internal-document-qa / import_data.py
Vincent Claes
first working end to end version
13b16b6
raw
history blame
No virus
2.59 kB
import os
import weaviate
from llama_index import download_loader
from llama_index.vector_stores import WeaviateVectorStore
from llama_index import VectorStoreIndex, StorageContext
from pathlib import Path
import argparse
def get_pdf_files(base_path, loader):
"""
Get paths to all PDF files in a directory and its subdirectories.
Parameters:
- base_path (str): The path to the starting directory.
Returns:
- list of str: A list of paths to all PDF files found.
"""
pdf_paths = []
# Check if the base path exists and is a directory
if not os.path.exists(base_path):
raise FileNotFoundError(f"The specified base path does not exist: {base_path}")
if not os.path.isdir(base_path):
raise NotADirectoryError(
f"The specified base_path is not a directory: {base_path}"
)
# Loop through all directories and files starting from the base path
for root, dirs, files in os.walk(base_path):
for filename in files:
# If a file has a .pdf extension, add its path to the list
if filename.endswith(".pdf"):
pdf_file = loader.load_data(file=Path(root, filename))
pdf_paths.extend(pdf_file)
return pdf_paths
def main(args):
PDFReader = download_loader("PDFReader")
loader = PDFReader()
documents = get_pdf_files(args.pdf_dir, loader)
client = weaviate.Client(
url=os.environ["WEAVIATE_URL"],
auth_client_secret=weaviate.AuthApiKey(api_key=os.environ["WEAVIATE_API_KEY"]),
additional_headers={"X-OpenAI-Api-Key": os.environ["OPENAI_API_KEY"]},
)
# construct vector store
vector_store = WeaviateVectorStore(
weaviate_client=client, index_name=args.customer, text_key="content"
)
# setting up the storage for the embeddings
storage_context = StorageContext.from_defaults(vector_store=vector_store)
# set up the index
index = VectorStoreIndex(documents, storage_context=storage_context)
query_engine = index.as_query_engine()
response = query_engine.query(args.query)
print(response)
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Process and query PDF files.")
parser.add_argument("--customer", default="Ausy", help="Customer name")
parser.add_argument("--pdf_dir", default="./data", help="Directory containing PDFs")
parser.add_argument(
"--query",
default="What is CX0 customer exprience office?",
help="Query to execute",
)
args = parser.parse_args()
main(args)