@comate/zulu 1.4.0-beta.2 → 1.4.0-beta.4

package/comate-engine/assets/skills/auto-commit/scripts/git_diff_cli.py

@@ -0,0 +1,201 @@
+"""命令行工具：获取 workspace 的 git diff 摘要
+
+用法:
+    python3 git_diff_cli.py --workspace /path/to/project
+
+输出:
+    JSON 格式的 diff 摘要，包含 changed_files、stat_summary、diff_content 等字段。
+    失败时输出 {"error": "错误信息"}。
+"""
+
+import argparse
+import json
+import os
+import subprocess
+import sys
+
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+from cache_manager import update_git_diff
+
+# tracked diff 最大行数
+MAX_TRACKED_DIFF_LINES = 300
+# untracked diff 最大行数
+MAX_UNTRACKED_DIFF_LINES = 200
+# 合并后 diff 最大总行数
+MAX_TOTAL_DIFF_LINES = 500
+
+
+def run_git(args, cwd, timeout=10):
+    """执行 git 命令，返回 stdout 字符串。失败返回空字符串。"""
+    try:
+        result = subprocess.run(
+            ["git"] + args,
+            capture_output=True, text=True, encoding="utf-8", errors="replace",
+            timeout=timeout, cwd=cwd
+        )
+        if result.returncode == 0:
+            return result.stdout
+        return ""
+    except (subprocess.TimeoutExpired, FileNotFoundError, OSError):
+        return ""
+
+
+def parse_numstat(cwd):
+    """通过 git diff HEAD --numstat 获取 tracked 文件变更统计。"""
+    output = run_git(["diff", "HEAD", "--numstat"], cwd)
+    files = []
+    for line in output.strip().splitlines():
+        if not line:
+            continue
+        parts = line.split("\t")
+        if len(parts) != 3:
+            continue
+        ins_str, del_str, file_path = parts
+        # binary 文件的 numstat 输出是 - - filename
+        insertions = int(ins_str) if ins_str != "-" else 0
+        deletions = int(del_str) if del_str != "-" else 0
+        files.append({
+            "file": file_path,
+            "insertions": insertions,
+            "deletions": deletions,
+        })
+    return files
+
+
+def get_untracked_files(cwd):
+    """获取 untracked 文件列表。"""
+    output = run_git(["ls-files", "--others", "--exclude-standard"], cwd)
+    return [f for f in output.strip().splitlines() if f]
+
+
+def get_tracked_diff(cwd):
+    """获取 tracked 文件的 diff 内容，截断到 MAX_TRACKED_DIFF_LINES 行。"""
+    output = run_git(["diff", "HEAD"], cwd, timeout=30)
+    lines = output.splitlines()
+    truncated = len(lines) > MAX_TRACKED_DIFF_LINES
+    if truncated:
+        lines = lines[:MAX_TRACKED_DIFF_LINES]
+    return "\n".join(lines), truncated
+
+
+def get_untracked_diff_and_stats(cwd, untracked_files):
+    """获取 untracked 文件的 diff 和变更统计。
+
+    Returns:
+        (changed_files, diff_content, truncated)
+    """
+    changed_files = []
+    diff_lines = []
+    total_lines = 0
+    truncated = False
+
+    for file_path in untracked_files:
+        full_path = os.path.join(cwd, file_path)
+        try:
+            with open(full_path, "r", encoding="utf-8", errors="replace") as f:
+                lines = f.readlines()
+        except (OSError, IOError):
+            continue
+
+        insertions = len(lines)
+        if insertions > 0:
+            changed_files.append({
+                "file": file_path,
+                "insertions": insertions,
+                "deletions": 0,
+            })
+
+        # 构造 unified diff 格式的输出
+        file_diff_lines = [
+            f"diff --git a/{file_path} b/{file_path}",
+            "new file mode 100644",
+            f"--- /dev/null",
+            f"+++ b/{file_path}",
+            f"@@ -0,0 +1,{insertions} @@",
+        ] + [f"+{line.rstrip()}" for line in lines]
+
+        # 检查总行数上限
+        remaining = MAX_UNTRACKED_DIFF_LINES - total_lines
+        if remaining <= 0:
+            truncated = True
+            break
+        if len(file_diff_lines) > remaining:
+            file_diff_lines = file_diff_lines[:remaining]
+            truncated = True
+        diff_lines.extend(file_diff_lines)
+        total_lines += len(file_diff_lines)
+
+    return changed_files, "\n".join(diff_lines), truncated
+
+
