ed
de3dfec4de
- context_comp_decouple_20260510: Decouple Files & Media from Context Composition, directory grouping, file stats - context_comp_slices_20260510: Slice visualization with annotations, view presets - context_comp_presets_20260510: Context preset save/load with validation, context preview - Depends_on chain: presets -> slices -> decouple
2.3 KiB
2.3 KiB
Implementation Plan: Context Composition Decoupling
Phase 1: Core Data Model Changes
Focus: Add view_mode field to FileItem, understand current coupling
- Task 1.1: Audit FileItem model in models.py - add view_mode and custom_slices fields
- Task 1.2: Audit _render_context_composition_panel() to understand current coupling
- Task 1.3: Audit _render_files_panel() to understand how Files & Media populates context
- Task 1.4: Write tests for FileItem with view_mode and custom_slices
Phase 2: Decouple Context Composition from Files & Media
Focus: Remove auto-population inheritance, make Context Composition independent
- Task 2.1: Remove auto-population of context from Files & Media in context composition
- Task 2.2: Add manual "Add Files" button to Context Composition (file picker from project whitelist)
- Task 2.3: Implement "Add All" batch operation
- Task 2.4: Write tests for decoupled context composition state
Phase 3: Directory Grouping + File Stats
Focus: Compact file listing with stats
- Task 3.1: Implement directory grouping helper to group files by relative path prefix
- Task 3.2: Add file stats computation (line count, AST element count) - async
- Task 3.3: Render file list with collapsible directory headers
- Task 3.4: Display aggregate stats (total files, lines, AST elements)
- Task 3.5: Write tests for directory grouping and stats
Phase 4: View Mode Selection UI
Focus: Per-file view mode dropdown (full/sig/def/custom)
- Task 4.1: Add view_mode dropdown to each file entry in Context Composition
- Task 4.2: Implement custom view mode indicator (enabled when custom slices exist)
- Task 4.3: Batch view mode change operations
- Task 4.4: Write tests for view mode selection
Phase 5: Context Presets Infrastructure
Focus: Data structures for save/load (without UI)
- Task 5.1: Create ContextPreset and FileViewPreset data models
- Task 5.2: Implement serialization for context presets (TOML)
- Task 5.3: Write tests for context preset models
Phase 6: Integration + Bug Fixes
Focus: Ensure aggregate respects new view modes, fix any issues
- Task 6.1: Verify aggregate.py respects view_mode when composing context
- Task 6.2: Test with gencpp project files
- Task 6.3: Conductor - User Manual Verification