awesome-copilot-index copilot-instructions.md

Copilot Instructions for Awesome Copilot Index

Project Architecture

This project is a multi-stage content indexing pipeline that downloads, parses, and publishes GitHub repository content. The architecture consists of:

1. Content Extraction: Downloads GitHub repos as ZIP files (not git clones) to avoid history
2. Markdown Parsing: Parses frontmatter and content using custom trimming logic for edge cases
3. Index Generation: Transforms raw markdown into structured JSON/EDN for consumption
4. Jekyll Publishing: Renders a searchable web interface with responsive CSS
5. VS Code Integration: Provides ClojureScript-based Joyride script for in-editor access

Key Data Flow: GitHub ZIP → Babashka Parser → JSON/EDN Index → Jekyll Site → Joyride Menu

Always show what you evaluate in the repl to the user, using a codeblock that starts with the namespace in a (in-ns ...) form.

Critical Workflows

Main Build Pipeline

bb download-awesome        # Downloads awesome-copilot repo as ZIP
bb generate-index         # Parses markdown, generates site/awesome-copilot.{json,edn}
bb download-cursorrules   # Downloads awesome-cursorrules repo
bb generate-cursorrules-index  # Generates cursor rules index

Development REPL Workflow

(in-ns 'pez.index)
(require 'pez.index :reload)
;; Test parsing with problematic content
(def test-content "---\ntitle: test\n---  \n\n# Content")
(parse-markdown-file test-content)

Jekyll Site Generation

• Templates use Selmer syntax: {{instruction-count}}, {{formatted-date}}
• Site builds from site/ directory with generated JSON/EDN indices
• CSS supports both ul (bullet) and ol (numbered) lists with blue styling

Development Template Workflow

When editing site/index-template.md during development:

1. Edit the template file: Make changes to site/index-template.md
2. Regenerate index: Run bb generate-index to process template with real data
3. Jekyll auto-reload: Jekyll's live reload automatically picks up the generated index.md changes