@pennyfarthing/core 7.9.0 → 7.9.2

package/pennyfarthing-dist/scripts/sprint/sprint-status.sh

@@ -1,134 +0,0 @@
-#!/bin/bash
-# Display current sprint status
-# Usage: .pennyfarthing/scripts/core/run.sh sprint/sprint-status.sh [filter]
-#    or: Invoked with PROJECT_ROOT already set
-#
-# Filters:
-#   (none)       - Show all stories
-#   todo         - Show only backlog stories
-#   in-progress  - Show only in_progress stories
-#   done         - Show only done/cancelled stories
-
-set -euo pipefail
-
-# Load shared sprint functions
-SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
-source "$SCRIPT_DIR/sprint-common.sh"
-
-# Parse filter argument
-FILTER="${1:-}"
-case "$FILTER" in
-  todo|backlog)
-    STATUS_FILTER="backlog"
-    FILTER_LABEL="Todo"
-    ;;
-  in-progress|wip)
-    STATUS_FILTER="in_progress"
-    FILTER_LABEL="In Progress"
-    ;;
-  done|completed)
-    STATUS_FILTER="done"
-    FILTER_LABEL="Done"
-    ;;
-  cancelled|canceled)
-    STATUS_FILTER="cancelled"
-    FILTER_LABEL="Cancelled"
-    ;;
-  "")
-    STATUS_FILTER=""
-    FILTER_LABEL=""
-    ;;
-  *)
-    echo "Error: Unknown filter '$FILTER'"
-    echo "Valid filters: todo, in-progress, done, cancelled"
-    exit 1
-    ;;
-esac
-
-SPRINT_FILE=$(get_sprint_file)
-
-if [[ ! -f "$SPRINT_FILE" ]]; then
-  echo "Error: Sprint file not found at $SPRINT_FILE"
-  exit 1
-fi
-
-check_yq || exit 1
-
-# Extract sprint metadata using shared function pattern
-SPRINT_NUM=$(get_sprint_metadata "number")
-SPRINT_NAME="TO Sprint $SPRINT_NUM"
-SPRINT_GOAL=$(get_sprint_metadata "goal")
-START_DATE=$(get_sprint_metadata "start_date")
-END_DATE=$(get_sprint_metadata "end_date")
-STATUS=$(get_sprint_metadata "status")
-
-echo "# Sprint Status"
-echo ""
-echo "**Sprint:** $SPRINT_NAME"
-echo "**Goal:** $SPRINT_GOAL"
-echo "**Dates:** $START_DATE to $END_DATE"
-echo "**Status:** $STATUS"
-echo ""
-
-# Get counts using shared functions
-STORY_COUNTS=$(get_story_counts)
-BACKLOG=$(echo "$STORY_COUNTS" | grep -o 'backlog:[0-9]*' | cut -d: -f2)
-IN_PROGRESS=$(echo "$STORY_COUNTS" | grep -o 'in_progress:[0-9]*' | cut -d: -f2)
-TOTAL_STORIES=$((BACKLOG + IN_PROGRESS))
-
-POINT_COUNTS=$(get_point_counts)
-BACKLOG_POINTS=$(echo "$POINT_COUNTS" | grep -o 'backlog:[0-9]*' | cut -d: -f2)
-IN_PROGRESS_POINTS=$(echo "$POINT_COUNTS" | grep -o 'in_progress:[0-9]*' | cut -d: -f2)
-TOTAL_POINTS=$(echo "$POINT_COUNTS" | grep -o 'total:[0-9]*' | cut -d: -f2)
-
-echo "## Summary"
-echo ""
-echo "| Status | Stories | Points |"
-echo "|--------|---------|--------|"
-echo "| Backlog | $BACKLOG | $BACKLOG_POINTS |"
-echo "| In Progress | $IN_PROGRESS | $IN_PROGRESS_POINTS |"
-echo "| **Total Remaining** | **$TOTAL_STORIES** | **$TOTAL_POINTS** |"
-echo ""
-
-# List stories by epic (grouped - epic name appears once)
-if [[ -n "$FILTER_LABEL" ]]; then
-  echo "## Stories by Epic (${FILTER_LABEL} only)"
-else
-  echo "## Stories by Epic"
-fi
-echo ""
-
-# Process each epic, output header once then all its stories
-# Apply status filter if specified
-if [[ -n "$STATUS_FILTER" ]]; then
-  yq eval -o=json '.epics[]' "$SPRINT_FILE" 2>/dev/null | jq -r --arg status "$STATUS_FILTER" '
-    .title as $title |
-    [.stories[] | select(.status == $status)] |
-    if length > 0 then
-      "### " + $title + "\n" +
-      (map("- [" + .status + "] " + .id + ": " + .title + " (" + (.points | tostring) + " pts)") | join("\n")) + "\n"
-    else
-      empty
-    end
-  ' 2>/dev/null || echo "No stories found"
-else
-  yq eval -o=json '.epics[]' "$SPRINT_FILE" 2>/dev/null | jq -r '
-    "### " + .title + "\n" +
-    (.stories | map("- [" + .status + "] " + .id + ": " + .title + " (" + (.points | tostring) + " pts)") | join("\n")) + "\n"
-  ' 2>/dev/null || echo "No stories found"
-fi
-
-# Check archive for completed count
-SPRINT_NUM=$(echo "$SPRINT_NAME" | sed 's/TO Sprint //')
-ARCHIVE_FILE="$PROJECT_ROOT/sprint/archive/sprint-${SPRINT_NUM}-completed.yaml"
-
-if [[ -f "$ARCHIVE_FILE" ]]; then
-  COMPLETED_COUNT=$(yq '.completed | length' "$ARCHIVE_FILE" 2>/dev/null || echo "0")
-  COMPLETED_POINTS=$(sum_points "$(yq '.completed[].points' "$ARCHIVE_FILE" 2>/dev/null)")
-  echo ""
-  echo "## Completed This Sprint"
-  echo ""
-  echo "**Stories:** $COMPLETED_COUNT | **Points:** $COMPLETED_POINTS"
-  echo ""
-  echo "See: sprint/archive/sprint-${SPRINT_NUM}-completed.yaml"
-fi