+def build_stat_summary(changed_files):
+    """构建统计摘要字符串，如 '3 files, +130, -78'。"""
+    total_ins = sum(f["insertions"] for f in changed_files)
+    total_del = sum(f["deletions"] for f in changed_files)
+    n = len(changed_files)
+    return f"{n} file{'s' if n != 1 else ''}, +{total_ins}, -{total_del}"
+
+
+def main():
+    """解析命令行参数并输出 workspace 的 git diff 摘要 JSON。"""
+    parser = argparse.ArgumentParser(description="获取 workspace 的 git diff 摘要")
+    parser.add_argument("--workspace", required=True, help="workspace 路径")
+    parser.add_argument("--max-diff-lines", type=int, default=MAX_TOTAL_DIFF_LINES,
+                        help="diff 内容最大总行数")
+    args = parser.parse_args()
+
+    workspace = os.path.abspath(args.workspace)
+    if not os.path.isdir(workspace):
+        print(json.dumps({"error": f"workspace 路径不存在: {workspace}"}))
+        sys.exit(1)
+
+    max_total = args.max_diff_lines
+
+    try:
+        # 1. tracked 文件变更统计
+        tracked_files = parse_numstat(workspace)
+
+        # 2. untracked 文件列表
+        untracked_files = get_untracked_files(workspace)
+
+        # 3. tracked diff 内容
+        tracked_diff, tracked_truncated = get_tracked_diff(workspace)
+
+        # 4. untracked diff 和统计
+        untracked_changed, untracked_diff, untracked_truncated = \
+            get_untracked_diff_and_stats(workspace, untracked_files)
+
+        # 5. 合并 changed_files
+        all_changed = tracked_files + untracked_changed
+
+        # 6. 合并 diff 内容，检查总行数上限
+        tracked_lines = tracked_diff.splitlines() if tracked_diff else []
+        untracked_lines = untracked_diff.splitlines() if untracked_diff else []
+        all_lines = tracked_lines + untracked_lines
+        truncated = tracked_truncated or untracked_truncated or len(all_lines) > max_total
+        if len(all_lines) > max_total:
+            all_lines = all_lines[:max_total]
+        diff_content = "\n".join(all_lines)
+
+        # 7. 构建输出
+        has_changes = bool(all_changed) or bool(untracked_files)
+        output = {
+            "workspace": workspace,
+            "changed_files": all_changed,
+            "untracked_files": untracked_files,
+            "stat_summary": build_stat_summary(all_changed) if all_changed else "",
+            "has_changes": has_changes,
+            "diff_content": diff_content,
+            "truncated": truncated,
+        }
+        # 写入统一缓存
+        update_git_diff(workspace, output)
+        print(json.dumps(output, ensure_ascii=False))
+
+    except Exception as e:
+        print(json.dumps({"error": f"获取 diff 摘要异常: {type(e).__name__}: {e}"}))
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()

package/comate-engine/assets/skills/auto-commit/scripts/git_utils.py

@@ -0,0 +1,230 @@
+"""Git 工具函数"""
+
+import os
+import subprocess
+import re
+from typing import Dict, List, Any, Optional
+
+from icafe.client import (
+    DEFAULT_MAX_COMMITS,
+    DEFAULT_MAX_DIFF_LINES,
+    GIT_LOG_FALSE_POSITIVE_PREFIXES,
+)
+from logger import get_logger
+
+logger = get_logger()
+
+
+def _get_username_from_arg_or_env(username: Optional[str] = None) -> Optional[str]:
+    """从命令行参数或环境变量获取用户名
+
+    优先使用传入的 username 参数（来自命令行 --username），
+    回退到环境变量 COMATE_USERNAME。
+
+    Args:
+        username: 通过命令行参数传入的用户名
+
+    Returns:
+        用户名字符串；未设置或未替换时返回 None
+    """
+    # 优先使用命令行参数
+    if username and 'COMATE_USERNAME' not in username:
+        return username
+    # 回退到环境变量
+    value = os.environ.get('COMATE_USERNAME', '')
+    if value and 'COMATE_USERNAME' not in value:
+        return value
+    return None
+
+
+def get_current_user(username: Optional[str] = None) -> Optional[str]:
+    """获取当前用户 ID
+
+    优先级：
+    1. 通过参数传入的用户名（来自命令行 --username）
+    2. 环境变量 COMATE_USERNAME
+    3. git config user.email 提取 @ 前的用户名
+    4. git config user.name
+
+    Args:
+        username: 通过命令行参数传入的用户名
+
+    Returns:
+        用户 ID 字符串，如 "dongkexin01"；无法获取时返回 None
