1
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 1 year has passed since last update.

LangChain の DocumentLoader のように HTMLファイルの一部を読み取っていくサンプル

Posted at

概要

手順

  • BeautifulSoup を使って HTML ファイルの一部を読み取る
  • LangChain の Document クラスを手動で生成する

サンプルスクリプト

import os
from pathlib import Path
from bs4 import BeautifulSoup
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.docstore.document import Document

# 特定ディレクトリのサブディレクトリも含むファイル一覧を取得
html_dir = Path("どこかのディレクトリ")
html_files = list(html_dir.glob("**/*.htm"))

# HTML内のデータを分割する用のSplitterを用意
text_splitter = RecursiveCharacterTextSplitter(chunk_size=200, chunk_overlap=0)

all_documents = []
for p in html_files:
    # HTMLファイルを BeautifulSoup で読み込み
    with open(p, "r", encoding="UTF-8") as f:
        soup = BeautifulSoup(f)

    # HTMLの一部を取得(今回は id 指定)
    content = soup.find(id="main-content").text
    # タイトルを取得(今回は h1 タグのテキスト)
    title = soup.find("h1").text

    # langchain の DocumentLoader を使わず、自前で Document を作成
    # メタデータにファイルのパス、タイトルを設定
    doc = Document(page_content=content, metadata={"source": p, "title": title})

    # documentをsplitterで分割してリストに追加
    splited = text_splitter.split_documents([doc])
    all_documents += splited

# LangChain のサンプルを参考にベクトル化するなど
embeddings = OpenAIEmbeddings(deployment="text-embedding-ada-002")
db = Chroma(persist_directory="_vecorstore", embedding_function=embeddings)
for doc in all_documents:
    doc.metadata["source"] = str(doc.metadata["source"])
    db.add_texts([doc.page_content], [doc.metadata])

1
1
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
1
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?