30 lines
935 B
Markdown
30 lines
935 B
Markdown
# Sprint 193 Plan: Intake Requirement Extraction Hardening II
|
|
|
|
## Context
|
|
Extraction fixes need robust test corpus and regression protection.
|
|
|
|
## Goals
|
|
1. Add intake regression corpus for real project specs
|
|
2. Track per-section recall/precision
|
|
3. Add strict intake gate for low recall
|
|
|
|
## Steps
|
|
|
|
### Step 2004: Intake regression corpus loader (10 tests)
|
|
Load fixture specs with labeled requirement outputs.
|
|
|
|
### Step 2005: Per-section recall/precision metrics (10 tests)
|
|
Compute deterministic extraction quality metrics.
|
|
|
|
### Step 2006: Strict intake quality gate (8 tests)
|
|
Block downstream generation on low requirement recall.
|
|
|
|
### Step 2007: Intake quality summary export (8 tests)
|
|
Emit quality metrics into pipeline summaries.
|
|
|
|
### Step 2008: Sprint 193 Integration Summary (8 tests)
|
|
Add `editor/src/Sprint193IntegrationSummary.h`.
|
|
|
|
## Architecture Gate
|
|
- Intake quality must be measurable and enforceable before task generation.
|