scripts/parse-phases.py - master - PavelNikolaeW/worms-wars

scripts/parse-phases.py
82 строки2 KB
Павел Николаев
EPIC-02 phase-1: safety commit
31 янв 2026, 16:14
31 янв 2026, 16:14132b965
#!/usr/bin/env python3
"""
parse-phases.py — парсит plan.md и извлекает фазы.

Использование:
    python3 scripts/parse-phases.py <path-to-plan.md>

Вывод: JSON-массив в stdout:
[
  {"number": 1, "name": "...", "stories": ["US-001"], "raw_markdown": "..."},
  ...
]
"""

import json
import re
import sys


def parse_phases(plan_path: str) -> list[dict]:
    with open(plan_path) as f:
        content = f.read()

    # Pattern: ### Фаза N: <name> (US-XXX, US-YYY)
    # Also handles the variant without parenthesized stories
    phase_pattern = re.compile(
        r'^###\s+Фаза\s+(\d+)\s*:\s*(.+?)(?:\s*\(([^)]+)\))?\s*$',
        re.MULTILINE,
    )

    matches = list(phase_pattern.finditer(content))
    if not matches:
        print("Ошибка: фазы не найдены в plan.md", file=sys.stderr)
        sys.exit(1)

    phases = []
    for i, match in enumerate(matches):
        number = int(match.group(1))
        name = match.group(2).strip()

        # Extract stories from parentheses
        stories_str = match.group(3) or ""
        stories = []
        if stories_str:
            stories = [s.strip() for s in stories_str.split(",") if s.strip()]

        # Extract raw markdown: from this header to the next phase header (or end)
        start = match.start()
        if i + 1 < len(matches):
            end = matches[i + 1].start()
        else:
            # Until the next ## header or end of file
            next_h2 = re.search(r'^## ', content[match.end():], re.MULTILINE)
            if next_h2:
                end = match.end() + next_h2.start()
            else:
                end = len(content)

        raw_markdown = content[start:end].rstrip()

        phases.append({
            "number": number,
            "name": name,
            "stories": stories,
            "raw_markdown": raw_markdown,
        })

    return phases


def main():
    if len(sys.argv) != 2:
        print(f"Использование: {sys.argv[0]} <plan.md>", file=sys.stderr)
        sys.exit(1)

    plan_path = sys.argv[1]
    phases = parse_phases(plan_path)
    print(json.dumps(phases, ensure_ascii=False, indent=2))


if __name__ == "__main__":
    main()