package/pennyfarthing-dist/scripts/story/README.md

@@ -1,23 +0,0 @@
-# Story Scripts
-
-Scripts for story creation and management.
-
-## Scripts
-
-| Script | Purpose |
-|--------|---------|
-| `create-story.sh` | Create a new story in sprint YAML |
-| `size-story.sh` | Estimate story points |
-| `story-template.sh` | Generate story template |
-
-## Usage
-
-```bash
-.pennyfarthing/scripts/core/run.sh story/create-story.sh
-.pennyfarthing/scripts/core/run.sh story/size-story.sh
-```
-
-## Ownership
-
-- **Primary users:** SM agent, `/story` skill
-- **Maintained by:** Core Pennyfarthing team

package/pennyfarthing-dist/scripts/story/create-story.sh

@@ -1,19 +0,0 @@
-#!/usr/bin/env zsh
-# Generate a story YAML block for adding to sprint
-# Usage: create-story.sh <epic-id> "<title>" <points> [options]
-#   --type bug|feature|refactor|chore  Story type (default: feature)
-#   --workflow <name>                  Workflow override
-#   --priority P0|P1|P2|P3             Priority (default: P2)
-#
-# Thin wrapper that delegates to Python CLI:
-#   python -m pennyfarthing_scripts.story create <epic-id> "<title>" <points> [options]
-
-set -e
-
-# Source common functions for Python discovery
-SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
-PARENT_DIR="$(dirname "$SCRIPT_DIR")"
-source "${PARENT_DIR}/lib/common.sh"
-
-# Delegate to Python CLI
-run_python_module story create "$@"

