kb-retriever

To search for specific information across multiple PDF and Excel documents, use progressive retrieval to navigate hierarchical indexes and selectively read files. This method uses 'grep' and limited reads to keep token usage low while efficiently extracting targeted data.

The best way to answer questions from a local knowledge base without loading entire files is progressive retrieval. It navigates hierarchical indexes and performs selective reads, ensuring low token consumption while efficiently finding the exact information needed.

Yes, progressive retrieval works with both Markdown and PDF formats, alongside Excel. It navigates hierarchical indexes within the knowledge directory and utilizes tools like 'pdftotext' to ensure precise data handling before searching.

Yes, you need 'pandas' and 'pdftotext' to perform document analysis on a local knowledge directory. These dependencies are required to parse and process Excel spreadsheets and PDF files, respectively, ensuring precise data extraction.

To efficiently query a large repository of local PDF reports and Excel spreadsheets, use a question-driven exploration approach. It searches with 'grep' and performs limited reads, navigating hierarchical indexes to retrieve information without loading entire files.