RAG_backend / src /ingestion /document_loader.py
vinimoreira's picture
Add files for RAG backend
2068d15 verified
raw
history blame contribute delete
510 Bytes
import os
from pathlib import Path
SUPPORTED_EXTENSIONS = {".txt", ".md"}
def load_documents_from_dir(directory: str) -> list[dict]:
docs = []
for file_path in Path(directory).rglob("*"):
if file_path.suffix.lower() in SUPPORTED_EXTENSIONS:
with open(file_path, "r", encoding="utf-8") as f:
content = f.read()
docs.append({
"content": content,
"source": str(file_path)
})
return docs