@trac3r/oh-my-god 2.2.11

package/tools/commit_splitter.py

@@ -0,0 +1,749 @@
+#!/usr/bin/env python3
+"""
+AI Commit Splitter for OMG
+
+Analyzes git changes and groups them into logical atomic commits
+with hunk-level staging support. Read-only analysis — never runs git commit.
+
+Feature flag: OMG_AI_COMMIT_ENABLED (default: False)
+"""
+
+import json
+import os
+import re
+import shlex
+import subprocess
+import sys
+from typing import Any, Dict, List, Optional
+
+# Lazy imports for git_inspector and feature flags
+_git_inspector = None
+_get_feature_flag = None
+
+
+def _ensure_imports():
+    """Lazy import git_inspector and feature flag helper."""
+    global _git_inspector, _get_feature_flag
+    if _git_inspector is None:
+        sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+        from tools import git_inspector as _gi
+        from hooks._common import get_feature_flag as _gff
+        _git_inspector = _gi
+        _get_feature_flag = _gff
+
+
+def _is_enabled() -> bool:
+    """Check if AI commit splitter feature is enabled."""
+    _ensure_imports()
+    if _get_feature_flag is None:
+        return False
+    return bool(_get_feature_flag("GIT_WORKFLOW", default=False))
+
+
+# --- File Classification ---
+
+# Extension → category mapping
+_EXT_CATEGORY = {
+    # Python
+    ".py": "python",
+    ".pyi": "python",
+    ".pyx": "python",
+    # JavaScript/TypeScript
+    ".js": "javascript",
+    ".jsx": "javascript",
+    ".ts": "javascript",
+    ".tsx": "javascript",
+    ".mjs": "javascript",
+    ".cjs": "javascript",
+    # Config
+    ".json": "config",
+    ".yaml": "config",
+    ".yml": "config",
+    ".toml": "config",
+    ".ini": "config",
+    ".cfg": "config",
+    ".env": "config",
+    ".conf": "config",
+    ".properties": "config",
+    # Docs
+    ".md": "docs",
+    ".rst": "docs",
+    ".txt": "docs",
+    ".adoc": "docs",
+    # Shell
+    ".sh": "shell",
+    ".bash": "shell",
+    ".zsh": "shell",
+    # CSS/Styles
+    ".css": "styles",
+    ".scss": "styles",
+    ".less": "styles",
+    ".sass": "styles",
+    # HTML/Templates
+    ".html": "markup",
+    ".htm": "markup",
+    ".xml": "markup",
+    ".svg": "markup",
+}
+
+# Test path indicators
+_TEST_INDICATORS = (
+    "test_",
+    "_test.",
+    "tests/",
+    "test/",
+    "__tests__/",
+    ".test.",
+    ".spec.",
+    "spec/",
+    "conftest.py",
+)
+
+# Category → default suggested commit type
+_CATEGORY_DEFAULT_TYPE = {
+    "python": "feat",
+    "javascript": "feat",
+    "shell": "chore",
+    "config": "chore",
+    "docs": "docs",
+    "styles": "style",
+    "markup": "feat",
+    "tests": "test",
+    "other": "chore",
+}
+
+_QUALITY_STEPS = ["format", "lint", "typecheck", "test"]
+
+_ALLOWED_PREFIXES = [
+    ("npm", "test"),
+    ("yarn", "test"),
+    ("pnpm", "test"),
+    ("bun", "test"),
+    ("npx", "--no-install", "prettier"),
+    ("npx", "--no-install", "eslint"),
+    ("npx", "--no-install", "tsc"),
+    ("npx", "--no-install", "jest"),
+    ("npx", "--no-install", "vitest"),
+    ("npx", "--no-install", "biome"),
+    ("jest",),
+    ("vitest",),
+    ("eslint",),
+    ("prettier",),
+    ("tsc",),
+    ("biome",),
+    ("pytest",),
+    ("python", "-m", "pytest"),
+    ("python3", "-m", "pytest"),
+    ("ruff",),
+    ("mypy",),
+    ("flake8",),
+    ("black",),
+    ("isort",),
+    ("bandit",),
+    ("pylint",),
+    ("go", "test"),
+    ("go", "vet"),
+    ("go", "build"),
+    ("golangci-lint",),
+    ("cargo", "test"),
+    ("cargo", "check"),
+    ("cargo", "build"),
+    ("cargo", "clippy"),
+    ("cargo", "fmt"),
+    ("shellcheck",),
+]
+
+_BLOCKED_PATTERNS = [
+    "&&", "||", "|", ";", "`", "$(", "${", ">", "<", "\n",
+    "rm ", "curl ", "wget ", "eval ", "exec ", "sudo ",
+]
+
+_CONVENTIONAL_COMMIT_RE = re.compile(r"^(feat|fix|chore|refactor|docs|test|ci|style|perf|build)(\([^)]+\))?: .+")
+
+
+def _classify_file(file_path: str) -> str:
+    """Classify a file path into a category.
