Create an integration between PDF parsing tools and vector databases to automate the pipeline from raw file to knowledge graph. Ensure high accuracy for complex table layouts.