package/pennyfarthing-dist/scripts/story/size-story.sh

@@ -1,18 +0,0 @@
-#!/usr/bin/env zsh
-# Display story sizing guidelines
-# Usage: size-story.sh [points]
-#   No args: Show all sizing guidelines
-#   points:  Show specific guidance for that point value
-#
-# Thin wrapper that delegates to Python CLI:
-#   python -m pennyfarthing_scripts.story size [points]
-
-set -e
-
-# Source common functions for Python discovery
-SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
-PARENT_DIR="$(dirname "$SCRIPT_DIR")"
-source "${PARENT_DIR}/lib/common.sh"
-
-# Delegate to Python CLI
-run_python_module story size "$@"

package/pennyfarthing-dist/scripts/story/story-template.sh

@@ -1,18 +0,0 @@
-#!/usr/bin/env zsh
-# Show story templates by type
-# Usage: story-template.sh [bug|feature|refactor|chore]
-#   No args: Show all templates
-#   type:    Show specific template
-#
-# Thin wrapper that delegates to Python CLI:
-#   python -m pennyfarthing_scripts.story template [type]
-
-set -e
-
-# Source common functions for Python discovery
-SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
-PARENT_DIR="$(dirname "$SCRIPT_DIR")"
-source "${PARENT_DIR}/lib/common.sh"
-
-# Delegate to Python CLI
-run_python_module story template "$@"

package/pennyfarthing-dist/scripts/test/README.md

@@ -1,23 +0,0 @@
-# Test Scripts
-
-Scripts for test infrastructure and benchmarking.
-
-## Scripts
-
-| Script | Purpose |
-|--------|---------|
-| `test-setup.sh` | Set up test environment |
-| `test-cache.sh` | Manage test result caching |
-| `ground-truth-judge.py` | Python judge for ground truth evaluation |
-| `swebench-judge.py` | SWE-bench style evaluation |
-
-## Usage
-
-```bash
-.pennyfarthing/scripts/core/run.sh test/test-setup.sh
-```
-
-## Ownership
-
-- **Primary users:** TEA agent, testing-runner subagent
-- **Maintained by:** Core Pennyfarthing team

package/pennyfarthing-dist/scripts/test/ensure-swebench-data.sh

@@ -1,59 +0,0 @@
-#!/usr/bin/env bash
-# ensure-swebench-data.sh - Downloads SWE-bench data if not present
-#
-# Usage: ensure-swebench-data.sh [--force]
-#
-# Downloads SWE-bench Verified dataset from HuggingFace to /tmp/swebench_all.json
-# This is a dependency for:
-#   - swebench-judge.py
-#   - ground-truth-judge.py
-#
-# Options:
-#   --force    Re-download even if file exists
-
-set -euo pipefail
-
-CACHE_PATH="/tmp/swebench_all.json"
-DATASET_URL="https://huggingface.co/datasets/princeton-nlp/SWE-bench_Verified/resolve/main/data/test.jsonl"
-
-force=false
-if [[ "${1:-}" == "--force" ]]; then
-    force=true
-fi
-
-# Check if already present
-if [[ -f "$CACHE_PATH" ]] && [[ "$force" == "false" ]]; then
-    echo "SWE-bench data already cached at $CACHE_PATH"
-    exit 0
-fi
-
-echo "Downloading SWE-bench Verified dataset..."
-
-# Download JSONL and convert to JSON array
-if command -v curl &>/dev/null; then
-    curl -sL "$DATASET_URL" | python3 -c "
-import json
-import sys
-lines = [json.loads(line) for line in sys.stdin if line.strip()]
-print(json.dumps(lines, indent=2))
-" > "$CACHE_PATH"
-elif command -v wget &>/dev/null; then
-    wget -qO- "$DATASET_URL" | python3 -c "
-import json
-import sys
-lines = [json.loads(line) for line in sys.stdin if line.strip()]
-print(json.dumps(lines, indent=2))
-" > "$CACHE_PATH"
-else
-    echo "Error: curl or wget required to download SWE-bench data"
-    exit 1
-fi
-
-# Verify download
-if [[ -f "$CACHE_PATH" ]]; then
-    count=$(python3 -c "import json; print(len(json.load(open('$CACHE_PATH'))))")
-    echo "Downloaded $count SWE-bench scenarios to $CACHE_PATH"
-else
-    echo "Error: Failed to download SWE-bench data"
-    exit 1
-fi