+
+    Test files are always classified as 'tests' regardless of extension.
+
+    Args:
+        file_path: Relative file path from git diff.
+
+    Returns:
+        Category string: 'python', 'javascript', 'config', 'docs',
+        'tests', 'shell', 'styles', 'markup', or 'other'.
+    """
+    if file_path is None:
+        return "other"
+
+    lower_path = file_path.lower()
+
+    # Check for test files first — they always get their own group
+    for indicator in _TEST_INDICATORS:
+        if indicator in lower_path:
+            return "tests"
+
+    # Classify by extension
+    _, ext = os.path.splitext(lower_path)
+    return _EXT_CATEGORY.get(ext, "other")
+
+
+def _derive_scope(files: List[str]) -> str:
+    """Derive a scope name from a list of files.
+
+    Uses the most common parent directory or module name.
+
+    Args:
+        files: List of file paths.
+
+    Returns:
+        Scope string suitable for conventional commit format.
+    """
+    if not files:
+        return "general"
+
+    # Collect directory components
+    dirs: List[str] = []
+    for f in files:
+        parts = f.replace("\\", "/").split("/")
+        if len(parts) > 1:
+            dirs.append(parts[0])
+        else:
+            # Single file at root — use filename without extension
+            name, _ = os.path.splitext(parts[0])
+            dirs.append(name)
+
+    if not dirs:
+        return "general"
+
+    # Most common directory
+    from collections import Counter
+    counts = Counter(dirs)
+    scope = counts.most_common(1)[0][0]
+    return scope
+
+
+def _derive_scope_from_prefix(files: List[str], prefix_dirs: tuple[str, ...]) -> str:
+    """Derive scope from subdirectory after a type-matched prefix."""
+    from collections import Counter
+
+    sub_dirs: List[str] = []
+    for f in files:
+        norm = f.replace("\\", "/")
+        matched_prefix = ""
+        for prefix in prefix_dirs:
+            if norm.startswith(prefix):
+                matched_prefix = prefix
+                break
+
+        if matched_prefix:
+            remainder = norm[len(matched_prefix):]
+            parts = remainder.split("/")
+            if len(parts) > 1:
+                sub_dirs.append(parts[0])
+            else:
+                sub_dirs.append(matched_prefix.rstrip("/"))
+        else:
+            parts = norm.split("/")
+            if len(parts) > 1:
+                sub_dirs.append(parts[0])
+            else:
+                name, _ = os.path.splitext(parts[0])
+                sub_dirs.append(name)
+
+    if not sub_dirs:
+        return "general"
+
+    counts = Counter(sub_dirs)
+    return counts.most_common(1)[0][0]
+
+
+# --- Commit type detection by path ---
+
+# Prefix directories → commit type
+_PATH_PREFIX_TYPE = {
+    "src/": "feat",
+    "lib/": "feat",
+    "app/": "feat",
+    "hooks/": "feat",
+    "tests/": "test",
+    "test/": "test",
+    "spec/": "test",
+    "docs/": "docs",
+}
+
+# Fix-related path keywords
+_FIX_KEYWORDS = ("fix", "bug", "patch", "hotfix")
+
+# Config file extensions (checked when no prefix matches)
+_CONFIG_EXTENSIONS = (".json", ".yaml", ".yml", ".toml", ".cfg")
+
+# Config file name patterns (basename startswith or exact match)
+_CONFIG_BASENAMES = ("setup.", "makefile")
+
+
+def _detect_commit_type(file_path: str) -> str:
+    """Detect commit type from file path. Priority: fix keywords → prefix → docs → config → chore."""
