upload_knowledge.py

from api.pinecone_func import upsert_texts
from pathlib import Path
from utils.chunking import chunk_text
import uuid
from dotenv import load_dotenv
load_dotenv()

def load_documents(folder="docs"):
    texts = []
    ids = []
    for file in Path(folder).glob("*.txt"):
        with open(file, "r", encoding="utf-8") as f:
            content = f.read()
            chunks = chunk_text(content)
            for chunk in chunks:
                texts.append(chunk)
                ids.append(str(uuid.uuid4()))
    return texts, ids

if __name__ == "__main__":
    texts, ids = load_documents()
    print(f"{len(texts)}件の知識チャンクをアップロード中")
    upsert_texts(texts, ids)
    print("アップロード完了")