>\n<td>\u6279\u91cf\u5bfc\u5165\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u672c\u5730\u6587\u4ef6<\/td>\n<td><code>CSVLoader<\/code><\/td>\n<td>\u6309\u884c\u6216\u6309\u5b57\u6bb5\u8bfb\u53d6 CSV \u6570\u636e\uff0c\u5e76\u8f6c\u6362\u4e3a\u6587\u6863<\/td>\n<td>FAQ\u3001\u5546\u54c1\u8868\u3001\u6570\u636e\u8868<\/td>\n<\/tr>\n<tr>\n<td>\u672c\u5730\u6587\u4ef6<\/td>\n<td><code>JSONLoader<\/code><\/td>\n<td>\u4ece JSON \u4e2d\u63d0\u53d6\u6307\u5b9a\u5b57\u6bb5\u6216\u5bf9\u8c61<\/td>\n<td>API \u6570\u636e\u3001\u914d\u7f6e\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u672c\u5730\u6587\u4ef6<\/td>\n<td><code>PythonLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Python \u6e90\u4ee3\u7801<\/td>\n<td>\u4ee3\u7801\u95ee\u7b54\u3001\u4ee3\u7801\u68c0\u7d22<\/td>\n<\/tr>\n<tr>\n<td>\u901a\u7528\u6587\u4ef6<\/td>\n<td><code>UnstructuredFileLoader<\/code> \/ <code>UnstructuredLoader<\/code><\/td>\n<td>\u501f\u52a9 Unstructured \u89e3\u6790\u591a\u79cd\u975e\u7ed3\u6784\u5316\u6587\u4ef6<\/td>\n<td>\u6df7\u5408\u683c\u5f0f\u6587\u6863\u5e93<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PyPDFLoader<\/code><\/td>\n<td>\u4f7f\u7528 <code>pypdf<\/code> \u63d0\u53d6 PDF \u6587\u672c\uff0c\u901a\u5e38\u6309\u9875\u751f\u6210\u6587\u6863<\/td>\n<td>\u666e\u901a\u6587\u672c\u578b PDF<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PyMuPDFLoader<\/code><\/td>\n<td>\u4f7f\u7528 PyMuPDF \u8bfb\u53d6 PDF \u9875\u9762\u548c\u6587\u672c<\/td>\n<td>\u8ffd\u6c42\u901f\u5ea6\u548c\u517c\u5bb9\u6027\u7684 PDF \u5904\u7406<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PDFPlumberLoader<\/code><\/td>\n<td>\u4f7f\u7528 pdfplumber \u63d0\u53d6\u6587\u672c\u3001\u9875\u9762\u4f4d\u7f6e\u548c\u8868\u683c\u4fe1\u606f<\/td>\n<td>\u8868\u683c\u578b PDF\u3001\u7248\u9762\u5206\u6790<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PDFMinerLoader<\/code><\/td>\n<td>\u4f7f\u7528 PDFMiner \u8fdb\u884c\u6587\u672c\u63d0\u53d6<\/td>\n<td>\u9700\u8981 PDFMiner \u89e3\u6790\u80fd\u529b\u7684\u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>UnstructuredPDFLoader<\/code><\/td>\n<td>\u4f7f\u7528 Unstructured \u89e3\u6790 PDF \u7684\u6587\u672c\u548c\u7ed3\u6784\u5143\u7d20<\/td>\n<td>\u590d\u6742\u6392\u7248 PDF<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>DoclingLoader<\/code> \/ <code>Docling<\/code><\/td>\n<td>\u89e3\u6790\u6587\u6863\u7ed3\u6784\u3001\u8868\u683c\u3001\u6807\u9898\u548c\u7248\u9762<\/td>\n<td>\u590d\u6742 PDF\u3001\u4f01\u4e1a\u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PyMuPDF4LLMLoader<\/code><\/td>\n<td>\u4f7f\u7528\u9762\u5411 LLM \u7684 PyMuPDF \u5904\u7406 PDF<\/td>\n<td>\u9700\u8981 Markdown \u5316\u7684 PDF \u5185\u5bb9<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PDFParser<\/code><\/td>\n<td>\u901a\u8fc7\u5916\u90e8 PDF \u89e3\u6790\u670d\u52a1\u63d0\u53d6\u6587\u6863\u5185\u5bb9<\/td>\n<td>\u4e91\u7aef PDF \u89e3\u6790<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>MinerULoader<\/code><\/td>\n<td>\u4f7f\u7528 MinerU \u89e3\u6790 PDF \u548c\u5176\u4ed6\u590d\u6742\u6587\u6863<\/td>\n<td>\u5b66\u672f\u8bba\u6587\u3001\u590d\u6742\u7248\u9762<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PdfInspectorLoader<\/code><\/td>\n<td>\u68c0\u67e5\u548c\u63d0\u53d6 PDF \u5185\u7684\u6587\u672c\u3001\u5e03\u5c40\u6216\u5143\u7d20\u4fe1\u606f<\/td>\n<td>PDF \u8d28\u91cf\u68c0\u67e5\u3001\u7ed3\u6784\u5316\u63d0\u53d6<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>PdfmuseLoader<\/code><\/td>\n<td>\u63d0\u53d6 PDF\/DOCX \u7684\u6587\u672c\u3001\u8868\u683c\u3001\u5750\u6807\u548c\u533a\u5757\u5143\u6570\u636e<\/td>\n<td>\u9700\u8981\u7cbe\u786e\u5f15\u7528\u4f4d\u7f6e\u7684 RAG<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>OpenDataLoader PDF<\/code><\/td>\n<td>\u901a\u8fc7 OpenDataLoader \u89e3\u6790 PDF<\/td>\n<td>\u4f01\u4e1a\u6587\u6863\u89e3\u6790<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>OxidizePdfLoader<\/code><\/td>\n<td>\u4f7f\u7528 Rust \u5f15\u64ce\u89e3\u6790 PDF\uff0c\u5e76\u8f85\u52a9\u751f\u6210\u5206\u5757\u6570\u636e<\/td>\n<td>\u5927\u91cf PDF\u3001\u7a33\u5b9a\u89e3\u6790<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>CVFileLoader<\/code><\/td>\n<td>\u8bfb\u53d6\u5e26\u6709\u5d4c\u5165 Markdown\u3001HTML \u6216 JSON \u7684\u7b80\u5386 PDF<\/td>\n<td>\u7b80\u5386\u548c\u7ed3\u6784\u5316 PDF<\/td>\n<\/tr>\n<tr>\n<td>PDF<\/td>\n<td><code>HwpHwpxLoader<\/code><\/td>\n<td>\u8bfb\u53d6 HWP\/HWPX \u6587\u6863<\/td>\n<td>\u97e9\u6587\u529e\u516c\u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>PDF \/ \u6587\u6863<\/td>\n<td><code>Upstage Document Parse Loader<\/code><\/td>\n<td>\u4f7f\u7528 Upstage \u6587\u6863\u89e3\u6790\u670d\u52a1\u63d0\u53d6\u5185\u5bb9\u548c\u7ed3\u6784<\/td>\n<td>\u4f01\u4e1a\u7ea7\u590d\u6742\u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>PDF \/ \u6587\u6863<\/td>\n<td><code>UnDatasIO<\/code><\/td>\n<td>\u901a\u8fc7 UnDatasIO \u670d\u52a1\u89e3\u6790\u6587\u6863<\/td>\n<td>\u4e91\u7aef\u6587\u6863\u89e3\u6790<\/td>\n<\/tr>\n<tr>\n<td>PDF \/ \u6587\u6863<\/td>\n<td><code>PolarisAIDataInsightLoader<\/code><\/td>\n<td>\u4f7f\u7528 Polaris AI \u6587\u6863\u62bd\u53d6\u670d\u52a1\u63d0\u53d6\u7ed3\u6784\u5316\u5185\u5bb9<\/td>\n<td>Office\u3001PDF \u7b49\u590d\u6742\u6587\u4ef6<\/td>\n<\/tr>\n<tr>\n<td>Office<\/td>\n<td><code>Docx2txtLoader<\/code><\/td>\n<td>\u63d0\u53d6 Word \u6587\u6863\u4e2d\u7684\u6b63\u6587\u6587\u672c<\/td>\n<td>\u7b80\u5355 <code>.docx<\/code> \u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>Office<\/td>\n<td><code>UnstructuredWordDocumentLoader<\/code><\/td>\n<td>\u89e3\u6790 Word \u6587\u6863\u7684\u6bb5\u843d\u3001\u6807\u9898\u548c\u7ed3\u6784\u5143\u7d20<\/td>\n<td>\u4f01\u4e1a Word \u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>Office<\/td>\n<td><code>UnstructuredPowerPointLoader<\/code><\/td>\n<td>\u63d0\u53d6 PowerPoint \u5e7b\u706f\u7247\u4e2d\u7684\u6587\u672c\u548c\u7ed3\u6784<\/td>\n<td>\u57f9\u8bad\u6750\u6599\u3001\u6f14\u793a\u6587\u7a3f<\/td>\n<\/tr>\n<tr>\n<td>Office<\/td>\n<td><code>UnstructuredExcelLoader<\/code><\/td>\n<td>\u63d0\u53d6 Excel \u5de5\u4f5c\u7c3f\u3001\u5de5\u4f5c\u8868\u548c\u8868\u683c\u5185\u5bb9<\/td>\n<td>\u8868\u683c\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>WebBaseLoader<\/code><\/td>\n<td>\u52a0\u8f7d\u5e76\u89e3\u6790\u5355\u4e2a\u7f51\u9875<\/td>\n<td>\u6587\u6863\u9875\u9762\u3001\u535a\u5ba2\u3001\u5e2e\u52a9\u4e2d\u5fc3<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>RecursiveUrlLoader<\/code><\/td>\n<td>\u4ece\u5165\u53e3 URL \u9012\u5f52\u53d1\u73b0\u548c\u52a0\u8f7d\u7f51\u9875<\/td>\n<td>\u5c0f\u578b\u7f51\u7ad9\u3001\u6587\u6863\u7ad9<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>SitemapLoader<\/code><\/td>\n<td>\u6839\u636e Sitemap \u6279\u91cf\u52a0\u8f7d\u7f51\u9875<\/td>\n<td>\u7f51\u7ad9\u77e5\u8bc6\u5e93\u3001\u5e2e\u52a9\u4e2d\u5fc3<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>FirecrawlLoader<\/code><\/td>\n<td>\u901a\u8fc7 Firecrawl \u6293\u53d6\u3001\u6e05\u6d17\u548c\u8f6c\u6362\u7f51\u9875<\/td>\n<td>\u52a8\u6001\u7f51\u7ad9\u3001\u6574\u7ad9\u91c7\u96c6<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>ApifyCrawlLoader<\/code><\/td>\n<td>\u901a\u8fc7 Apify \u722c\u866b\u6293\u53d6\u7f51\u9875\u5e76\u8f6c\u6210\u6587\u6863<\/td>\n<td>\u5927\u89c4\u6a21\u7f51\u7ad9\u6293\u53d6<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>ApifyDatasetLoader<\/code><\/td>\n<td>\u4ece Apify \u6570\u636e\u96c6\u8bfb\u53d6\u5df2\u7ecf\u6293\u53d6\u7684\u6570\u636e<\/td>\n<td>\u590d\u7528\u722c\u866b\u7ed3\u679c<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>UnstructuredURLLoader<\/code><\/td>\n<td>\u4f7f\u7528 Unstructured \u89e3\u6790 URL \u5185\u5bb9<\/td>\n<td>\u901a\u7528\u7f51\u9875\u6587\u672c\u63d0\u53d6<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>DomPrunerLoader<\/code><\/td>\n<td>\u4fee\u526a\u7f51\u9875 DOM\uff0c\u53bb\u9664\u65e0\u5173\u5185\u5bb9\u5e76\u538b\u7f29\u6210 Markdown<\/td>\n<td>\u964d\u4f4e\u7f51\u9875\u566a\u58f0\u548c Token \u6d88\u8017<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>DiffbotCrawlLoader<\/code><\/td>\n<td>\u4f7f\u7528 Diffbot \u722c\u53d6\u7f51\u7ad9\u5e76\u63d0\u53d6\u7ed3\u6784\u5316\u5185\u5bb9<\/td>\n<td>\u65b0\u95fb\u3001\u5546\u54c1\u3001\u7f51\u7ad9\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>DiffbotExtractLoader<\/code><\/td>\n<td>\u4f7f\u7528 Diffbot \u4ece\u7f51\u9875\u4e2d\u63d0\u53d6\u7ed3\u6784\u5316\u5b9e\u4f53\u4fe1\u606f<\/td>\n<td>\u7f51\u9875\u4fe1\u606f\u62bd\u53d6<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>AgentQLLoader<\/code><\/td>\n<td>\u4f7f\u7528\u81ea\u7136\u8bed\u8a00\u6216\u67e5\u8be2\u4ece\u7f51\u9875\u4e2d\u63d0\u53d6\u7ed3\u6784\u5316\u6570\u636e<\/td>\n<td>\u52a8\u6001\u7f51\u9875\u6570\u636e\u63d0\u53d6<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>HyperbrowserLoader<\/code><\/td>\n<td>\u4f7f\u7528\u65e0\u5934\u6d4f\u89c8\u5668\u8bbf\u95ee\u548c\u6293\u53d6\u7f51\u9875<\/td>\n<td>\u9700\u8981\u6d4f\u89c8\u5668\u6267\u884c\u80fd\u529b\u7684\u7f51\u9875<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875<\/td>\n<td><code>PlasmateSOMLoader<\/code><\/td>\n<td>\u4ece\u6d4f\u89c8\u5668\u9875\u9762\u63d0\u53d6\u5e26\u6709 Set of Mark \u7684\u7ed3\u6784\u5316\u5185\u5bb9<\/td>\n<td>Agent \u6d4f\u89c8\u5668\u4efb\u52a1<\/td>\n<\/tr>\n<tr>\n<td>\u7f51\u9875 \/ \u4e91\u7aef<\/td>\n<td><code>HydrafetchLoader<\/code><\/td>\n<td>\u901a\u8fc7 Hydrafetch \u83b7\u53d6\u7f51\u9875\u6216\u8fdc\u7a0b\u5185\u5bb9<\/td>\n<td>\u8fdc\u7a0b\u5185\u5bb9\u91c7\u96c6<\/td>\n<\/tr>\n<tr>\n<td>\u4e91\u5b58\u50a8<\/td>\n<td><code>GoogleDriveLoader<\/code><\/td>\n<td>\u4ece Google Drive \u8bfb\u53d6 Google \u6587\u6863\u7b49\u5185\u5bb9<\/td>\n<td>\u56e2\u961f\u4e91\u7aef\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u4e91\u5b58\u50a8<\/td>\n<td><code>GCSFileLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Google Cloud Storage \u4e2d\u7684\u5355\u4e2a\u5bf9\u8c61<\/td>\n<td>\u4e91\u7aef\u6587\u4ef6\u5bfc\u5165<\/td>\n<\/tr>\n<tr>\n<td>\u4e91\u5b58\u50a8<\/td>\n<td><code>GCSDirectoryLoader<\/code><\/td>\n<td>\u6279\u91cf\u8bfb\u53d6 GCS Bucket \u4e2d\u7684\u6587\u4ef6<\/td>\n<td>\u4e91\u7aef\u6279\u91cf\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u4e91\u5b58\u50a8<\/td>\n<td><code>Azure Blob Storage Loader<\/code><\/td>\n<td>\u4ece Azure Blob Storage \u5bb9\u5668\u6216 Blob \u8bfb\u53d6\u6587\u6863<\/td>\n<td>Azure \u6587\u4ef6\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u4e91\u5b58\u50a8<\/td>\n<td><code>XNSBlobLoader<\/code><\/td>\n<td>\u4ece XNS Blob \u5b58\u50a8\u8bfb\u53d6\u5bf9\u8c61<\/td>\n<td>XNS \u4e91\u7aef\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u4e91\u5b58\u50a8<\/td>\n<td><code>langchain_box<\/code><\/td>\n<td>\u4ece Box \u4e91\u76d8\u8bfb\u53d6\u548c\u540c\u6b65\u6587\u4ef6<\/td>\n<td>\u4f01\u4e1a\u4e91\u76d8\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u534f\u4f5c\u5de5\u5177<\/td>\n<td><code>SlackLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Slack \u9891\u9053\u548c\u6d88\u606f<\/td>\n<td>\u56e2\u961f\u804a\u5929\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u534f\u4f5c\u5de5\u5177<\/td>\n<td><code>TrelloLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Trello \u770b\u677f\u3001\u5217\u8868\u548c\u5361\u7247<\/td>\n<td>\u9879\u76ee\u7ba1\u7406\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u534f\u4f5c\u5de5\u5177<\/td>\n<td><code>Notion<\/code> \u76f8\u5173 Loader<\/td>\n<td>\u8bfb\u53d6 Notion \u9875\u9762\u3001\u6570\u636e\u5e93\u548c\u5757\u5185\u5bb9<\/td>\n<td>\u56e2\u961f\u6587\u6863\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u534f\u4f5c\u5de5\u5177<\/td>\n<td><code>ConfluenceLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Confluence \u9875\u9762\u548c\u7a7a\u95f4\u5185\u5bb9<\/td>\n<td>\u4f01\u4e1a Wiki<\/td>\n<\/tr>\n<tr>\n<td>\u534f\u4f5c\u5de5\u5177<\/td>\n<td><code>GoogleClassroomLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Google Classroom \u8bfe\u7a0b\u8d44\u6599<\/td>\n<td>\u6559\u5b66\u8d44\u6599\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u534f\u4f5c\u5de5\u5177<\/td>\n<td><code>AgentMailLoader<\/code><\/td>\n<td>\u8bfb\u53d6 AgentMail \u90ae\u4ef6\u5185\u5bb9<\/td>\n<td>\u90ae\u4ef6\u77e5\u8bc6\u5e93\u3001\u90ae\u4ef6 Agent<\/td>\n<\/tr>\n<tr>\n<td>\u77e5\u8bc6\u5e93\u5e73\u53f0<\/td>\n<td><code>Outline<\/code> Loader<\/td>\n<td>\u8bfb\u53d6 Outline \u77e5\u8bc6\u5e93\u6587\u6863<\/td>\n<td>\u56e2\u961f Wiki<\/td>\n<\/tr>\n<tr>\n<td>\u77e5\u8bc6\u5e93\u5e73\u53f0<\/td>\n<td><code>Docugami<\/code> Loader<\/td>\n<td>\u4ece Docugami \u6587\u6863\u7ba1\u7406\u7cfb\u7edf\u8bfb\u53d6\u7ed3\u6784\u5316\u6587\u6863<\/td>\n<td>\u4f01\u4e1a\u6587\u6863\u7ba1\u7406<\/td>\n<\/tr>\n<tr>\n<td>\u77e5\u8bc6\u5e93\u5e73\u53f0<\/td>\n<td><code>LangSmithLoader<\/code><\/td>\n<td>\u8bfb\u53d6 LangSmith \u4e2d\u7684\u8ffd\u8e2a\u3001\u6570\u636e\u6216\u76f8\u5173\u8bb0\u5f55<\/td>\n<td>\u8c03\u8bd5\u6570\u636e\u3001\u8bc4\u4f30\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u4ee3\u7801\u4ed3\u5e93<\/td>\n<td><code>GitLoader<\/code><\/td>\n<td>\u4ece Git \u4ed3\u5e93\u52a0\u8f7d\u6307\u5b9a\u6587\u4ef6<\/td>\n<td>\u672c\u5730\u4ee3\u7801\u5e93\u95ee\u7b54<\/td>\n<\/tr>\n<tr>\n<td>\u4ee3\u7801\u4ed3\u5e93<\/td>\n<td><code>GithubFileLoader<\/code><\/td>\n<td>\u4ece GitHub \u4ed3\u5e93\u8bfb\u53d6\u6587\u4ef6<\/td>\n<td>GitHub \u4ee3\u7801\u68c0\u7d22<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>SnowflakeLoader<\/code><\/td>\n<td>\u4ece Snowflake \u6570\u636e\u4ed3\u5e93\u8bfb\u53d6\u6570\u636e<\/td>\n<td>\u4f01\u4e1a\u6570\u636e\u4ed3\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Google BigQuery<\/code> Loader<\/td>\n<td>\u4ece BigQuery \u67e5\u8be2\u6216\u8bfb\u53d6\u6570\u636e<\/td>\n<td>\u4e91\u7aef\u5206\u6790\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Google Cloud SQL<\/code> Loader<\/td>\n<td>\u4ece Cloud SQL \u7684 PostgreSQL\u3001MySQL \u6216 SQL Server \u8bfb\u53d6\u6570\u636e<\/td>\n<td>\u4e91\u6570\u636e\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Google AlloyDB for PostgreSQL<\/code> Loader<\/td>\n<td>\u4ece AlloyDB PostgreSQL \u6570\u636e\u5e93\u8bfb\u53d6\u6570\u636e<\/td>\n<td>\u4e91\u7aef PostgreSQL<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Google Spanner<\/code> Loader<\/td>\n<td>\u4ece Spanner \u5206\u5e03\u5f0f\u6570\u636e\u5e93\u8bfb\u53d6\u6570\u636e<\/td>\n<td>\u5927\u89c4\u6a21\u5173\u7cfb\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Google Firestore<\/code> Loader<\/td>\n<td>\u4ece Firestore Native \u6216 Datastore \u6a21\u5f0f\u8bfb\u53d6\u6587\u6863<\/td>\n<td>NoSQL \u6587\u6863\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Google Bigtable<\/code> Loader<\/td>\n<td>\u4ece Bigtable \u8bfb\u53d6\u5bbd\u5217\u6570\u636e<\/td>\n<td>\u5927\u89c4\u6a21\u952e\u503c\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Oracle Autonomous Database<\/code> Loader<\/td>\n<td>\u4ece Oracle Autonomous Database \u8bfb\u53d6\u6570\u636e<\/td>\n<td>Oracle \u4e91\u6570\u636e\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Google El Carro for Oracle<\/code> Loader<\/td>\n<td>\u8bfb\u53d6\u8fd0\u884c\u5728 Google Cloud \u4e0a\u7684 Oracle \u6570\u636e<\/td>\n<td>Oracle \u4e91\u8fc1\u79fb\u573a\u666f<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>Kinetica Document Loader<\/code><\/td>\n<td>\u4ece Kinetica \u6570\u636e\u5e73\u53f0\u8bfb\u53d6\u6587\u6863\u6216\u6570\u636e<\/td>\n<td>\u5b9e\u65f6\u5206\u6790\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>SingleStoreLoader<\/code><\/td>\n<td>\u4ece SingleStore \u6570\u636e\u5e93\u8bfb\u53d6\u6570\u636e<\/td>\n<td>\u5b9e\u65f6\u6570\u636e\u5e93\u3001\u6df7\u5408\u68c0\u7d22<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>AstraDB<\/code> Loader<\/td>\n<td>\u4ece Astra DB \u8bfb\u53d6\u6587\u6863\u6216\u5411\u91cf\u76f8\u5173\u6570\u636e<\/td>\n<td>Cassandra \u4e91\u670d\u52a1<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5e93<\/td>\n<td><code>PowerScaleDocumentLoader<\/code><\/td>\n<td>\u4ece Dell PowerScale \u6587\u4ef6\u548c\u6587\u6863\u5b58\u50a8\u8bfb\u53d6\u5185\u5bb9<\/td>\n<td>\u4f01\u4e1a\u6587\u4ef6\u7cfb\u7edf<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5904\u7406<\/td>\n<td><code>PySparkDataFrameLoader<\/code><\/td>\n<td>\u628a PySpark DataFrame \u7684\u884c\u8f6c\u6362\u6210 <code>Document<\/code><\/td>\n<td>\u5927\u6570\u636e\u5904\u7406\u7ed3\u679c<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5904\u7406<\/td>\n<td><code>AirbyteLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Airbyte \u540c\u6b65\u5230\u7684\u6570\u636e\u6e90<\/td>\n<td>\u591a\u6e90\u6570\u636e\u540c\u6b65<\/td>\n<\/tr>\n<tr>\n<td>\u6570\u636e\u5904\u7406<\/td>\n<td><code>Google Memorystore for Redis<\/code> Loader<\/td>\n<td>\u4ece Redis \u6570\u636e\u670d\u52a1\u8bfb\u53d6\u8bb0\u5f55<\/td>\n<td>\u7f13\u5b58\u6216\u952e\u503c\u6570\u636e<\/td>\n<\/tr>\n<tr>\n<td>\u97f3\u9891 \/ \u89c6\u9891<\/td>\n<td><code>YoutubeLoader<\/code> \/ <code>YoutubeLoaderDL<\/code><\/td>\n<td>\u83b7\u53d6 YouTube \u5b57\u5e55\u6216\u89c6\u9891\u8f6c\u5f55\u6587\u672c<\/td>\n<td>\u89c6\u9891\u95ee\u7b54\u3001\u8bfe\u7a0b\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u97f3\u9891 \/ \u89c6\u9891<\/td>\n<td><code>Google Speech-to-Text Audio