+    norm = file_path.replace("\\", "/").lower()
+
+    for keyword in _FIX_KEYWORDS:
+        if ("/" + keyword) in norm or norm.startswith(keyword):
+            return "fix"
+
+    for prefix, commit_type in _PATH_PREFIX_TYPE.items():
+        if norm.startswith(prefix):
+            return commit_type
+
+    basename = os.path.basename(norm)
+    if basename.startswith("readme") or basename == "changelog.md":
+        return "docs"
+    _, ext = os.path.splitext(norm)
+    if ext == ".md":
+        return "docs"
+
+    if ext in _CONFIG_EXTENSIONS:
+        return "chore"
+    for config_name in _CONFIG_BASENAMES:
+        if basename.startswith(config_name) or basename == config_name.rstrip("."):
+            return "chore"
+
+    return "chore"
+
+
+def _detect_type_majority(files: List[str]) -> str:
+    """Detect commit type by majority vote across files."""
+    if not files:
+        return "chore"
+
+    from collections import Counter
+    types = [_detect_commit_type(f) for f in files]
+    counts = Counter(types)
+    return counts.most_common(1)[0][0]
+
+
+def _prefix_dirs_for_type(commit_type: str) -> tuple:
+    result = [p for p, t in _PATH_PREFIX_TYPE.items() if t == commit_type]
+    return tuple(result) if result else ("",)
+
+
+def generate_commit_message(diff_stats: Dict[str, Any]) -> str:
+    """Generate conventional commit message: type(scope): description.
+
+    Args:
+        diff_stats: Dict with ``files`` (list[str]), optional ``description``
+            (str), and optional ``breaking_change`` (str).
+    """
+    files: List[str] = diff_stats.get("files", [])
+    description: str = diff_stats.get("description", "")
+    breaking_change: str = diff_stats.get("breaking_change", "")
+
+    commit_type = _detect_type_majority(files)
+
+    prefix_dirs = _prefix_dirs_for_type(commit_type)
+    scope = _derive_scope_from_prefix(files, prefix_dirs) if files else "general"
+
+    if not description:
+        if files:
+            category = _classify_file(files[0])
+            description = _derive_description(category, files)
+        else:
+            description = "update project"
+
+    prefix = f"{commit_type}({scope}): "
+    max_desc_len = 72 - len(prefix)
+    if max_desc_len < 1:
+        max_desc_len = 1
+    if len(description) > max_desc_len:
+        description = description[:max_desc_len - 3].rstrip() + "..."
+
+    subject = f"{prefix}{description}"
+
+    if breaking_change:
+        return f"{subject}\n\nBREAKING CHANGE: {breaking_change}"
+
+    return subject
+
+
+def _derive_description(category: str, files: List[str]) -> str:
+    """Generate a short description for a commit group.
+
+    Args:
+        category: File category (e.g., 'python', 'tests').
+        files: List of affected files.
+
+    Returns:
+        Human-readable description string.
