hopefully done refining

This commit is contained in:
2026-03-06 16:14:31 -05:00
parent 88e27ae414
commit 1294104f7f
20 changed files with 1736 additions and 734 deletions

View File

@@ -1,25 +1,107 @@
# Implementation Plan: Tool Usage Analytics (tool_usage_analytics_20260306)
## Phase 1: Data Collection
- [ ] Task: Initialize MMA Environment
- [ ] Task: Verify tool_log_callback
- WHERE: src/ai_client.py
- WHAT: Check existing logging
> **Reference:** [Spec](./spec.md) | [Architecture Guide](../../../docs/guide_architecture.md)
## Phase 2: Aggregation
- [ ] Task: Implement usage aggregation
- WHERE: src/gui_2.py or new module
- WHAT: Count tools, avg times, failures
- HOW: Process tool_log entries
- SAFETY: Efficient data structures
## Phase 1: Data Collection
Focus: Add tool execution tracking
- [ ] Task 1.1: Initialize MMA Environment
- Run `activate_skill mma-orchestrator` before starting
- [ ] Task 1.2: Add tool stats state
- WHERE: `src/app_controller.py` or `src/gui_2.py`
- WHAT: Add `_tool_stats: dict[str, dict]` state
- HOW:
```python
self._tool_stats: dict[str, dict] = {}
# Structure: {tool_name: {"count": 0, "total_time_ms": 0, "failures": 0}}
```
- CODE STYLE: 1-space indentation
- [ ] Task 1.3: Hook into tool execution
- WHERE: `src/ai_client.py` in tool execution path
- WHAT: Track tool name, time, success/failure
- HOW:
```python
start_time = time.time()
try:
result = mcp_client.dispatch(name, args)
success = True
except Exception:
success = False
finally:
elapsed_ms = (time.time() - start_time) * 1000
# Update stats via callback or direct update
```
- SAFETY: Don't impact tool execution performance
## Phase 2: Aggregation Logic
Focus: Calculate derived metrics
- [ ] Task 2.1: Implement stats update function
- WHERE: `src/app_controller.py`
- WHAT: Function to update tool stats
- HOW:
```python
def _update_tool_stats(self, tool_name: str, elapsed_ms: float, success: bool) -> None:
if tool_name not in self._tool_stats:
self._tool_stats[tool_name] = {"count": 0, "total_time_ms": 0.0, "failures": 0}
self._tool_stats[tool_name]["count"] += 1
self._tool_stats[tool_name]["total_time_ms"] += elapsed_ms
if not success:
self._tool_stats[tool_name]["failures"] += 1
```
- [ ] Task 2.2: Calculate average time and failure rate
- WHERE: `src/gui_2.py` in render function
- WHAT: Derive avg_time and failure_rate from stats
- HOW:
```python
for tool, stats in self._tool_stats.items():
count = stats["count"]
avg_time = stats["total_time_ms"] / count if count > 0 else 0
failure_rate = (stats["failures"] / count * 100) if count > 0 else 0
```
## Phase 3: Visualization
- [ ] Task: Render analytics
- WHERE: src/gui_2.py
- WHAT: Charts and tables
- HOW: imgui tables, plot_lines
- SAFETY: Handle empty data
Focus: Display analytics in GUI
## Phase 4: Verification
- [ ] Task: Test analytics
- [ ] Task: Conductor - Phase Verification
- [ ] Task 3.1: Add analytics panel
- WHERE: `src/gui_2.py` in MMA Dashboard or Operations
- WHAT: Table showing tool stats
- HOW:
```python
if imgui.collapsing_header("Tool Usage Analytics"):
if imgui.begin_table("tool_stats", 4):
imgui.table_setup_column("Tool")
imgui.table_setup_column("Count")
imgui.table_setup_column("Avg Time (ms)")
imgui.table_setup_column("Failure %")
imgui.table_headers_row()
for tool, stats in sorted(self._tool_stats.items(), key=lambda x: -x[1]["count"]):
imgui.table_next_row()
imgui.table_set_column_index(0)
imgui.text(tool)
# ... other columns
imgui.end_table()
```
## Phase 4: Reset on Session Clear
Focus: Clear stats on new session
- [ ] Task 4.1: Clear stats on session reset
- WHERE: `src/gui_2.py` or `src/app_controller.py` reset handler
- WHAT: Clear `_tool_stats` dict
- HOW: `self._tool_stats.clear()`
## Phase 5: Testing
Focus: Verify all functionality
- [ ] Task 5.1: Write unit tests
- WHERE: `tests/test_tool_analytics.py` (new file)
- WHAT: Test stats accumulation, avg calculation
- HOW: Mock tool execution, verify stats update
- [ ] Task 5.2: Conductor - Phase Verification
- Run: `uv run pytest tests/test_tool_analytics.py -v`
- Manual: Verify analytics panel displays in GUI

