File Loader Configuration

DeepSearcher supports various file loaders to extract and process content from different file formats.

📝 Basic Configuration

config.set_provider_config("file_loader", "(FileLoaderName)", "(Arguments dict)")

📋 Available File Loaders

Loader	Description	Supported Formats
UnstructuredLoader	General purpose document loader with broad format support	PDF, DOCX, PPT, HTML, etc.
DoclingLoader	Document processing library with extraction capabilities	See documentation

🔍 File Loader Options

Unstructured

Unstructured is a powerful library for extracting content from various document formats.

config.set_provider_config("file_loader", "UnstructuredLoader", {})

Setup Instructions

You can use Unstructured in two ways:

1. With API (recommended for production)

2. Set environment variables:

   • UNSTRUCTURED_API_KEY
   • UNSTRUCTURED_API_URL

3. Local Processing

4. Simply don't set the API environment variables
5. Install required dependencies:

   # Install core dependencies
   pip install unstructured-ingest
   
   # For all document formats
   pip install "unstructured[all-docs]"
   
   # For specific formats (e.g., PDF only)
   pip install "unstructured[pdf]"
   

For more information: - Unstructured Documentation - Installation Guide

Docling

Docling provides document processing capabilities with support for multiple formats.

config.set_provider_config("file_loader", "DoclingLoader", {})

Setup Instructions

1. Install Docling:

   pip install docling
   

2. For information on supported formats, see the Docling documentation.