Transcripts<\/code> Loader<\/td>\n<td>\u8bfb\u53d6\u8bed\u97f3\u8f6c\u6587\u5b57\u7ed3\u679c<\/td>\n<td>\u97f3\u9891\u8f6c\u5f55\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u97f3\u9891 \/ \u89c6\u9891<\/td>\n<td><code>Soniox<\/code> Loader<\/td>\n<td>\u4f7f\u7528 Soniox \u8bed\u97f3\u670d\u52a1\u83b7\u53d6\u8f6c\u5f55\u6587\u672c<\/td>\n<td>\u97f3\u9891\u3001\u4f1a\u8bae\u8bb0\u5f55<\/td>\n<\/tr>\n<tr>\n<td>OCR \/ \u89c6\u89c9<\/td>\n<td><code>PaddleOCR-VL<\/code> Loader<\/td>\n<td>\u4f7f\u7528 OCR \u548c\u89c6\u89c9\u6a21\u578b\u89e3\u6790\u6587\u5b57\u3001\u7248\u9762\u548c\u56fe\u7247<\/td>\n<td>\u626b\u63cf\u4ef6\u3001\u56fe\u6587\u6df7\u6392\u6587\u6863<\/td>\n<\/tr>\n<tr>\n<td>\u4f01\u4e1a\u6587\u4ef6<\/td>\n<td><code>AlfrescoLoader<\/code><\/td>\n<td>\u4ece Alfresco \u5185\u5bb9\u7ba1\u7406\u7cfb\u7edf\u8bfb\u53d6\u6587\u6863<\/td>\n<td>\u4f01\u4e1a\u5185\u5bb9\u7ba1\u7406<\/td>\n<\/tr>\n<tr>\n<td>\u4f01\u4e1a\u6587\u4ef6<\/td>\n<td><code>OpenDMALoader<\/code><\/td>\n<td>\u4ece OpenDMA \u6587\u6863\u7ba1\u7406\u7cfb\u7edf\u8bfb\u53d6\u5185\u5bb9<\/td>\n<td>\u4f01\u4e1a\u6587\u6863\u5f52\u6863<\/td>\n<\/tr>\n<tr>\n<td>\u4f01\u4e1a\u6587\u4ef6<\/td>\n<td><code>CapslaneLoader<\/code><\/td>\n<td>\u4ece Capslane \u5185\u5bb9\u670d\u52a1\u8bfb\u53d6\u6587\u6863<\/td>\n<td>\u4f01\u4e1a\u5185\u5bb9\u63a5\u5165<\/td>\n<\/tr>\n<tr>\n<td>\u4f01\u4e1a\u6587\u4ef6<\/td>\n<td><code>SecureLangChainLoader<\/code><\/td>\n<td>\u5728\u5b89\u5168\u63a7\u5236\u73af\u5883\u4e2d\u52a0\u8f7d\u5916\u90e8\u6587\u6863<\/td>\n<td>\u53d7\u63a7\u6570\u636e\u63a5\u5165<\/td>\n<\/tr>\n<tr>\n<td>\u4f01\u4e1a\u6587\u4ef6<\/td>\n<td><code>ReplyLayerLoader<\/code><\/td>\n<td>\u4ece ReplyLayer \u670d\u52a1\u8bfb\u53d6\u5185\u5bb9<\/td>\n<td>\u4f01\u4e1a\u6d88\u606f\u6216\u5185\u5bb9\u63a5\u5165<\/td>\n<\/tr>\n<tr>\n<td>\u4f01\u4e1a\u6587\u4ef6<\/td>\n<td><code>OpeddFeedLoader<\/code><\/td>\n<td>\u8bfb\u53d6\u5e26\u6388\u6743\u548c\u6765\u6e90\u4fe1\u606f\u7684 Opedd \u5185\u5bb9\u76ee\u5f55<\/td>\n<td>\u5408\u89c4\u5185\u5bb9\u63a5\u5165<\/td>\n<\/tr>\n<tr>\n<td>\u4f01\u4e1a\u6587\u4ef6<\/td>\n<td><code>EnconvertLoader<\/code><\/td>\n<td>\u901a\u8fc7 Enconvert \u670d\u52a1\u8f6c\u6362\u5e76\u8bfb\u53d6\u6587\u4ef6<\/td>\n<td>\u683c\u5f0f\u8f6c\u6362\u3001\u6587\u6863\u5bfc\u5165<\/td>\n<\/tr>\n<tr>\n<td>\u5176\u4ed6<\/td>\n<td><code>GutenbergLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Project Gutenberg \u7535\u5b50\u4e66<\/td>\n<td>\u516c\u7248\u4e66\u7c4d\u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u5176\u4ed6<\/td>\n<td><code>ConLLULoader<\/code><\/td>\n<td>\u8bfb\u53d6 CoNLL-U \u683c\u5f0f\u7684\u8bed\u8a00\u5b66\u6807\u6ce8\u6570\u636e<\/td>\n<td>NLP \u8bad\u7ec3\u548c\u5206\u6790<\/td>\n<\/tr>\n<tr>\n<td>\u5176\u4ed6<\/td>\n<td><code>NucliaLoader<\/code><\/td>\n<td>\u4ece Nuclia \u77e5\u8bc6\u5e73\u53f0\u8bfb\u53d6\u5185\u5bb9<\/td>\n<td>Nuclia \u77e5\u8bc6\u5e93<\/td>\n<\/tr>\n<tr>\n<td>\u5176\u4ed6<\/td>\n<td><code>IuguLoader<\/code><\/td>\n<td>\u4ece Iugu \u670d\u52a1\u8bfb\u53d6\u76f8\u5173\u4e1a\u52a1\u8bb0\u5f55<\/td>\n<td>Iugu \u6570\u636e\u63a5\u5165<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<blockquote>\n<p>\u6ce8\u610f\uff1a\u5b98\u65b9\u201c\u5168\u90e8\u52a0\u8f7d\u5668\u201d\u76ee\u5f55\u5305\u542b\u5927\u91cf\u793e\u533a\u548c\u5408\u4f5c\u65b9\u96c6\u6210\uff0c\u540d\u79f0\u548c\u6570\u91cf\u4f1a\u968f\u7248\u672c\u53d8\u5316\uff1b\u6709\u4e9b\u52a0\u8f7d\u5668\u5e76\u4e0d\u5728\u540c\u4e00\u4e2a Python \u5305\u4e2d\uff0c\u4e5f\u53ef\u80fd\u9700\u8981\u72ec\u7acb\u5b89\u88c5\u5bf9\u5e94\u96c6\u6210\u5305\u3002\u5b98\u65b9\u5f53\u524d\u5206\u7c7b\u548c\u5b8c\u6574\u76ee\u5f55\u53ef\u53c2\u8003 <a href=\"https:\/\/docs.langchain.com\/oss\/python\/integrations\/document_loaders\">Document loader integrations<\/a>\u3002<\/p>\n<\/blockquote>\n<h2>1. \u4ec0\u4e48\u662f\u6570\u636e\u52a0\u8f7d\u5668<\/h2>\n<p>\u6570\u636e\u52a0\u8f7d\u5668\u8d1f\u8d23\u4ece\u4e0d\u540c\u6765\u6e90\u8bfb\u53d6\u6570\u636e\uff0c\u5e76\u628a\u6570\u636e\u8f6c\u6362\u6210 LangChain \u7edf\u4e00\u7684 <code>Document<\/code> \u5bf9\u8c61\u3002<\/p>\n<pre><code class=\"language-text\">PDF \/ \u7f51\u9875 \/ TXT \/ \u6570\u636e\u5e93 \/ \u4e91\u76d8\n              \u2193\n          Document Loader\n              \u2193\nDocument(page_content, metadata)\n<\/code><\/pre>\n<p>\u4e00\u4e2a <code>Document<\/code> \u901a\u5e38\u5305\u542b\u4e24\u4e2a\u90e8\u5206\uff1a<\/p>\n<pre><code class=\"language-python\">Document(\n    page_content=&quot;\u8fd9\u662f\u6587\u6863\u7684\u6b63\u6587\u5185\u5bb9&quot;,\n    metadata={\n        &quot;source&quot;: &quot;manual.pdf&quot;,\n        &quot;page&quot;: 3\n    }\n)\n<\/code><\/pre>\n<ul>\n<li><code>page_content<\/code>\uff1a\u6b63\u6587\u5185\u5bb9\u3002<\/li>\n<li><code>metadata<\/code>\uff1a\u6765\u6e90\u3001\u9875\u7801\u3001\u6807\u9898\u3001\u521b\u5efa\u65f6\u95f4\u7b49\u8f85\u52a9\u4fe1\u606f\u3002<\/li>\n<\/ul>\n<p>\u8fd9\u4e9b\u4fe1\u606f\u540e\u7eed\u53ef\u4ee5\u4ea4\u7ed9\u6587\u672c\u5207\u5206\u5668\u3001Embedding \u6a21\u578b\u548c\u5411\u91cf\u6570\u636e\u5e93\uff0c\u7528\u4e8e\u6784\u5efa RAG \u77e5\u8bc6\u5e93\u3002<\/p>\n<h2>2. \u52a0\u8f7d\u5668\u7684\u7edf\u4e00\u63a5\u53e3<\/h2>\n<p>LangChain \u7684\u52a0\u8f7d\u5668\u901a\u5e38\u5b9e\u73b0 <code>BaseLoader<\/code> \u63a5\u53e3\u3002\u6700\u5e38\u89c1\u7684\u65b9\u6cd5\u6709\u4e24\u4e2a\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u65b9\u6cd5<\/th>\n<th>\u4f5c\u7528<\/th>\n<th>\u9002\u5408\u573a\u666f<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>load()<\/code><\/td>\n<td>\u4e00\u6b21\u6027\u52a0\u8f7d\u5168\u90e8\u6587\u6863<\/td>\n<td>\u6587\u4ef6\u8f83\u5c0f\u3001\u4ee3\u7801\u7b80\u5355<\/td>\n<\/tr>\n<tr>\n<td><code>lazy_load()<\/code><\/td>\n<td>\u9010\u4e2a\u6216\u5206\u6279\u8fd4\u56de\u6587\u6863<\/td>\n<td>\u5927\u6587\u4ef6\u3001\u5927\u89c4\u6a21\u6570\u636e<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h3>\u4f7f\u7528 <code>load()<\/code><\/h3>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import TextLoader\n\nloader = TextLoader(&quot;notes.txt&quot;, encoding=&quot;utf-8&quot;)\ndocuments = loader.load()\n\nfor document in documents:\n    print(document.page_content)\n    print(document.metadata)\n<\/code><\/pre>\n<h3>\u4f7f\u7528 <code>lazy_load()<\/code><\/h3>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import TextLoader\n\nloader = TextLoader(&quot;large-notes.txt&quot;, encoding=&quot;utf-8&quot;)\n\nfor document in loader.lazy_load():\n    print(document.page_content[:100])\n<\/code><\/pre>\n<h2>3. \u5e38\u89c1\u7684\u6570\u636e\u52a0\u8f7d\u5668\u5206\u7c7b<\/h2>\n<p>LangChain \u5b98\u65b9\u6587\u6863\u4e2d\u7684\u52a0\u8f7d\u5668\u6570\u91cf\u5f88\u591a\uff0c\u800c\u4e14\u4f1a\u968f\u7740\u7248\u672c\u548c\u96c6\u6210\u751f\u6001\u53d8\u5316\u3002\u5b66\u4e60\u65f6\u4e0d\u9700\u8981\u628a\u6240\u6709\u7c7b\u540d\u90fd\u80cc\u4e0b\u6765\uff0c\u66f4\u91cd\u8981\u7684\u662f\u6839\u636e\u6570\u636e\u6e90\u9009\u62e9\u5408\u9002\u7684\u7c7b\u522b\u3002<\/p>\n<h3>3.1 \u672c\u5730\u6587\u672c\u6587\u4ef6<\/h3>\n<table>\n<thead>\n<tr>\n<th>Loader<\/th>\n<th>\u7528\u9014<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>TextLoader<\/code><\/td>\n<td>TXT\u3001Markdown \u7b49\u7eaf\u6587\u672c\u6587\u4ef6<\/td>\n<\/tr>\n<tr>\n<td><code>DirectoryLoader<\/code><\/td>\n<td>\u6279\u91cf\u52a0\u8f7d\u4e00\u4e2a\u76ee\u5f55\u4e2d\u7684\u6587\u4ef6<\/td>\n<\/tr>\n<tr>\n<td><code>CSVLoader<\/code><\/td>\n<td>CSV \u8868\u683c\u6587\u4ef6<\/td>\n<\/tr>\n<tr>\n<td><code>JSONLoader<\/code><\/td>\n<td>JSON \u6587\u4ef6<\/td>\n<\/tr>\n<tr>\n<td><code>PythonLoader<\/code><\/td>\n<td>Python \u6e90\u4ee3\u7801\u6587\u4ef6<\/td>\n<\/tr>\n<tr>\n<td><code>UnstructuredFileLoader<\/code><\/td>\n<td>\u901a\u8fc7 Unstructured \u5904\u7406\u591a\u79cd\u6587\u4ef6\u683c\u5f0f<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h4>\u52a0\u8f7d TXT \u6587\u4ef6<\/h4>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import TextLoader\n\nloader = TextLoader(&quot;data\/intro.txt&quot;, encoding=&quot;utf-8&quot;)\ndocuments = loader.load()\n<\/code><\/pre>\n<h4>\u6279\u91cf\u52a0\u8f7d\u76ee\u5f55<\/h4>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import DirectoryLoader\nfrom langchain_community.document_loaders import TextLoader\n\nloader = DirectoryLoader(\n    &quot;data\/&quot;,\n    glob=&quot;**\/*.txt&quot;,\n    loader_cls=TextLoader,\n    loader_kwargs={&quot;encoding&quot;: &quot;utf-8&quot;}\n)\n\ndocuments = loader.load()\n<\/code><\/pre>\n<p><code>DirectoryLoader<\/code> \u672c\u8eab\u901a\u5e38\u4e0d\u8d1f\u8d23\u89e3\u6790\u6240\u6709\u683c\u5f0f\uff0c\u5b83\u66f4\u50cf\u4e00\u4e2a\u6279\u5904\u7406\u5668\uff1a\u904d\u5386\u76ee\u5f55\uff0c\u7136\u540e\u628a\u6bcf\u4e2a\u6587\u4ef6\u4ea4\u7ed9\u6307\u5b9a\u7684 <code>loader_cls<\/code>\u3002<\/p>\n<h3>3.2 PDF \u6587\u4ef6<\/h3>\n<p>PDF \u662f RAG \u4e2d\u6700\u5e38\u89c1\u3001\u4e5f\u6700\u5bb9\u6613\u51fa\u95ee\u9898\u7684\u6570\u636e\u6e90\u4e4b\u4e00\u3002<\/p>\n<table>\n<thead>\n<tr>\n<th>Loader<\/th>\n<th>\u7279\u70b9<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>PyPDFLoader<\/code><\/td>\n<td>\u4f7f\u7528 <code>pypdf<\/code>\uff0c\u9002\u5408\u666e\u901a\u6587\u672c\u578b PDF<\/td>\n<\/tr>\n<tr>\n<td><code>PyMuPDFLoader<\/code><\/td>\n<td>\u4f7f\u7528 PyMuPDF\uff0c\u901f\u5ea6\u548c\u517c\u5bb9\u6027\u8f83\u597d<\/td>\n<\/tr>\n<tr>\n<td><code>PDFPlumberLoader<\/code><\/td>\n<td>\u5bf9\u6587\u672c\u5b9a\u4f4d\u548c\u8868\u683c\u5904\u7406\u66f4\u53cb\u597d<\/td>\n<\/tr>\n<tr>\n<td><code>PDFMinerLoader<\/code><\/td>\n<td>\u4f7f\u7528 PDFMiner \u63d0\u53d6\u6587\u672c<\/td>\n<\/tr>\n<tr>\n<td><code>UnstructuredPDFLoader<\/code><\/td>\n<td>\u9002\u5408\u590d\u6742\u6587\u6863\u7ed3\u6784<\/td>\n<\/tr>\n<tr>\n<td><code>DoclingLoader<\/code><\/td>\n<td>\u9002\u5408\u590d\u6742\u6392\u7248\u3001\u8868\u683c\u548c\u7ed3\u6784\u5316\u89e3\u6790<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h4>\u4f7f\u7528 <code>PyPDFLoader<\/code><\/h4>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import PyPDFLoader\n\nloader = PyPDFLoader(&quot;manual.pdf&quot;)\ndocuments = loader.load()\n\nprint(len(documents))\nprint(documents[0].metadata)\n<\/code><\/pre>\n<p>\u666e\u901a PDF \u901a\u5e38\u4f1a\u6309\u9875\u751f\u6210\u591a\u4e2a <code>Document<\/code>\uff0c\u5143\u6570\u636e\u4e2d\u53ef\u80fd\u5305\u542b\u9875\u7801\u548c\u6587\u4ef6\u6765\u6e90\u3002<\/p>\n<h4>\u5982\u4f55\u9009\u62e9 PDF Loader<\/h4>\n<ul>\n<li>PDF \u4e3b\u8981\u662f\u8fde\u7eed\u6587\u5b57\uff1a\u5148\u5c1d\u8bd5 <code>PyPDFLoader<\/code>\u3002<\/li>\n<li>\u9700\u8981\u66f4\u597d\u7684\u901f\u5ea6\u6216\u9875\u9762\u5904\u7406\uff1a\u5c1d\u8bd5 <code>PyMuPDFLoader<\/code>\u3002<\/li>\n<li>\u5305\u542b\u8868\u683c\u3001\u590d\u6742\u5e03\u5c40\u3001\u591a\u680f\u6392\u7248\uff1a\u8003\u8651 <code>PDFPlumberLoader<\/code>\u3001<code>Unstructured<\/code> \u6216 <code>Docling<\/code>\u3002<\/li>\n<li>PDF \u662f\u626b\u63cf\u56fe\u7247\uff1a\u666e\u901a\u6587\u672c Loader \u901a\u5e38\u65e0\u6cd5\u8bfb\u53d6\uff0c\u9700\u8981 OCR \u6216\u652f\u6301\u89c6\u89c9\u89e3\u6790\u7684\u65b9\u6848\u3002<\/li>\n<\/ul>\n<h3>3.3 Word\u3001PowerPoint \u548c Excel<\/h3>\n<p>\u5e38\u89c1\u7684 Office \u6587\u6863\u52a0\u8f7d\u65b9\u5f0f\u5305\u62ec\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>Loader<\/th>\n<th>\u7528\u9014<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>Docx2txtLoader<\/code><\/td>\n<td>\u63d0\u53d6 Word \u6587\u6863\u6b63\u6587<\/td>\n<\/tr>\n<tr>\n<td><code>UnstructuredWordDocumentLoader<\/code><\/td>\n<td>\u89e3\u6790 Word \u7ed3\u6784<\/td>\n<\/tr>\n<tr>\n<td><code>UnstructuredPowerPointLoader<\/code><\/td>\n<td>\u63d0\u53d6 PowerPoint \u5185\u5bb9<\/td>\n<\/tr>\n<tr>\n<td><code>UnstructuredExcelLoader<\/code><\/td>\n<td>\u8bfb\u53d6 Excel \u5185\u5bb9<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u793a\u4f8b\uff1a<\/p>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import Docx2txtLoader\n\nloader = Docx2txtLoader(&quot;report.docx&quot;)\ndocuments = loader.load()\n<\/code><\/pre>\n<p>\u5bf9\u4e8e Office \u6587\u6863\uff0c\u91cd\u70b9\u4e0d\u662f\u201c\u80fd\u4e0d\u80fd\u52a0\u8f7d\u201d\uff0c\u800c\u662f\u662f\u5426\u80fd\u4fdd\u7559\u6807\u9898\u3001\u6bb5\u843d\u3001\u8868\u683c\u548c\u9875\u9762\u7ed3\u6784\u3002\u7ed3\u6784\u4e22\u5931\u540e\uff0c\u5373\u4f7f\u540e\u9762\u7684\u5411\u91cf\u68c0\u7d22\u6b63\u5e38\uff0c\u56de\u7b54\u8d28\u91cf\u4e5f\u53ef\u80fd\u4e0b\u964d\u3002<\/p>\n<h3>3.4 \u7f51\u9875\u548c\u7f51\u7ad9<\/h3>\n<table>\n<thead>\n<tr>\n<th>Loader<\/th>\n<th>\u7528\u9014<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>WebBaseLoader<\/code><\/td>\n<td>\u52a0\u8f7d\u5355\u4e2a\u7f51\u9875<\/td>\n<\/tr>\n<tr>\n<td><code>RecursiveUrlLoader<\/code><\/td>\n<td>\u4ece\u5165\u53e3\u5730\u5740\u9012\u5f52\u6293\u53d6\u7f51\u9875<\/td>\n<\/tr>\n<tr>\n<td><code>SitemapLoader<\/code><\/td>\n<td>\u6839\u636e\u7f51\u7ad9 Sitemap \u6279\u91cf\u52a0\u8f7d\u9875\u9762<\/td>\n<\/tr>\n<tr>\n<td><code>FirecrawlLoader<\/code><\/td>\n<td>\u4f7f\u7528 Firecrawl \u6293\u53d6\u548c\u6e05\u6d17\u7f51\u9875<\/td>\n<\/tr>\n<tr>\n<td><code>ApifyCrawlLoader<\/code><\/td>\n<td>\u4f7f\u7528 Apify \u6293\u53d6\u7f51\u7ad9\u5185\u5bb9<\/td>\n<\/tr>\n<tr>\n<td><code>UnstructuredURLLoader<\/code><\/td>\n<td>\u4f7f\u7528 Unstructured \u89e3\u6790\u7f51\u9875<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h4>\u52a0\u8f7d\u5355\u4e2a\u7f51\u9875<\/h4>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import WebBaseLoader\n\nloader = WebBaseLoader(&quot;https:\/\/example.com\/docs&quot;)\ndocuments = loader.load()\n<\/code><\/pre>\n<p>\u7f51\u9875\u52a0\u8f7d\u901a\u5e38\u8fd8\u9700\u8981\u5904\u7406\uff1a<\/p>\n<ul>\n<li>\u5bfc\u822a\u680f\u3001\u9875\u811a\u548c\u5e7f\u544a\u5185\u5bb9\u3002<\/li>\n<li>\u52a8\u6001 JavaScript \u9875\u9762\u3002<\/li>\n<li>\u8bbf\u95ee\u9891\u7387\u9650\u5236\u548c robots.txt\u3002<\/li>\n<li>\u7f51\u9875\u66f4\u65b0\u540e\u7684\u589e\u91cf\u540c\u6b65\u3002<\/li>\n<li>\u9875\u9762\u6807\u9898\u3001URL \u548c\u53d1\u5e03\u65f6\u95f4\u7b49\u5143\u6570\u636e\u3002<\/li>\n<\/ul>\n<p>\u5982\u679c\u53ea\u662f\u5b66\u4e60\uff0c\u53ef\u4ee5\u4ece <code>WebBaseLoader<\/code> \u5f00\u59cb\uff1b\u5982\u679c\u8981\u6293\u53d6\u4e00\u4e2a\u5b8c\u6574\u7f51\u7ad9\uff0c\u901a\u5e38\u9700\u8981\u4e13\u95e8\u7684\u722c\u866b\u6216\u7f51\u9875\u89e3\u6790\u670d\u52a1\u3002<\/p>\n<h3>3.5 \u4e91\u5b58\u50a8\u548c\u751f\u4ea7\u529b\u5de5\u5177<\/h3>\n<p>LangChain \u4e5f\u63d0\u4f9b\u4e86\u8fde\u63a5\u4e91\u5b58\u50a8\u548c\u534f\u4f5c\u5de5\u5177\u7684\u52a0\u8f7d\u5668\uff0c\u4f8b\u5982\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u6570\u636e\u6765\u6e90<\/th>\n<th>\u5e38\u89c1 Loader<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Google Drive<\/td>\n<td><code>GoogleDriveLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>Google Cloud Storage<\/td>\n<td><code>GCSFileLoader<\/code>\u3001<code>GCSDirectoryLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>Slack<\/td>\n<td><code>SlackLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>Trello<\/td>\n<td><code>TrelloLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>Notion<\/td>\n<td>Notion \u76f8\u5173 Loader<\/td>\n<\/tr>\n<tr>\n<td>Confluence<\/td>\n<td><code>ConfluenceLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>GitHub<\/td>\n<td><code>GithubFileLoader<\/code><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u8fd9\u7c7b Loader \u901a\u5e38\u9700\u8981\uff1a<\/p>\n<ul>\n<li>API Key \u6216 OAuth \u6388\u6743\u3002<\/li>\n<li>\u6307\u5b9a\u6587\u4ef6\u5939\u3001\u9891\u9053\u3001\u6570\u636e\u5e93\u6216\u9875\u9762\u8303\u56f4\u3002<\/li>\n<li>\u5904\u7406\u8bbf\u95ee\u6743\u9650\u548c\u7528\u6237\u9690\u79c1\u3002<\/li>\n<li>\u4fdd\u5b58\u6765\u6e90 ID \u548c\u66f4\u65b0\u65f6\u95f4\uff0c\u65b9\u4fbf\u540e\u7eed\u589e\u91cf\u66f4\u65b0\u3002<\/li>\n<\/ul>\n<h3>3.6 \u4ee3\u7801\u4ed3\u5e93<\/h3>\n<p>\u4ee3\u7801\u4e5f\u53ef\u4ee5\u4f5c\u4e3a\u77e5\u8bc6\u5e93\u6570\u636e\uff0c\u4f8b\u5982\u7528\u4e8e\u6784\u5efa\u4ee3\u7801\u95ee\u7b54\u52a9\u624b\u3002<\/p>\n<p>\u5e38\u89c1\u52a0\u8f7d\u65b9\u5f0f\u5305\u62ec\uff1a<\/p>\n<ul>\n<li><code>GitLoader<\/code>\uff1a\u52a0\u8f7d Git \u4ed3\u5e93\u4e2d\u7684\u6587\u4ef6\u3002<\/li>\n<li><code>GithubFileLoader<\/code>\uff1a\u4ece GitHub \u83b7\u53d6\u6587\u4ef6\u3002<\/li>\n<li><code>PythonLoader<\/code>\uff1a\u52a0\u8f7d Python \u6e90\u4ee3\u7801\u3002<\/li>\n<li><code>DirectoryLoader<\/code>\uff1a\u6279\u91cf\u52a0\u8f7d\u6307\u5b9a\u6269\u5c55\u540d\u7684\u4ee3\u7801\u6587\u4ef6\u3002<\/li>\n<\/ul>\n<p>\u4ee3\u7801\u52a0\u8f7d\u65f6\u901a\u5e38\u9700\u8981\u8fc7\u6ee4\uff1a<\/p>\n<ul>\n<li><code>.git<\/code> \u76ee\u5f55\u3002<\/li>\n<li>\u6784\u5efa\u4ea7\u7269\u548c\u4f9d\u8d56\u76ee\u5f55\u3002<\/li>\n<li>\u4e8c\u8fdb\u5236\u6587\u4ef6\u3002<\/li>\n<li>\u5bc6\u94a5\u3001\u914d\u7f6e\u6587\u4ef6\u548c\u73af\u5883\u53d8\u91cf\u6587\u4ef6\u3002<\/li>\n<\/ul>\n<h3>3.7 \u6570\u636e\u5e93\u548c\u6570\u636e\u5e73\u53f0<\/h3>\n<p>LangChain \u751f\u6001\u4e2d\u4e5f\u6709\u9762\u5411\u6570\u636e\u5e93\u3001\u6570\u636e\u4ed3\u5e93\u548c DataFrame \u7684\u52a0\u8f7d\u96c6\u6210\uff0c\u4f8b\u5982\uff1a<\/p>\n<ul>\n<li>Snowflake \u76f8\u5173 Loader\u3002<\/li>\n<li>BigQuery \u76f8\u5173\u96c6\u6210\u3002<\/li>\n<li><code>PySparkDataFrameLoader<\/code>\u3002<\/li>\n<li>\u5176\u4ed6\u6570\u636e\u5e93\u6216\u4e91\u6570\u636e\u5e73\u53f0 