AI Source Code Mapper

Role & Goal

You are an expert mobile developer and system architect. Your task is to build an Android application (APK) that analyzes local or GitHub-cloned repositories, generates a deep architectural understanding of the codebase, and renders an interactive, collapsible HTML summary report with integrated multi-LLM support.

Key Application Features

1. Repository Input & Processing

• Allow the user to either open a local repository or clone a public/private repository from GitHub.
• Perform a complete recursive code analysis to capture all active and passive codebase elements (UI components, files, functions, methods, triggers, etc.).

2. Multi-LLM Endpoint Configuration

Implement a configurable LLM provider settings screen allowing users to choose and switch between the following endpoints for repository analysis:

1. Primary Custom OpenAI-Compatible API:
2. Ollama Cloud Endpoint:
   • Supports user-provided API Key or OAuth authentication.
3. OpenCode Go Endpoint:
   • Supports user-provided API Key and custom model selection.
4. OpenCode Zen Endpoint:
   • Configured with the default free model name.

Report Generation Requirements

Generate a standalone, interactive HTML report with collapsible details (<details>/<summary> or custom JS collapse).

Column Structure (4 Columns):

1. Name / Identifier / Description (1st Column):
   • Must cover EVERY code element: UI elements, files, functions, methods, triggers, variables, and architectural entities.
   • Includes: Name + brief high-level explanation of what the element does.
2. Category / Technology / Type (2nd Column):
   • Detailed specification of method type, design pattern, class type, technology stack, or execution layer used.
3. Exact Code & Line Reference (3rd Column):
   • Direct line number(s) and the exact full source code snippet corresponding to the entry.
   • Completeness Rule: Every single line of code in the analyzed repository must eventually map to a corresponding row in this column.
4. Code Quality & Security Assessment (4th Column):
   • AI-driven evaluation for that specific snippet.
   • Categorize and badge each entry:
     • ⚠️ Problem / Warning (potential bug or performance bottleneck)
     • 🔒 Vulnerability (security risk)
     • ✅ Best Practice Passed (clean, optimal code)

UI / Layout & Collapse Constraints

• Responsive Width Handling: If any text or content block exceeds 25% (1/4) of the screen width, it MUST be automatically collapsed.
• Accordion / Collapsible UX:
  • In the collapsed state, only show a concise summary (1-2 sentences max).
  • The full, detailed content (full line code snippets, deep analysis) must only expand upon user click.
• Provide a clean, modern UI for the generated HTML with dark mode / light mode toggle and instant search/filter capabilities.

enhanced by gemini