+    """
+    n = len(files)
+    if category == "tests":
+        if n == 1:
+            return f"update test {os.path.basename(files[0])}"
+        return f"update {n} test files"
+    if category == "docs":
+        if n == 1:
+            return f"update {os.path.basename(files[0])}"
+        return f"update {n} documentation files"
+    if category == "config":
+        if n == 1:
+            return f"update {os.path.basename(files[0])}"
+        return f"update {n} config files"
+    # Source code
+    if n == 1:
+        return f"update {os.path.basename(files[0])}"
+    return f"update {n} {category} files"
+
+
+def _is_safe_command(cmd: str) -> tuple[bool, str, list[str]]:
+    cmd = cmd.strip()
+    cmd_lower = cmd.lower()
+
+    for pattern in _BLOCKED_PATTERNS:
+        target = cmd_lower if any(ch.isalpha() for ch in pattern) else cmd
+        if pattern in target:
+            return False, f"blocked pattern '{pattern}'", []
+
+    try:
+        argv = shlex.split(cmd)
+    except ValueError as exc:
+        return False, f"invalid command syntax: {exc}", []
+
+    if not argv:
+        return False, "empty command", []
+
+    normalized_argv = list(argv)
+    normalized_argv[0] = os.path.basename(normalized_argv[0])
+
+    for prefix in _ALLOWED_PREFIXES:
+        if len(argv) < len(prefix):
+            continue
+        if tuple(argv[: len(prefix)]) == prefix or tuple(normalized_argv[: len(prefix)]) == prefix:
+            return True, "", argv
+
+    return False, "not in allowed commands list", []
+
+
+def _load_quality_gate_config(project_dir: str) -> Optional[dict[str, str]]:
+    primary = os.path.join(project_dir, ".omg", "state", "quality-gate.json")
+    legacy = os.path.join(project_dir, ".omc", "quality-gate.json")
+    path = primary if os.path.exists(primary) else legacy
+
+    if not os.path.exists(path):
+        return None
+
+    with open(path, "r", encoding="utf-8") as file_obj:
+        loaded = json.load(file_obj)
+
+    if not isinstance(loaded, dict):
+        raise ValueError("quality-gate.json must be a JSON object")
+
+    config: dict[str, str] = {}
+    for key, value in loaded.items():
+        if isinstance(key, str) and isinstance(value, str):
+            config[key] = value
+    return config
+
+
+def _run_quality_gate(project_dir: str) -> Dict[str, Any]:
+    try:
+        config = _load_quality_gate_config(project_dir)
+    except (OSError, json.JSONDecodeError, ValueError) as exc:
+        return {"ok": False, "step": "config", "reason": str(exc), "results": []}
+
+    if not config:
+        return {"ok": True, "results": []}
+
+    results: list[str] = []
+    for step in _QUALITY_STEPS:
+        cmd = config.get(step)
+        if cmd is None or not cmd.strip():
+            continue
+
+        safe, reason, argv = _is_safe_command(cmd)
+        if not safe:
+            return {
+                "ok": False,
+                "step": step,
+                "reason": f"{cmd} ({reason})",
+                "results": results,
+            }
+
+        try:
+            run_result = subprocess.run(
+                argv,
+                capture_output=True,
+                text=True,
+                timeout=60,
+                cwd=project_dir,
+            )
+        except subprocess.TimeoutExpired:
+            return {
+                "ok": False,
+                "step": step,
+                "reason": f"TIMEOUT {step}: {cmd}",
+                "results": results,
+            }
+        except FileNotFoundError:
+            results.append(f"SKIP {step}: command not found ({cmd})")
+            continue
+
+        if run_result.returncode != 0:
+            snippet = (run_result.stderr or run_result.stdout).strip()[:300]
+            return {
+                "ok": False,
+                "step": step,
+                "reason": f"FAIL {step}: {cmd} (exit {run_result.returncode}) {snippet}",
+                "results": results,
+            }
+        results.append(f"PASS {step}: {cmd} (exit 0)")