Loader\u3002<\/li>\n<\/ul>\n<p>\u4e0d\u8fc7\uff0c\u6570\u636e\u5e93\u95ee\u7b54\u4e0d\u4e00\u5b9a\u8981\u628a\u6240\u6709\u6570\u636e\u8f6c\u6362\u6210\u5411\u91cf\u3002\u5bf9\u4e8e\u7ed3\u6784\u5316\u6570\u636e\uff0c\u76f4\u63a5\u751f\u6210 SQL \u67e5\u8be2\u5f80\u5f80\u66f4\u5408\u9002\uff1b\u53ea\u6709\u5f53\u6570\u636e\u5e93\u5b57\u6bb5\u4e2d\u5305\u542b\u5927\u91cf\u6587\u672c\u65f6\uff0c\u624d\u9002\u5408\u628a\u6587\u672c\u5b57\u6bb5\u52a0\u8f7d\u3001\u5207\u5206\u5e76\u5efa\u7acb\u5411\u91cf\u7d22\u5f15\u3002<\/p>\n<h3>3.8 \u97f3\u9891\u548c\u89c6\u9891<\/h3>\n<p>\u5e38\u89c1\u573a\u666f\u662f\u628a\u89c6\u9891\u5b57\u5e55\u6216\u97f3\u9891\u8f6c\u5f55\u7ed3\u679c\u52a0\u8f7d\u4e3a\u6587\u6863\uff0c\u4f8b\u5982\uff1a<\/p>\n<ul>\n<li><code>YoutubeLoader<\/code>\uff1a\u8bfb\u53d6 YouTube \u5b57\u5e55\u3002<\/li>\n<li>\u8bed\u97f3\u8f6c\u6587\u5b57\u670d\u52a1 Loader\uff1a\u5148\u8f6c\u5f55\u97f3\u9891\uff0c\u518d\u751f\u6210 <code>Document<\/code>\u3002<\/li>\n<li>\u89c6\u9891\u5185\u5bb9\u89e3\u6790\u670d\u52a1\uff1a\u63d0\u53d6\u5b57\u5e55\u3001\u753b\u9762\u63cf\u8ff0\u548c\u65f6\u95f4\u6233\u3002<\/li>\n<\/ul>\n<p>\u8fd9\u7c7b\u6570\u636e\u6700\u597d\u4fdd\u7559\u65f6\u95f4\u4fe1\u606f\uff0c\u4f8b\u5982\uff1a<\/p>\n<pre><code class=\"language-python\">{\n    &quot;source&quot;: &quot;video.mp4&quot;,\n    &quot;start_time&quot;: 120,\n    &quot;end_time&quot;: 160\n}\n<\/code><\/pre>\n<p>\u8fd9\u6837\u56de\u7b54\u65f6\u53ef\u4ee5\u5b9a\u4f4d\u5230\u89c6\u9891\u7684\u5177\u4f53\u65f6\u95f4\u6bb5\u3002<\/p>\n<h2>4. Loader\u3001Parser \u548c Splitter \u7684\u533a\u522b<\/h2>\n<p>\u8fd9\u51e0\u4e2a\u6982\u5ff5\u5bb9\u6613\u6df7\u6dc6\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u7ec4\u4ef6<\/th>\n<th>\u4e3b\u8981\u804c\u8d23<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Loader<\/td>\n<td>\u4ece\u6570\u636e\u6e90\u8bfb\u53d6\u5185\u5bb9<\/td>\n<\/tr>\n<tr>\n<td>Parser<\/td>\n<td>\u4ece\u590d\u6742\u683c\u5f0f\u4e2d\u63d0\u53d6\u6b63\u6587\u548c\u7ed3\u6784<\/td>\n<\/tr>\n<tr>\n<td>Splitter<\/td>\n<td>\u628a\u957f\u6587\u672c\u5207\u6210\u9002\u5408\u68c0\u7d22\u7684\u7247\u6bb5<\/td>\n<\/tr>\n<tr>\n<td>Embedding<\/td>\n<td>\u628a\u6587\u672c\u8f6c\u6362\u6210\u5411\u91cf<\/td>\n<\/tr>\n<tr>\n<td>Vector Store<\/td>\n<td>\u4fdd\u5b58\u5411\u91cf\u5e76\u6267\u884c\u76f8\u4f3c\u5ea6\u68c0\u7d22<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u5178\u578b\u6d41\u7a0b\u5982\u4e0b\uff1a<\/p>\n<pre><code class=\"language-text\">Loader\n  \u2193\nParser \/ Cleaner\n  \u2193\nText Splitter\n  \u2193\nEmbedding\n  \u2193\nVector Store\n<\/code><\/pre>\n<p>\u6709\u4e9b\u96c6\u6210\u4f1a\u628a Loader \u548c Parser \u5408\u5e76\uff0c\u56e0\u6b64\u4e00\u4e2a Loader \u53ef\u80fd\u540c\u65f6\u5b8c\u6210\u8bfb\u53d6\u548c\u89e3\u6790\u5de5\u4f5c\u3002<\/p>\n<h2>5. \u5982\u4f55\u9009\u62e9\u6570\u636e\u52a0\u8f7d\u5668<\/h2>\n<p>\u53ef\u4ee5\u4f7f\u7528\u4e0b\u9762\u7684\u51b3\u7b56\u65b9\u5f0f\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u4f60\u7684\u6570\u636e<\/th>\n<th>\u63a8\u8350\u8d77\u70b9<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>TXT\u3001Markdown<\/td>\n<td><code>TextLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>\u4e00\u6279\u672c\u5730\u6587\u4ef6<\/td>\n<td><code>DirectoryLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>\u666e\u901a PDF<\/td>\n<td><code>PyPDFLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>\u8868\u683c\u578b PDF<\/td>\n<td><code>PDFPlumberLoader<\/code> \u6216\u4e13\u95e8\u7684\u6587\u6863\u89e3\u6790\u5668<\/td>\n<\/tr>\n<tr>\n<td>\u626b\u63cf PDF<\/td>\n<td>OCR \u6216\u89c6\u89c9\u6587\u6863\u89e3\u6790\u5668<\/td>\n<\/tr>\n<tr>\n<td>Word\u3001PPT\u3001\u590d\u6742 Office<\/td>\n<td><code>Unstructured<\/code> \u6216 <code>Docling<\/code><\/td>\n<\/tr>\n<tr>\n<td>\u5355\u4e2a\u7f51\u9875<\/td>\n<td><code>WebBaseLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>\u6574\u4e2a\u7f51\u7ad9<\/td>\n<td>Sitemap\u3001\u722c\u866b\u6216\u7f51\u9875\u6293\u53d6\u670d\u52a1<\/td>\n<\/tr>\n<tr>\n<td>GitHub \u4ee3\u7801<\/td>\n<td><code>GithubFileLoader<\/code><\/td>\n<\/tr>\n<tr>\n<td>\u7ed3\u6784\u5316\u6570\u636e\u5e93<\/td>\n<td>\u4f18\u5148\u8003\u8651 SQL \u67e5\u8be2\uff0c\u800c\u4e0d\u662f\u76f4\u63a5\u5411\u91cf\u5316<\/td>\n<\/tr>\n<tr>\n<td>\u89c6\u9891\u5b57\u5e55<\/td>\n<td><code>YoutubeLoader<\/code> \u6216\u8f6c\u5f55\u670d\u52a1<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u5b66\u4e60\u9636\u6bb5\u5efa\u8bae\u5148\u638c\u63e1\u4ee5\u4e0b 6 \u4e2a\uff1a<\/p>\n<pre><code class=\"language-text\">TextLoader\nDirectoryLoader\nPyPDFLoader\nWebBaseLoader\nCSVLoader\nDoclingLoader\n<\/code><\/pre>\n<h2>6. \u5305\u548c\u7248\u672c\u95ee\u9898<\/h2>\n<p>\u8bb8\u591a\u5e38\u7528 Loader \u7684\u5bfc\u5165\u8def\u5f84\u662f\uff1a<\/p>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import TextLoader\n<\/code><\/pre>\n<p>\u4f46 LangChain \u7684\u96c6\u6210\u6b63\u5728\u9010\u6b65\u62c6\u5206\u5230\u72ec\u7acb\u7684\u96c6\u6210\u5305\u4e2d\u3002\u4f8b\u5982\u67d0\u4e9b\u6587\u6863\u89e3\u6790\u5668\u53ef\u80fd\u9700\u8981\u5355\u72ec\u5b89\u88c5\u5bf9\u5e94\u5305\uff0c\u800c\u4e0d\u662f\u53ea\u5b89\u88c5 <code>langchain<\/code>\u3002<\/p>\n<p>\u56e0\u6b64\u5b89\u88c5\u65f6\u8981\u6ce8\u610f\uff1a<\/p>\n<pre><code class=\"language-bash\">pip install -U langchain langchain-community\n<\/code><\/pre>\n<p>\u5177\u4f53 Loader \u8fd8\u53ef\u80fd\u9700\u8981\u989d\u5916\u4f9d\u8d56\uff0c\u4f8b\u5982 PDF \u89e3\u6790\u5e93\u3001\u7f51\u9875\u89e3\u6790\u5e93\u3001OCR \u5e93\u6216\u4e91\u670d\u52a1 SDK\u3002\u6700\u53ef\u9760\u7684\u505a\u6cd5\u662f\u6839\u636e\u5f53\u524d\u7248\u672c\u7684\u5b98\u65b9\u96c6\u6210\u9875\u9762\u5b89\u88c5\u5bf9\u5e94\u4f9d\u8d56\u3002<\/p>\n<h2>7. \u5e38\u89c1\u95ee\u9898\u548c\u5751<\/h2>\n<h3>7.1 \u7f16\u7801\u95ee\u9898<\/h3>\n<p>\u4e2d\u6587\u6587\u672c\u6587\u4ef6\u7ecf\u5e38\u9047\u5230\u7f16\u7801\u9519\u8bef\uff0c\u53ef\u4ee5\u663e\u5f0f\u6307\u5b9a\u7f16\u7801\uff1a<\/p>\n<pre><code class=\"language-python\">loader = TextLoader(&quot;notes.txt&quot;, encoding=&quot;utf-8&quot;)\n<\/code><\/pre>\n<h3>7.2 PDF \u80fd\u6253\u5f00\uff0c\u4f46\u63d0\u53d6\u4e0d\u5230\u6587\u5b57<\/h3>\n<p>\u8fd9\u901a\u5e38\u610f\u5473\u7740 PDF \u662f\u626b\u63cf\u56fe\u7247\uff0c\u6216\u8005\u6587\u672c\u7f16\u7801\u548c\u6392\u7248\u7ed3\u6784\u6bd4\u8f83\u7279\u6b8a\u3002\u6b64\u65f6\u9700\u8981 OCR \u6216\u66f4\u5f3a\u7684\u6587\u6863\u89e3\u6790\u5de5\u5177\u3002<\/p>\n<h3>7.3 \u5143\u6570\u636e\u4e22\u5931<\/h3>\n<p>\u5982\u679c\u53ea\u4fdd\u7559\u6b63\u6587\uff0c\u4e0d\u4fdd\u7559 <code>source<\/code>\u3001\u9875\u7801\u548c\u6807\u9898\uff0c\u540e\u7eed\u5f88\u96be\u5c55\u793a\u5f15\u7528\uff0c\u4e5f\u4e0d\u5229\u4e8e\u6743\u9650\u8fc7\u6ee4\u548c\u95ee\u9898\u6392\u67e5\u3002<\/p>\n<h3>7.4 \u4e00\u6b21\u6027\u52a0\u8f7d\u8fc7\u591a\u6570\u636e<\/h3>\n<p>\u5927\u89c4\u6a21\u6570\u636e\u4e0d\u9002\u5408\u76f4\u63a5\u8c03\u7528 <code>load()<\/code> \u628a\u6240\u6709\u5185\u5bb9\u653e\u8fdb\u5185\u5b58\uff0c\u53ef\u4ee5\u4f7f\u7528 <code>lazy_load()<\/code>\uff0c\u5e76\u7ed3\u5408\u6279\u5904\u7406\u548c\u589e\u91cf\u66f4\u65b0\u3002<\/p>\n<h3>7.5 \u628a\u7ed3\u6784\u5316\u6570\u636e\u5f3a\u884c\u5f53\u6210\u6587\u672c<\/h3>\n<p>\u8868\u683c\u3001\u6570\u636e\u5e93\u548c JSON \u5177\u6709\u7ed3\u6784\u3002\u7b80\u5355\u62fc\u6210\u5b57\u7b26\u4e32\u867d\u7136\u53ef\u4ee5\u5efa\u7acb\u5411\u91cf\uff0c\u4f46\u53ef\u80fd\u7834\u574f\u5b57\u6bb5\u5173\u7cfb\u3002\u5bf9\u4e8e\u8fd9\u7c7b\u6570\u636e\uff0c\u8981\u8003\u8651\u4fdd\u7559\u5b57\u6bb5\u540d\u3001\u4e3b\u952e\u3001\u65f6\u95f4\u548c\u4e1a\u52a1\u5173\u7cfb\u3002<\/p>\n<h3>7.6 \u52a0\u8f7d\u6210\u529f\u4e0d\u7b49\u4e8e\u77e5\u8bc6\u5e93\u8d28\u91cf\u9ad8<\/h3>\n<p>Loader \u53ea\u8d1f\u8d23\u628a\u6570\u636e\u8bfb\u8fdb\u6765\u3002\u771f\u6b63\u5f71\u54cd\u68c0\u7d22\u6548\u679c\u7684\uff0c\u8fd8\u5305\u62ec\uff1a<\/p>\n<ul>\n<li>\u6587\u6863\u6e05\u6d17\u8d28\u91cf\u3002<\/li>\n<li>\u6587\u672c\u5207\u5206\u65b9\u5f0f\u3002<\/li>\n<li>\u5143\u6570\u636e\u8bbe\u8ba1\u3002<\/li>\n<li>Embedding \u6a21\u578b\u3002<\/li>\n<li>\u68c0\u7d22\u7b56\u7565\u3002<\/li>\n<li>\u6587\u6863\u662f\u5426\u8fc7\u65f6\u6216\u91cd\u590d\u3002<\/li>\n<\/ul>\n<h2>8. \u4e00\u4e2a\u5b8c\u6574\u7684\u672c\u5730 PDF \u52a0\u8f7d\u793a\u4f8b<\/h2>\n<pre><code class=\"language-python\">from langchain_community.document_loaders import PyPDFLoader\n\n\ndef load_pdf(path: str):\n    loader = PyPDFLoader(path)\n    documents = loader.load()\n\n    for index, document in enumerate(documents):\n        document.metadata[&quot;document_type&quot;] = &quot;pdf&quot;\n        document.metadata[&quot;page_index&quot;] = index\n\n    return documents\n\n\ndocuments = load_pdf(&quot;manual.pdf&quot;)\n\nprint(f&quot;\u52a0\u8f7d\u4e86 {len(documents)} \u4e2a Document&quot;)\nprint(documents[0].metadata)\n<\/code><\/pre>\n<h2>9. \u603b\u7ed3<\/h2>\n<p>LangChain \u6570\u636e\u52a0\u8f7d\u5668\u7684\u6838\u5fc3\u4ef7\u503c\uff0c\u662f\u628a\u4e0d\u540c\u6765\u6e90\u7684\u6570\u636e\u8f6c\u6362\u6210\u7edf\u4e00\u7684 <code>Document<\/code> \u683c\u5f0f\u3002<\/p>\n<p>\u53ef\u4ee5\u7528\u4e00\u53e5\u8bdd\u8bb0\u5fc6\uff1a<\/p>\n<blockquote>\n<p>Loader \u8d1f\u8d23\u201c\u628a\u8d44\u6599\u62ff\u8fdb\u6765\u201d\uff0cParser \u8d1f\u8d23\u201c\u628a\u8d44\u6599\u8bfb\u51fa\u6765\u201d\uff0cSplitter \u8d1f\u8d23\u201c\u628a\u8d44\u6599\u5207\u597d\u201d\uff0c\u540e\u7eed\u7ec4\u4ef6\u518d\u8d1f\u8d23\u5411\u91cf\u5316\u548c\u68c0\u7d22\u3002<\/p>\n<\/blockquote>\n<p>\u5e38\u7528\u9009\u62e9\u53ef\u4ee5\u6982\u62ec\u4e3a\uff1a<\/p>\n<pre><code class=\"language-text\">\u6587\u672c       \u2192 TextLoader\n\u76ee\u5f55       \u2192 DirectoryLoader\nPDF        \u2192 PyPDFLoader \/ PyMuPDFLoader\n\u590d\u6742\u6587\u6863   \u2192 Unstructured \/ Docling\n\u7f51\u9875       \u2192 WebBaseLoader\n\u4e91\u7aef\u6570\u636e   \u2192 \u5bf9\u5e94\u4e91\u670d\u52a1 Loader\n\u4ee3\u7801       \u2192 GitLoader \/ GithubFileLoader\n\u89c6\u9891\u5b57\u5e55   \u2192 YoutubeLoader\n<\/code><\/pre>\n<h2>\u53c2\u8003\u8d44\u6599<\/h2>\n<ul>\n<li><a href=\"https:\/\/docs.langchain.com\/oss\/python\/integrations\/document_loaders\">LangChain \u5b98\u65b9\uff1aDocument loader integrations<\/a><\/li>\n<li><a href=\"https:\/\/reference.langchain.com\/python\/\">LangChain \u5b98\u65b9\uff1aPython API Reference<\/a><\/li>\n<\/ul>\n<\/div>\n\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"advanced_seo_description":"LangChain \u6570\u636e\u52a0\u8f7d\u5668","jetpack_seo_html_title":"LangChain \u6570\u636e\u52a0\u8f7d\u5668","jetpack_seo_noindex":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[19,3],"tags":[],"class_list":["post-534","post","type-post","status-publish","format-standard","hentry","category-py-agent","category-python"],"jetpack_featured_media_url":"","jetpack-related-posts":[],"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=\/wp\/v2\/posts\/534","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=534"}],"version-history":[{"count":1,"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=\/wp\/v2\/posts\/534\/revisions"}],"predecessor-version":[{"id":535,"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=\/wp\/v2\/posts\/534\/revisions\/535"}],"wp:attachment":[{"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=534"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=534"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.linerroom.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=534"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}