tianxuan/scripts/build_update.py

#!/usr/bin/env python
"""
Build incremental update package for 天璇 (TianXuan).

Usage:
    python scripts/build_update.py

Detects changes since the last git tag, creates a ZIP package containing
only modified/new files, ready to be applied on the offline machine.

Output: tianxuan_update_<from_version>-<to_version>.zip
"""
import argparse
import hashlib
import json
import os
import subprocess
import sys
import zipfile
from datetime import datetime, timezone
from pathlib import Path

# Add project root to path for sibling imports
PROJECT_DIR = Path(__file__).resolve().parent.parent
sys.path.insert(0, str(PROJECT_DIR))

from scripts.version_utils import (
    get_current_version,
    set_current_version,
    get_latest_tag,
    get_changed_files,
    bump_version,
    is_working_tree_clean,
)

# Files/dirs to exclude from update packages (in addition to .gitignore)
EXCLUDE_PATTERNS = {
    '.git', '.venv', '__pycache__', '*.pyc', '.pytest_cache',
    '.omo', '.codegraph', '.playwright-mcp',
}

GITIGNORE_PATTERNS = {
    'runtime/', '.venv', '*.sqlite3*', 'logs/', '*.log',
}


def should_exclude(path: str) -> bool:
    """Check if a file path should be excluded from the update package."""
    path_lower = path.lower()
    for pat in EXCLUDE_PATTERNS:
        if pat.endswith('/'):
            if path_lower.startswith(pat) or f'/{pat.lower()}' in path_lower:
                return True
        else:
            if path_lower == pat.lower():
                return True
    # Exclude .gitignore patterns that are not in git
    for pat in GITIGNORE_PATTERNS:
        if pat.endswith('/'):
            if path_lower.startswith(pat) or f'/{pat.lower()}' in path_lower:
                return True
        elif path_lower.endswith(pat.lstrip('*')):
            return True
    return False


def compute_sha256(filepath: Path) -> str:
    """Compute SHA256 hash of a file."""
    h = hashlib.sha256()
    with open(filepath, 'rb') as f:
        for chunk in iter(lambda: f.read(65536), b''):
            h.update(chunk)
    return h.hexdigest()


def has_migration_changes(added_modified: list[str]) -> bool:
    """Check if any migration files changed."""
    for path in added_modified:
        if '/migrations/' in path and path.endswith('.py'):
            return True
    return False


def has_tlsdb_changes(added_modified: list[str]) -> bool:
    """Check if TlsDB.csv changed."""
    for path in added_modified:
        if path == 'TlsDB.csv':
            return True
    return False


def has_dependency_changes(added_modified: list[str]) -> bool:
    """Check if pyproject.toml or requirements.txt changed."""
    for path in added_modified:
        if path in ('pyproject.toml', 'requirements.txt', 'uv.lock'):
            return True
    return False


def detect_runtime_changes(project_dir: Path) -> bool:
    """Detect if portable Python runtime packages have changed.
    
    Uses pip list output to snapshot installed packages, comparing
    against the previous build's snapshot. This avoids depending on
    git (which can't track runtime/ as it's gitignored).
    
    Returns True if any package changed.
    """
    python_exe = str(project_dir / 'runtime' / 'python' / 'python.exe')
    if not os.path.exists(python_exe):
        return False  # No runtime installed — skip
    
    # Get current pip list
    result = subprocess.run(
        [python_exe, '-m', 'pip', 'list', '--format=json'],
        capture_output=True, text=True, cwd=str(project_dir), timeout=60
    )
    if result.returncode != 0:
        return False  # Can't determine — assume no change
    
    try:
        current = json.loads(result.stdout)
    except json.JSONDecodeError:
        return False
    
    # Build a dict of name -> version
    current_packages = {pkg['name'].lower(): pkg['version'] for pkg in current}
    
    # Compare with previous snapshot
    snapshot_dir = project_dir / '.update_cache'
    snapshot_dir.mkdir(parents=True, exist_ok=True)
    snapshot_file = snapshot_dir / 'runtime_snapshot.json'
    
    if not snapshot_file.exists():
        # First build — save and report no change
        json.dump(current_packages, open(snapshot_file, 'w'), indent=2)
        return False
    
    try:
        previous = json.loads(open(snapshot_file).read())
    except (json.JSONDecodeError, OSError):
        json.dump(current_packages, open(snapshot_file, 'w'), indent=2)
        return False
    
