Skip to content
Logo CodeGraphContext / Documentation
CodeGraphContext/CodeGraphContext
v0.4.7 3.2k 564
Home › User Guides › Indexing Source Code

Indexing Source Code

Indexing is the process of parsing your source code and materializing it into the code graph. CodeGraphContext (CGC) supports various indexing strategies to suit different workflows.

1. Local Repository Indexing

The most common use case is indexing the project you are currently working on.

cd /path/to/project
cgc index

Partial Scans

If you only want to index a specific subdirectory or file:

cgc index ./src/core

Incremental Updates

CGC is smart enough to skip files that haven't changed since the last index. Simply run cgc index again to sync your changes.

2. Dependency Indexing

To get a complete picture of your application, you may want to index its dependencies.

Python Packages

CGC can automatically find and index installed Python packages:

cgc index-package requests

Generic Folders

You can also index any arbitrary folder and mark it as a dependency:

cgc index --path /path/to/lib --dependency

3. Real-time Monitoring (watch)

For projects under active development, use the watch command. CGC will monitor your filesystem for changes (create, update, delete) and keep the graph in sync in the background.

cgc watch
  • Background Jobs: The watch command runs as a background process.
  • Status: Check the status of your watchers with cgc list-watchers.

4. Advanced Indexing Options

Option Description
--deep Performs a deep scan, following imports and resolving external symbols.
--scip (Experimental) Uses SCIP index data if available for higher precision.
--exclude Glob patterns to ignore (e.g., **/tests/**).

The .cgcignore File

Similar to .gitignore, you can create a .cgcignore file in your repository root to permanently exclude files or directories from the index.

# Example .cgcignore
node_modules/
dist/
*.pyc

Performance Tips

  • Memory: Large repositories may require significant memory during the initial parse phase.
  • Storage: Ensure you have enough disk space for the graph database (usually 2-5x the size of the source code).
CodeGraphContext v0.1.0 · Documentation

Indexing Source Code

Indexing is the process of parsing your source code and materializing it into the code graph. CodeGraphContext (CGC) supports various indexing strategies to suit different workflows.

1. Local Repository Indexing

The most common use case is indexing the project you are currently working on.

cd /path/to/project
cgc index

Partial Scans

If you only want to index a specific subdirectory or file:

cgc index ./src/core

Incremental Updates

CGC is smart enough to skip files that haven't changed since the last index. Simply run cgc index again to sync your changes.

2. Dependency Indexing

To get a complete picture of your application, you may want to index its dependencies.

Python Packages

CGC can automatically find and index installed Python packages:

cgc index-package requests

Generic Folders

You can also index any arbitrary folder and mark it as a dependency:

cgc index --path /path/to/lib --dependency

3. Real-time Monitoring (watch)

For projects under active development, use the watch command. CGC will monitor your filesystem for changes (create, update, delete) and keep the graph in sync in the background.

cgc watch
  • Background Jobs: The watch command runs as a background process.
  • Status: Check the status of your watchers with cgc list-watchers.

4. Advanced Indexing Options

Option Description
--deep Performs a deep scan, following imports and resolving external symbols.
--scip (Experimental) Uses SCIP index data if available for higher precision.
--exclude Glob patterns to ignore (e.g., **/tests/**).

The .cgcignore File

Similar to .gitignore, you can create a .cgcignore file in your repository root to permanently exclude files or directories from the index.

# Example .cgcignore
node_modules/
dist/
*.pyc

Performance Tips

  • Memory: Large repositories may require significant memory during the initial parse phase.
  • Storage: Ensure you have enough disk space for the graph database (usually 2-5x the size of the source code).