package/pennyfarthing-dist/scripts/test/ground-truth-judge.py

@@ -1,220 +0,0 @@
-#!/usr/bin/env python3
-"""
-Ground-truth judge for SWE-bench scenarios.
-
-Compares Claude's proposed fix against the actual SWE-bench patch.
-Scores based on:
-- File identification (20%)
-- Function/location identification (20%)
-- Fix logic match (40%)
-- Completeness (20%)
-"""
-
-import json
-import re
-import sys
-from pathlib import Path
-from difflib import SequenceMatcher
-
-# Add parent to path for pennyfarthing_scripts imports
-sys.path.insert(0, str(Path(__file__).resolve().parents[3]))
-
-from pennyfarthing_scripts.swebench import (
-    extract_patch_info,
-    extract_problem_keywords,
-    find_scenario,
-    get_meaningful_patterns,
-    load_swebench_data,
-)
-
-
-def score_response(response_text, ground_truth):
-    """Score a response against ground truth patch."""
-    patch_info = extract_patch_info(ground_truth['patch'])
-
-    scores = {
-        'file_identification': 0,
-        'location_identification': 0,
-        'fix_logic_match': 0,
-        'completeness': 0,
-        'details': {}
-    }
-
-    response_lower = response_text.lower()
-
-    # 1. FILE IDENTIFICATION (20 points)
-    files_found = 0
-    for f in patch_info.files:
-        # Check various forms of the filename
-        filename = Path(f).name
-        if filename.lower() in response_lower or f.lower() in response_lower:
-            files_found += 1
-
-    if patch_info.files:
-        file_score = (files_found / len(patch_info.files)) * 20
-        scores['file_identification'] = min(20, file_score)
-        scores['details']['files_expected'] = patch_info.files
-        scores['details']['files_found'] = files_found
-    else:
-        scores['file_identification'] = 20  # No specific file in patch
-
-    # 2. LOCATION IDENTIFICATION (20 points)
-    # Look for function/class names mentioned in the patch
-    locations_found = 0
-    for func in patch_info.functions:
-        # Extract the function/class name
-        func_match = re.search(r'(def|class)\s+(\w+)', func)
-        if func_match:
-            func_name = func_match.group(2)
-            if func_name.lower() in response_lower:
-                locations_found += 1
-        elif func.strip() and func.strip().split()[0] in response_lower:
-            locations_found += 1
-
-    if patch_info.functions:
-        loc_score = (locations_found / len(patch_info.functions)) * 20
-        scores['location_identification'] = min(20, loc_score)
-        scores['details']['locations_expected'] = patch_info.functions[:3]
-        scores['details']['locations_found'] = locations_found
-    else:
-        scores['location_identification'] = 10  # Partial credit
-
-    # 3. FIX LOGIC MATCH (40 points)
-    # Check if key code patterns from the fix appear in the response
-    meaningful_patterns = get_meaningful_patterns(patch_info.key_patterns)
-
-    patterns_found = 0
-    for pattern in meaningful_patterns:
-        if pattern.lower() in response_lower:
-            patterns_found += 1
-
-    if meaningful_patterns:
-        pattern_score = (patterns_found / len(meaningful_patterns)) * 20
-        scores['details']['patterns_expected'] = meaningful_patterns[:10]
-        scores['details']['patterns_found'] = patterns_found
-    else:
-        pattern_score = 10
-
-    # Check for actual code additions
-    additions_matched = 0
-    for addition in patch_info.additions[:5]:  # Check first 5 additions
-        # Normalize and check
-        addition_normalized = re.sub(r'\s+', ' ', addition.lower())
-        response_normalized = re.sub(r'\s+', ' ', response_lower)
-
-        # Use fuzzy matching
-        similarity = SequenceMatcher(None, addition_normalized, response_normalized).ratio()
