feat: CLI & MCP progress tracking for `parse-prd` command #251

New issue

Closed

opened 2025-10-14 15:46:59 -06:00 by navan · 0 comments

navan commented

2025-10-14 15:46:59 -06:00

Owner

Originally created by @joedanz on 7/25/2025

This PR introduces progress-tracking capabilities for the parse-prd command and establishes a standardized foundation for all other long-running AI operations.

Key Features

1. Integrated CLI Progress UI

• Added dedicated progress tracker for PRD parsing: ParsePrdTracker
• Multi-bar displays powered by cli-progress featuring:

Dynamic ETA, elapsed time, and token I/O metrics
Per-task titles with priority/complexity indicators
Live aggregated counters and status information
Real-time tracking with running totals

2. FastMCP-Compliant Progress Events

• All commands emit FastMCP-compliant reportProgress events following the MCP 2025-03-26 specification
• Progress notifications include:

Progress tracking: { progress: number, total: number, message: string }
Task metadata: Titles, priority icons, complexity scores, and token counts
Granular updates: Real-time progress during streaming operations
• Ensures parity between CLI visuals and programmatic MCP feedback

3. Advanced Progress Features

• Real-time subtask counting - Updates during individual subtask generation
• Token tracking - Live monitoring of AI usage
• Intelligent ETAs - Adaptive time estimates based on recent performance
• Status indicators - Visual feedback for success/failure/skipped operations

4. Robust Fallback Logic

• Automatic fallback to non-streaming mode if providers cannot supply async text streams
• Graceful degradation preserves functionality across all environments
• Error recovery maintains progress state during temporary failures

5. Reusable Progress-Tracking Foundation

• Modular tracker classes can be imported by any future long-running command
• Standardized UI components in src/ui/ for consistent visual patterns
• Factory pattern for easy extension to new command types

Why It Matters

– Enhanced user experience with real-time visibility into all lengthy AI operations and detailed metrics
– Standardized patterns for consistent progress experience across long-running Task Master commands
– Scalable architecture supporting future commands with minimal implementation effort

Summary by CodeRabbit

New Features
- Live streaming PRD parsing with CLI/MCP progress, priority/complexity indicators, and rich start/summary displays; structured object streaming for AI responses.
Improvements
- Better cost telemetry (includes currency), role-based provider selection, retries, timeout handling, and quieter CLI (legacy spinner removed).
Chores
- Added runtime deps: @streamparser/json, cli-progress.
Tests
- Expanded integration and unit tests plus progress testing guides and harnesses.

*Originally created by @joedanz on 7/25/2025* This PR introduces progress-tracking capabilities for the `parse-prd` command and establishes a standardized foundation for all other long-running AI operations. ## Key Features ### 1. **Integrated CLI Progress UI** • Added dedicated progress tracker for PRD parsing: `ParsePrdTracker` • Multi-bar displays powered by `cli-progress` featuring: - Dynamic ETA, elapsed time, and token I/O metrics - Per-task titles with priority/complexity indicators - Live aggregated counters and status information - Real-time tracking with running totals ### 2. **FastMCP-Compliant Progress Events** • All commands emit **FastMCP-compliant** `reportProgress` events following the MCP 2025-03-26 specification • Progress notifications include: - **Progress tracking**: `{ progress: number, total: number, message: string }` - **Task metadata**: Titles, priority icons, complexity scores, and token counts - **Granular updates**: Real-time progress during streaming operations • Ensures parity between CLI visuals and programmatic MCP feedback ### 3. **Advanced Progress Features** • **Real-time subtask counting** - Updates during individual subtask generation • **Token tracking** - Live monitoring of AI usage • **Intelligent ETAs** - Adaptive time estimates based on recent performance • **Status indicators** - Visual feedback for success/failure/skipped operations ### 4. **Robust Fallback Logic** • Automatic fallback to non-streaming mode if providers cannot supply async text streams • Graceful degradation preserves functionality across all environments • Error recovery maintains progress state during temporary failures ### 5. **Reusable Progress-Tracking Foundation** • Modular tracker classes can be imported by any future long-running command • Standardized UI components in `src/ui/` for consistent visual patterns • Factory pattern for easy extension to new command types ## Why It Matters – **Enhanced user experience** with real-time visibility into all lengthy AI operations and detailed metrics – **Standardized patterns** for consistent progress experience across long-running Task Master commands – **Scalable architecture** supporting future commands with minimal implementation effort  ## Summary by CodeRabbit * **New Features** * Live streaming PRD parsing with CLI/MCP progress, priority/complexity indicators, and rich start/summary displays; structured object streaming for AI responses. * **Improvements** * Better cost telemetry (includes currency), role-based provider selection, retries, timeout handling, and quieter CLI (legacy spinner removed). * **Chores** * Added runtime deps: @streamparser/json, cli-progress. * **Tests** * Expanded integration and unit tests plus progress testing guides and harnesses.