+
+    return {"ok": True, "results": results}
+
+
+def execute_commit_plan(plan: List[Dict[str, Any]], project_dir: str, dry_run: bool = False) -> Dict[str, Any]:
+    if not isinstance(plan, list):
+        raise ValueError("plan must be a list")
+
+    for idx, group in enumerate(plan):
+        if not isinstance(group, dict):
+            raise ValueError(f"plan[{idx}] must be a dict")
+        files = group.get("files")
+        message = group.get("message")
+        if not isinstance(files, list) or not files or not all(isinstance(p, str) and p for p in files):
+            raise ValueError(f"plan[{idx}] must include non-empty files list")
+        if not isinstance(message, str) or not message.strip():
+            raise ValueError(f"plan[{idx}] must include non-empty message")
+        if _CONVENTIONAL_COMMIT_RE.match(message.strip()) is None:
+            raise ValueError(f"plan[{idx}] message must be conventional commit format")
+
+    result: Dict[str, Any] = {
+        "succeeded": [],
+        "failed": None,
+        "aborted": [],
+    }
+
+    if dry_run:
+        for group in plan:
+            files = group["files"]
+            message = group["message"].strip()
+            print(f"[DRY-RUN] git add {' '.join(files)}")
+            print(f"[DRY-RUN] git commit -m {message}")
+        return result
+
+    for idx, group in enumerate(plan):
+        files = group["files"]
+        message = group["message"].strip()
+
+        gate = _run_quality_gate(project_dir)
+        if not gate.get("ok", False):
+            result["failed"] = {
+                "message": message,
+                "files": files,
+                "stage": "quality_gate",
+                "reason": gate.get("reason", "quality gate failed"),
+            }
+            result["aborted"] = [remaining["message"] for remaining in plan[idx + 1:]]
+            return result
+
+        add_cmd = ["git", "-C", project_dir, "add", *files]
+        add_result = subprocess.run(
+            add_cmd,
+            capture_output=True,
+            text=True,
+            timeout=60,
+            cwd=project_dir,
+        )
+        if add_result.returncode != 0:
+            result["failed"] = {
+                "message": message,
+                "files": files,
+                "stage": "git_add",
+                "reason": (add_result.stderr or add_result.stdout).strip(),
+            }
+            result["aborted"] = [remaining["message"] for remaining in plan[idx + 1:]]
+            return result
+
+        commit_cmd = ["git", "-C", project_dir, "commit", "-m", message]
+        commit_result = subprocess.run(
+            commit_cmd,
+            capture_output=True,
+            text=True,
+            timeout=60,
+            cwd=project_dir,
+        )
+        if commit_result.returncode != 0:
+            result["failed"] = {
+                "message": message,
+                "files": files,
+                "stage": "git_commit",
+                "reason": (commit_result.stderr or commit_result.stdout).strip(),
+            }
+            result["aborted"] = [remaining["message"] for remaining in plan[idx + 1:]]
+            return result
+
+        result["succeeded"].append(message)
+
+    return result
+
+
+# --- Public API ---
+
+
+def analyze_changes(cwd: str = ".") -> List[Dict[str, Any]]:
+    """Analyze git changes and group hunks by logical concern.
+
+    Groups changes by file type/category, always separating test files
+    from source code. Returns an empty list when the feature flag
+    ``OMG_AI_COMMIT_ENABLED`` is ``False``.
+
+    Args:
+        cwd: Working directory (default: current directory).
+
+    Returns:
+        List of dicts, each with keys:
+            - group_name (str): Human-readable group name.
+            - files (list[str]): Affected file paths.
+            - hunks (list[dict]): Raw hunk dicts from git_hunk().
+            - suggested_type (str): Conventional commit type.