    # Detect changes
    changed = {}
    for name, version in current_packages.items():
        prev_ver = previous.get(name)
        if prev_ver is None:
            changed[name] = f'+{version}'  # New package
        elif prev_ver != version:
            changed[name] = f'{prev_ver}→{version}'  # Version changed
    
    for name in previous:
        if name not in current_packages:
            changed[name] = '-removed'  # Package removed
    
    # Update snapshot
    json.dump(current_packages, open(snapshot_file, 'w'), indent=2)
    
    if changed:
        print(f"\n[RUNTIME] {len(changed)} 个依赖包变更:")
        for name, diff in sorted(changed.items()):
            print(f"    {name}: {diff}")
        return True
    
    return False


def build_update_package(auto_mode=False):
    print("=" * 50)
    print("  天璇 增量更新包构建工具")
    print("=" * 50)

    # Step 1: Check working tree
    if not is_working_tree_clean():
        if auto_mode:
            # Step A: Auto-commit any uncommitted changes
            print("\n[AUTO] 检测到未提交的变更，自动提交中...")
            subprocess.run(['git', 'add', '-A'], cwd=PROJECT_DIR, check=True)
            status = subprocess.run(
                ['git', 'status', '--short'],
                capture_output=True, text=True, cwd=PROJECT_DIR
            )
            file_count = len([l for l in status.stdout.strip().split('\n') if l.strip()])
            version = get_current_version()
            commit_msg = f"auto: update for v{version}"
            subprocess.run(
                ['git', 'commit', '-m', commit_msg],
                cwd=PROJECT_DIR, check=True
            )
            print(f"  ✅ 已提交 {file_count} 个文件: {commit_msg}")
        else:
            print("\n[ERROR] 工作树有未提交的变更！")
            print("  请先提交或贮藏 (git stash) 后再构建")
            subprocess.run(['git', 'status', '--short'], cwd=PROJECT_DIR)
            sys.exit(1)

    # Step 1.5: Auto-tag (auto mode only)
    if auto_mode:
        current_ver = get_current_version()
        result = subprocess.run(
            ['git', 'tag', '-l', current_ver],
            capture_output=True, text=True, cwd=PROJECT_DIR
        )
        if current_ver not in result.stdout:
            subprocess.run(['git', 'tag', current_ver], cwd=PROJECT_DIR, check=True)
            print(f"  ✅ 已创建 tag: {current_ver}")
        else:
            print(f"  ℹ️  Tag {current_ver} 已存在，跳过")

    # Step 2: Determine versions
    current_version = get_current_version()
    latest_tag = get_latest_tag()

    print(f"\n当前 VERSION: {current_version}")
    print(f"最新 git tag: {latest_tag or '(无)'}")

    # If no tags exist, use VERSION content
    from_version = latest_tag if latest_tag else current_version
    to_version = bump_version(current_version)

    # Step 3: Get changed files
    if latest_tag:
        added_modified, deleted = get_changed_files(latest_tag)
    else:
        # First build — include everything except gitignored/excluded
        print("\n[INFO] 首次构建，包含所有被 git 跟踪的文件")
        result = subprocess.run(
            ['git', 'ls-files'],
            capture_output=True, text=True, cwd=PROJECT_DIR, timeout=30
        )
        added_modified = [f.strip() for f in result.stdout.strip().split('\n') if f.strip()]
        deleted = []

    # Filter excluded
    added_modified = [f for f in added_modified if not should_exclude(f)]
    deleted = [f for f in deleted if not should_exclude(f)]

    if not added_modified and not deleted:
        print("\n[NO_CHANGES] 自上次构建以来没有变更文件")
        print(f"  VERSION: {current_version}")
        sys.exit(0)

    print(f"\n变更文件: {len(added_modified)} 个修改/新增, {len(deleted)} 个删除")