View File

@@ -1,35 +1,99 @@
# Track Specification: Tool Usage Analytics (tool_usage_analytics_20260306)
## Overview
Analytics panel showing most-used tools, average execution time, failure rates.
Analytics panel showing most-used tools, average execution time, and failure rates. Uses existing tool execution data from ai_client.
## Current State Audit
### Already Implemented
- **`ai_client.tool_log_callback`**: Called when tool executes
- **`mcp_client.dispatch()`**: Routes tool calls
- **No aggregation or storage**
### Already Implemented (DO NOT re-implement)
### Gaps to Fill
- No tool usage tracking
- No execution time tracking
#### Tool Execution (src/ai_client.py)
- **Tool dispatch in `_execute_tool_calls_concurrently()`**: Executes tools via `mcp_client.dispatch()`
- **`pre_tool_callback`**: Optional callback before tool execution
- **No built-in tracking or aggregation**
#### MCP Client (src/mcp_client.py)
- **`dispatch(name, args)`**: Routes tool calls to implementations
- **26 tools available** (run_powershell, read_file, py_get_skeleton, etc.)
- **`MUTATING_TOOLS`**: Set of tools that modify files
### Gaps to Fill (This Track's Scope)
- No tool usage tracking (count per tool)
- No execution time tracking per tool
- No failure rate tracking
- No analytics display in GUI
## Architectural Constraints
### Efficient Aggregation
- Track tool stats in lightweight data structure
- Don't impact tool execution performance
- Use dict: `{tool_name: {count, total_time, failures}}`
### Memory Bounds
- Only track stats, not full history
- Reset on session reset
## Architecture Reference
### Key Integration Points
| File | Lines | Purpose |
|------|-------|---------|
| `src/ai_client.py` | ~500-550 | Tool execution - add tracking |
| `src/gui_2.py` | ~2700-2800 | Analytics panel |
### Proposed Tracking Structure
```python
# In AppController or App:
self._tool_stats: dict[str, dict] = {}
# Structure: {"read_file": {"count": 10, "total_time_ms": 150, "failures": 0}, ...}
```
## Functional Requirements
- Track tool name, execution time, success/failure
- Aggregate by tool name
- Display ranking by usage count
- Show average time per tool
- Show failure percentage
## Key Integration Points
| File | Purpose |
|-----|---------|
| `src/ai_client.py` | Hook into tool_log_callback |
| `src/gui_2.py` | Analytics panel rendering |
### FR1: Tool Usage Tracking
- Track tool name, execution time, success/failure
- Store in `_tool_stats` dict
- Update on each tool execution
### FR2: Aggregation by Tool
- Count total calls per tool
- Calculate average execution time
- Track failure count and rate
### FR3: Analytics Display
- Table showing tool name, count, avg time, failure rate
- Sort by usage count (most used first)
- Show in MMA Dashboard or Operations panel
## Non-Functional Requirements
| Requirement | Constraint |
|-------------|------------|
| Tracking Overhead | <1ms per tool call |
| Memory | <1KB for stats dict |
## Testing Requirements
### Unit Tests
- Test tracking updates correctly
- Test failure rate calculation
### Integration Tests
- Execute tools, verify stats accumulate
- Reset session, verify stats cleared
## Out of Scope
- Historical analytics across sessions
- Export to file
- Per-ticket tool breakdown
## Acceptance Criteria
- [ ] Tool ranking displayed
- [ ] Average times accurate
- [ ] Failure rates tracked
- [ ] 1-space indentation
- [ ] Tool execution tracked
- [ ] Count per tool accurate
- [ ] Average time calculated
- [ ] Failure rate shown
- [ ] Display in GUI panel
- [ ] Reset on session clear
- [ ] 1-space indentation maintained