-        if similarity > 0.6 or addition_normalized in response_normalized:
-            additions_matched += 1
-
-    if patch_info.additions:
-        addition_score = (additions_matched / min(5, len(patch_info.additions))) * 20
-        scores['details']['additions_matched'] = additions_matched
-    else:
-        addition_score = 10
-
-    scores['fix_logic_match'] = min(40, pattern_score + addition_score)
-
-    # 4. COMPLETENESS (20 points)
-    # Does the response have all the elements of a good fix?
-    completeness_score = 0
-
-    # Has code block?
-    if '```' in response_text:
-        completeness_score += 5
-
-    # Has test considerations?
-    if 'test' in response_lower:
-        completeness_score += 5
-
-    # Mentions the specific error/issue?
-    problem_keywords = extract_problem_keywords(ground_truth.get('problem_statement', ''))
-    keywords_found = sum(1 for kw in problem_keywords if kw.lower() in response_lower)
-    if problem_keywords:
-        completeness_score += min(5, (keywords_found / len(problem_keywords)) * 5)
-    else:
-        completeness_score += 2.5
-
-    # Has explanation of why fix works?
-    explanation_words = ['because', 'this fixes', 'this resolves', 'the issue', 'the problem', 'solution']
-    if any(word in response_lower for word in explanation_words):
-        completeness_score += 5
-
-    scores['completeness'] = min(20, completeness_score)
-
-    # Total
-    scores['total'] = round(
-        scores['file_identification'] +
-        scores['location_identification'] +
-        scores['fix_logic_match'] +
-        scores['completeness']
-    , 1)
-
-    return scores
-
-
-def main():
-    if len(sys.argv) < 3:
-        print("Usage: ground-truth-judge.py <scenario_name> <response_file>")
-        print("Example: ground-truth-judge.py flask-5014 run_20260102T134237Z.json")
-        sys.exit(1)
-
-    scenario_name = sys.argv[1]
-    response_file = sys.argv[2]
-
-    # Load SWE-bench data
-    swebench_data = load_swebench_data()
-
-    # Find scenario
-    scenario = find_scenario(swebench_data, scenario_name)
-    if not scenario:
-        print(f"Error: Scenario '{scenario_name}' not found in SWE-bench data")
-        sys.exit(1)
-
-    # Load response
-    with open(response_file, 'r') as f:
-        response_data = json.load(f)
-
-    response_text = response_data.get('result', '')
-    if not response_text:
-        print("Error: No 'result' field in response file")
-        sys.exit(1)
-
-    # Score
-    scores = score_response(response_text, scenario)
-
-    # Output
-    print(f"\n{'='*60}")
-    print(f"GROUND TRUTH EVALUATION: {scenario_name}")
-    print(f"{'='*60}")
-    print(f"\nScores:")
-    print(f"  File Identification:     {scores['file_identification']:5.1f}/20")
-    print(f"  Location Identification: {scores['location_identification']:5.1f}/20")
-    print(f"  Fix Logic Match:         {scores['fix_logic_match']:5.1f}/40")
-    print(f"  Completeness:            {scores['completeness']:5.1f}/20")
-    print(f"  {'─'*40}")
-    print(f"  TOTAL:                   {scores['total']:5.1f}/100")
-
-    print(f"\nDetails:")
-    for key, value in scores['details'].items():
-        print(f"  {key}: {value}")
-
-    # Output JSON for programmatic use
-    output = {
-        'scenario': scenario_name,
-        'instance_id': scenario.get('instance_id'),
-        'scores': scores,
-        'ground_truth_patch_preview': scenario.get('patch', '')[:300]
-    }
-
-    # Save judge output
-    output_path = response_file.replace('run_', 'gt_judge_')
-    with open(output_path, 'w') as f:
-        json.dump(output, f, indent=2)
-    print(f"\nSaved to: {output_path}")
-
-    return scores
-
-
-if __name__ == '__main__':
-    main()