    # Step 4: Build manifest
    files_manifest = {}
    for rel_path in added_modified:
        abs_path = PROJECT_DIR / rel_path
        if abs_path.exists() and abs_path.is_file():
            files_manifest[rel_path] = {
                'sha256': compute_sha256(abs_path),
                'size': abs_path.stat().st_size,
            }

    # Runtime change detection (runtime/ is gitignored, so git can't track it)
    runtime_changed = detect_runtime_changes(PROJECT_DIR)
    if not runtime_changed:
        # Also check if pyproject.toml was modified (even if pip list didn't change yet)
        runtime_changed = has_dependency_changes(added_modified)

    manifest = {
        'from_version': from_version,
        'to_version': to_version,
        'built_at': datetime.now(timezone.utc).strftime('%Y-%m-%dT%H:%M:%SZ'),
        'files': files_manifest,
        'deleted': deleted,
        'run_migrate': has_migration_changes(added_modified),
        'run_import_tlsdb': has_tlsdb_changes(added_modified),
        'runtime_changed': runtime_changed,
    }

    # Step 5: Build ZIP
    zip_name = f'tianxuan_update_{from_version}-{to_version}.zip'
    zip_path = PROJECT_DIR / zip_name

    with zipfile.ZipFile(zip_path, 'w', zipfile.ZIP_DEFLATED) as zf:
        # Write manifest
        zf.writestr('version.json', json.dumps(manifest, indent=2, ensure_ascii=False))

        # Write changed files
        for rel_path in added_modified:
            abs_path = PROJECT_DIR / rel_path
            if abs_path.exists() and abs_path.is_file():
                arcname = f'files/{rel_path}'
                zf.write(abs_path, arcname)

        # Write runtime_changed marker if dependencies changed
        if manifest['runtime_changed']:
            runtime_msg = (
                "[RUNTIME] 依赖已变更！\n"
                f"pyproject.toml 或 requirements.txt 已修改。\n"
                f"请手动将 runtime/ 目录同步到离线机。\n"
            )
            zf.writestr('runtime_changed.txt', runtime_msg)

    # Step 6: Update VERSION
    set_current_version(to_version)

    # Step 7: Print summary
    total_size_mb = zip_path.stat().st_size / (1024 * 1024)
    print(f"\n{'=' * 50}")
    print(f"✅ 增量更新包构建完成！")
    print(f"  包: {zip_name}")
    print(f"  大小: {total_size_mb:.2f} MB")
    print(f"  从: {from_version}")
    print(f"  到: {to_version}")
    print(f"  文件数: {len(files_manifest)}")
    print(f"  删除: {len(deleted)}")
    if manifest['run_migrate']:
        print(f"  ⚠️ 包含数据库迁移 (schema 变更)")
    if manifest['run_import_tlsdb']:
        print(f"  ⚠️ TlsDB.csv 已变更，需重新导入引用数据")
    if manifest['runtime_changed']:
        # Write a visible marker file
        runtime_marker = PROJECT_DIR / 'RUNTIME_CHANGED.txt'
        runtime_marker.write_text(
            "依赖已变更！请手动将 runtime/ 目录同步到离线机。\n"
            f"请在开发机上压缩 runtime/ 目录，然后用 U 盘复制到离线机。\n"
            f"也可直接复制整个 runtime/ 目录覆盖离线机的对应目录。\n",
            encoding='utf-8'
        )
        print(f"  ⚠️ 依赖已变更！请手动将 runtime/ 同步到离线机")
    else:
        # Clean up stale marker if no longer changed
        stale_marker = PROJECT_DIR / 'RUNTIME_CHANGED.txt'
        if stale_marker.exists():
            stale_marker.unlink()
    print(f"\n  复制 {zip_name} 到离线机，然后双击 update.bat")
    print(f"{'=' * 50}")


if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='Build incremental update package for 天璇')
    parser.add_argument('--auto', '-a', action='store_true', help='One-click mode: auto-commit + auto-tag + build')
    args = parser.parse_args()
    build_update_package(auto_mode=args.auto)