File extensions and libraries for llamaindex

services/rag/llamaindex/EXTENSIONS.md

@@ -0,0 +1,46 @@
+# Supported File Extensions and LlamaIndex Loaders
+
+This document lists the file extensions found in the `./../../../data` directory and the corresponding LlamaIndex loaders that can be used to process them.
+
+## Document Formats
+
+| Extension | File Type | LlamaIndex Loader | Installation Package |
+|-----------|-----------|-------------------|---------------------|
+| `.pdf` | Portable Document Format | `PDFReader` or `SimpleDirectoryReader` | `llama-index-readers-file` |
+| `.docx` | Microsoft Word Document | `DocxReader` or `SimpleDirectoryReader` | `llama-index-readers-file` |
+| `.xlsx` | Microsoft Excel Spreadsheet | `PandasExcelReader` or `SimpleDirectoryReader` | `llama-index-readers-file` |
+| `.pptx` | Microsoft PowerPoint Presentation | `PptxReader` or `SimpleDirectoryReader` | `llama-index-readers-file` |
+| `.odt` | OpenDocument Text | `SimpleDirectoryReader` with `UnstructuredReader` | `llama-index-readers-file` |
+
+## Image Formats
+
+| Extension | File Type | LlamaIndex Loader | Installation Package |
+|-----------|-----------|-------------------|---------------------|
+| `.png` | Portable Network Graphics | `ImageReader` | `llama-index-readers-file` |
+| `.jpg` | JPEG Image | `ImageReader` | `llama-index-readers-file` |
+
+## Archive Formats
+
+| Extension | File Type | LlamaIndex Loader | Installation Package |
+|-----------|-----------|-------------------|---------------------|
+| `.zip` | ZIP Archive | `SimpleDirectoryReader` with archive support | `llama-index-readers-file` |
+
+## System/Special Files (Ignored)
+
+- `.DS_Store` - macOS system file
+- `.gitignore` - Git configuration file
+
+## Audio/Video Formats (Skipped as per requirements)
+
+- `.m4a` - Audio file
+- `.mp3` - Audio file
+- `.mp4` - Video file
+- `.ogg` - Audio/Video file
+
+## Notes
+
+1. Many file types can be loaded using the `SimpleDirectoryReader` which automatically detects and handles multiple file formats.
+2. For advanced document parsing, specific readers might offer better performance or more features.
+3. All required dependencies have been installed with `llama-index-readers-file` and `patool` for archive support.
+4. No external API keys are required for the supported file types, as we're using local processing solutions.
+5. The system prioritizes local processing over cloud services as per requirements.