Skip to content

Source Path Examples

This document provides detailed examples for configuring SOURCE_PATHS in Flexible GraphRAG across different operating systems and scenarios.

📁 Basic Syntax

SOURCE_PATHS=["path1", "path2", "path3"]

The SOURCE_PATHS configuration accepts a JSON array of strings, where each string can be: - A single file path - A directory path (processes ALL files in the directory) - Mixed files and directories

🪟 Windows Examples

Single File

# Using forward slashes (recommended)
SOURCE_PATHS=["C:/Documents/report.pdf"]

# Using backslashes (escape required)
SOURCE_PATHS=["C:\\Documents\\report.pdf"]

# Windows "Copy as path" format (paste directly!)
SOURCE_PATHS=["C:\Users\John Doe\Documents\My Report.pdf"]

Multiple Files

SOURCE_PATHS=["C:/file1.pdf", "D:/folder/file2.docx", "E:/data/file3.txt"]

# With spaces in paths
SOURCE_PATHS=["C:\1 sample files\cmispress.pdf", "D:\My Files\reports\annual.docx"]

Whole Directory

# Processes ALL files in the directory
SOURCE_PATHS=["C:/Documents/reports"]

# Multiple directories
SOURCE_PATHS=["C:/Documents/reports", "D:/Projects/data", "E:/Archive"]

Mixed Files and Directories

SOURCE_PATHS=["C:/specific-file.pdf", "C:/Documents/folder", "D:/another-file.docx"]

Relative Paths (Windows)

# Relative to the project root
SOURCE_PATHS=["./sample-docs", "./data/reports"]
SOURCE_PATHS=["..\\parent-folder\\documents"]

UNC Network Paths

# Using forward slashes
SOURCE_PATHS=["//server/share/folder"]

# Using backslashes
SOURCE_PATHS=["\\\\server\\share\\folder"]

🍎 macOS Examples

Single File

SOURCE_PATHS=["/Users/username/Documents/report.pdf"]
SOURCE_PATHS=["/Applications/MyApp/data/file.txt"]

Multiple Files

SOURCE_PATHS=["/Users/john/Desktop/file1.pdf", "/Users/john/Downloads/file2.docx"]

Whole Directory

SOURCE_PATHS=["/Users/username/Documents/work-files"]
SOURCE_PATHS=["/Volumes/ExternalDrive/data"]

Relative Paths (macOS)

SOURCE_PATHS=["./sample-docs", "../shared-data"]
SOURCE_PATHS=["~/Documents/my-files"]  # Home directory shortcut

🐧 Linux Examples

Single File

SOURCE_PATHS=["/home/username/documents/report.pdf"]
SOURCE_PATHS=["/opt/data/analysis.txt"]

Multiple Files

SOURCE_PATHS=["/home/user/file1.pdf", "/var/data/file2.csv", "/tmp/temp-file.txt"]

Whole Directory

SOURCE_PATHS=["/home/username/project-data"]
SOURCE_PATHS=["/mnt/shared/documents"]

Relative Paths (Linux)

SOURCE_PATHS=["./local-data", "../shared-folder"]
SOURCE_PATHS=["~/documents/work"]  # Home directory shortcut

⚠️ Important Notes

Directory Processing Warning

When you specify a directory path, the system will process ALL files in that directory: 

# This processes EVERY file in the reports directory
SOURCE_PATHS=["C:/Documents/reports"]

If you only want specific files, list them individually: 

# This processes only these specific files
SOURCE_PATHS=["C:/Documents/reports/q1.pdf", "C:/Documents/reports/q2.pdf"]

File Types Supported

The system supports these file formats: - Documents: PDF, DOCX, PPTX, TXT, MD - Spreadsheets: XLSX, CSV
- Web: HTML - Images: PNG, JPG (with OCR) - Archive: ASCIIDOC

Path Encoding

  • Windows: Use forward slashes / or double backslashes \\
  • Spaces: Paths with spaces are supported, no escaping needed in JSON array
  • Special characters: Most Unicode characters are supported

🖥️ UI Client Differences

Different UI clients use different environment variable names:

Backend (FastAPI)

SOURCE_PATHS=["./sample-docs/file.pdf"]

Angular Frontend

PROCESS_FOLDER_PATH=./sample-docs

React/Vue Frontends

VITE_PROCESS_FOLDER_PATH=./sample-docs

Note: Frontend environment variables typically expect a single directory path, while the backend SOURCE_PATHS accepts multiple files and directories.

🗄️ Repository Path Examples (CMIS/Alfresco)

CMIS Repository Paths

CMIS uses standard CMIS path format: 

# CMIS paths start with /
CMIS_FOLDER_PATH=/Shared/Documents
CMIS_FOLDER_PATH=/Sites/my-site/documentLibrary/folder

Alfresco Repository Paths

NEW (python-alfresco-api 1.1.5+): Alfresco now uses native Alfresco paths with flexible format:

# Short format (recommended - matches what you see in Alfresco Share)
ALFRESCO_PATH=/Shared/GraphRAG
ALFRESCO_PATH=/Sites/my-site/documentLibrary/Reports
ALFRESCO_PATH=/User Homes/admin/My Files

# Full format (also works - system automatically strips /Company Home prefix)
ALFRESCO_PATH=/Company Home/Shared/GraphRAG
ALFRESCO_PATH=/Company Home/Sites/my-site/documentLibrary/Reports

Both formats work! The system automatically strips /Company Home prefix if present, since the root node already represents Company Home.

Benefits of Native Alfresco Paths: - ✅ More intuitive - matches what you see in Alfresco Share UI - ✅ Consistent with Alfresco Content Services API - ✅ Works with relative_path feature for better performance - ✅ Flexible - use short format (/Shared) or full format (/Company Home/Shared) - ✅ Backward compatible - both formats supported

Path Examples: 

# Shared folder (short format - recommended)
/Shared/GraphRAG/documents

# Sites (short format - recommended)
/Sites/engineering/documentLibrary/specs

# User Homes (short format - recommended)
/User Homes/admin/My Files

# Data Dictionary (short format - recommended)
/Data Dictionary/Scripts

# Full format also works (optional)
/Company Home/Shared/GraphRAG/documents
/Company Home/Sites/engineering/documentLibrary/specs

💡 Best Practices

  1. Use relative paths when possible for portability
  2. Start small - test with one file before processing large directories
  3. Use forward slashes on Windows for consistency
  4. Check file permissions ensure the application can read the specified paths
  5. Avoid system directories stick to user documents and data folders
  6. Test paths verify paths exist and are accessible before running
  7. Alfresco paths - Use native /Company Home/... format for clarity (python-alfresco-api 1.1.5+)