v1.5.0

Full Changelog: v1.4.3...v1.5.0

In this version:

• Added configuration option to ignore file type enforcement
• Added file decorations for marked files in the explorer
• Improved file marking logic to better handle single file selections
  • Individual files that were ignored/excluded can be manually marked for context if they are marked by themselves (via command palette or via context menu)
• Now the extension processes all ignore files that exist in a workspace matching a filename in the config
  • i.e. a file at /foo/bar/.gitignore will not be properly parsed and applied
• Fixed a place where the config value for token warnings was not properly implemented
• Added a new icon

v1.4.3

• Remove extra command/context item

Full Changelog: v1.4.2...v1.4.3

v1.4.2

• Refactor default file extensions in package.json to support more items out of the box
• Show live token count for marked items and a generate button in the Explorer view
• Provide more useful UX feedback when marking files that have excluded file extensions or match and ignore pattern

Full Changelog: v1.4.1...v1.4.2

v1.4.1

What's New in 1.4.1

• New Versioning: This extension now aligns with semantic versioning and is bumping from 0.3 to 1.4 on a minor release, and then to .1 for a final fix for a last minute WASM build copy issue.
• New Name: The name change to "LLM Context Generator" makes it clear that the extension can be used with various LLMs, not just GPT models.
• Improved File Exclusions: Adds a new .ignore configuration to add additional .ignore pattern lists.
  • Closes File exclusion settings #2
• Improved File Marking System: Adds multi-file/folder marking system using the Explorer menu to easily select specific files for context generation. Automatic file tracking handles when marked files are moved or deleted.
• Enhanced File Extension Handling: Broadened support for more diverse file types and programming languages out of the box.
• Improved Token Count Estimation: Switched from gpt-3-encoder to @dqbd/tiktoken for enhanced token counting accuracy and efficiency. for the generated context to help manage usage within LLM token limits, with a configurable warning if the context exceeds a predefined number of tokens (default: 8,000).
• Codebase Refactor: Improved code structure and maintainability using ESBuild, TypeScript, and a more modular design. Removed unused test files and outdated dependencies.
• Updated README: Revised documentation to reflect the new name, features, and usage instructions.

New Contributors

• @dependabot made their first contribution in #4

Full Changelog: v0.3...v1.4.1

v1.4.0 - Pre Release

Note: v1.4.0 was immediately superseded by v1.4.1 due to an issue with the copying of tiktoken's WASM module

What's Changed

• Bump micromatch from 4.0.5 to 4.0.8 by @dependabot in #4
• Bump braces from 3.0.2 to 3.0.3 by @dependabot in #5
• Update Dependencies by @codybrom in #14
• Bump typescript from 5.6.3 to 5.7.2 by @dependabot in #21
• Bump @types/node from 22.8.2 to 22.10.2 by @dependabot in #27
• Bump ignore from 6.0.2 to 7.0.0 by @dependabot in #28
• Bump mocha from 10.7.3 to 11.0.1 by @dependabot in #25
• Bump @types/vscode from 1.94.0 to 1.96.0 by @dependabot in #26
• Bump cross-spawn from 7.0.3 to 7.0.6 in the npm_and_yarn group by @dependabot in #19
• Bump prettier from 3.3.3 to 3.4.2 by @dependabot in #24
• Bump @types/node from 22.10.2 to 22.10.5 by @dependabot in #29
• Bump @types/mocha from 10.0.9 to 10.0.10 by @dependabot in #30

Full Changelog: v0.3...v1.4.0

v0.3

• Added the ability to add item to "Marked Files" via Right Click
• Updated some package internals

v0.2.1

This release includes:

• Add options for more flexibility and mark/unmark individual files for context generation #1 (Major thanks to @Aventuum!)

v0.1.1

In this release I've fixed import context generation

v0.1 - configuration in settings

In this release, you now have control over how the extension shares the generated output and what files can be included. You also have the option to include/exclude package.json (default is on).

This felt like such a major improvement I've upped the release to v0.1. Thanks to everyone joining me on this journey!

Coming Soon

• Configurable token warning

Full Changelog: v0.0.2...v0.1

v0.0.3

Switch to the gtp-3-encoder package for token estimation
Code now styled by prettier in airbnb-style