+    """
+    if not _is_enabled():
+        return []
+
+    _ensure_imports()
+    if _git_inspector is None:
+        return []
+    hunks = _git_inspector.git_hunk(cwd)
+
+    if not hunks:
+        return []
+
+    # Bucket hunks by category
+    buckets: Dict[str, Dict[str, Any]] = {}
+    for hunk in hunks:
+        file_path = hunk.get("file", "")
+        category = _classify_file(file_path)
+
+        if category not in buckets:
+            buckets[category] = {
+                "files_set": set(),
+                "hunks": [],
+            }
+        buckets[category]["files_set"].add(file_path)
+        buckets[category]["hunks"].append(hunk)
+
+    # Build result groups
+    groups: List[Dict[str, Any]] = []
+    for category, data in sorted(buckets.items()):
+        files = sorted(data["files_set"])
+        groups.append({
+            "group_name": category,
+            "files": files,
+            "hunks": data["hunks"],
+            "suggested_type": _CATEGORY_DEFAULT_TYPE.get(category, "chore"),
+        })
+
+    return groups
+
+
+def generate_commit_plan(cwd: str = ".") -> Dict[str, Any]:
+    """Generate a full commit plan with proposed messages.
+
+    Calls ``analyze_changes()`` and builds conventional commit messages
+    for each group. Returns an empty plan when the feature flag is off.
+
+    Args:
+        cwd: Working directory (default: current directory).
+
+    Returns:
+        Dict with keys:
+            - groups (list[dict]): Raw groups from analyze_changes().
+            - proposed_commits (list[dict]): Each with ``message``,
+              ``files``, and ``hunks``.
+            - total_commits (int): Number of proposed commits.
+    """
+    groups = analyze_changes(cwd)
+
+    if not groups:
+        return {
+            "groups": [],
+            "proposed_commits": [],
+            "total_commits": 0,
+        }
+
+    proposed: List[Dict[str, Any]] = []
+    for group in groups:
+        commit_type = group["suggested_type"]
+        scope = _derive_scope(group["files"])
+        description = _derive_description(group["group_name"], group["files"])
+        message = f"{commit_type}({scope}): {description}"
+
+        proposed.append({
+            "message": message,
+            "files": group["files"],
+            "hunks": group["hunks"],
+        })
+
+    return {
+        "groups": groups,
+        "proposed_commits": proposed,
+        "total_commits": len(proposed),
+    }
+
+
+def preview_commit_plan(cwd: str = ".") -> str:
+    """Human-readable preview of the commit plan.
+
+    Args:
+        cwd: Working directory (default: current directory).
+
+    Returns:
+        Formatted string showing each proposed commit and affected files.
+        Returns a notice string if the feature flag is off or no changes found.
+    """
+    plan = generate_commit_plan(cwd)
+
+    if plan["total_commits"] == 0:
+        if not _is_enabled():
+            return "[OMG] AI Commit Splitter is disabled. Set OMG_AI_COMMIT_ENABLED=1 to enable."
+        return "[OMG] No uncommitted changes found."
+
+    lines: List[str] = []
+    lines.append("=" * 60)
+    lines.append("  OMG AI Commit Splitter — Proposed Commit Plan")
+    lines.append("=" * 60)
+    lines.append("")
+    lines.append(f"  Total proposed commits: {plan['total_commits']}")
+    lines.append("")
+
+    for idx, commit in enumerate(plan["proposed_commits"], 1):
+        lines.append(f"  Commit {idx}: {commit['message']}")
+        lines.append(f"  {'─' * 50}")
+        for f in commit["files"]:
+            lines.append(f"    • {f}")
+        hunk_count = len(commit["hunks"])
+        lines.append(f"    ({hunk_count} hunk{'s' if hunk_count != 1 else ''})")
+        lines.append("")
+
+    lines.append("=" * 60)
+    lines.append("  NOTE: This is a preview only. No commits were made.")
+    lines.append("=" * 60)
+
+    return "\n".join(lines)
+
+
+# --- CLI ---
+
+
+def main():
+    """CLI entry point."""
+    if len(sys.argv) < 2 or sys.argv[1] != "--dry-run":
+        print("Usage:", file=sys.stderr)
+        print("  python3 tools/commit_splitter.py --dry-run", file=sys.stderr)
+        sys.exit(1)
+
+    cwd = os.getcwd()
+    output = preview_commit_plan(cwd)
+    print(output)
+
+
+if __name__ == "__main__":
+    main()