+    """
+    # 1. 优先从参数或环境变量获取用户名
+    comate_username = _get_username_from_arg_or_env(username)
+    logger.info("skill.md username: %s", comate_username)
+    if comate_username:
+        logger.info("用户识别: source=arg_or_env, user=%s", comate_username)
+        return comate_username
+
+    # 2. 从 git config user.email 提取
+    try:
+        result = subprocess.run(
+            ["git", "config", "user.email"],
+            capture_output=True, text=True, encoding="utf-8", errors="replace",
+            timeout=5
+        )
+        if result.returncode == 0 and result.stdout.strip():
+            username = result.stdout.strip().split("@")[0]
+            if username:
+                logger.info("用户识别: source=email, user=%s", username)
+                return username
+    except (subprocess.TimeoutExpired, FileNotFoundError):
+        pass
+
+    # 3. 回退到 git config user.name
+    try:
+        result = subprocess.run(
+            ["git", "config", "user.name"],
+            capture_output=True, text=True, encoding="utf-8", errors="replace",
+            timeout=5
+        )
+        if result.returncode == 0 and result.stdout.strip():
+            username = result.stdout.strip()
+            logger.info("用户识别: source=name, user=%s", username)
+            return username
+    except (subprocess.TimeoutExpired, FileNotFoundError):
+        pass
+
+    logger.warning("无法获取用户 ID，所有来源均失败")
+    return None
+
+
+def get_git_diff_summary(max_diff_lines: int = DEFAULT_MAX_DIFF_LINES) -> Optional[Dict[str, Any]]:
+    """获取当前工作区的 git diff 摘要
+
+    执行 git diff HEAD 获取所有未提交的变更（包括暂存和未暂存），
+    并提取变更文件列表和 diff 内容摘要。
+
+    Args:
+        max_diff_lines: diff 输出的最大行数，超出部分截断。默认 500。
+
+    Returns:
+        包含 diff 信息的字典：
+        {
+            "changed_files": ["path/to/file1.ts", ...],
+            "stat_summary": "git diff --stat 输出的摘要行",
+            "diff_content": "截断后的 diff 全文",
+            "truncated": bool
+        }
+        非 git 仓库或无变更时返回 None
+    """
+    try:
+        stat_result = subprocess.run(
+            ["git", "diff", "HEAD", "--stat"],
+            capture_output=True, text=True, encoding="utf-8", errors="replace",
+            timeout=10
+        )
+        if stat_result.returncode != 0 or not stat_result.stdout.strip():
+            logger.info("git diff 无变更")
+            return None
+    except (subprocess.TimeoutExpired, FileNotFoundError):
+        logger.info("git diff 无变更")
+        return None
+
+    stat_lines = stat_result.stdout.strip().splitlines()
+    changed_files = []
+    for line in stat_lines[:-1]:
+        file_path = line.split("|")[0].strip()
+        if file_path:
+            changed_files.append(file_path)
+    stat_summary = stat_lines[-1].strip() if stat_lines else ""
+
+    if not changed_files:
+        logger.info("git diff 无变更")
+        return None
+
+    try:
+        diff_result = subprocess.run(
+            ["git", "diff", "HEAD"],
+            capture_output=True, text=True, encoding="utf-8", errors="replace",
+            timeout=30
+        )
+        diff_content = diff_result.stdout if diff_result.returncode == 0 else ""
+    except (subprocess.TimeoutExpired, FileNotFoundError):
+        diff_content = ""
+
+    lines = diff_content.splitlines()
+    truncated = len(lines) > max_diff_lines
+    if truncated:
+        diff_content = "\n".join(lines[:max_diff_lines]) + "\n... (truncated)"
+
+    logger.info("git diff: %d 个文件, %s, truncated=%s", len(changed_files), stat_summary, truncated)
+
+    return {
+        "changed_files": changed_files,
+        "stat_summary": stat_summary,
+        "diff_content": diff_content,
+        "truncated": truncated,
+    }
+
+
+def extract_space_ids_from_git_log(
+    max_commits: int = DEFAULT_MAX_COMMITS,
+    known_prefixes: Optional[List[str]] = None,
+) -> List[str]:
+    """从 git commit message 中提取 iCafe 空间前缀
+
+    优先使用已知的空间前缀列表进行精确匹配（支持含连字符的前缀如 DevOps-iScan），
+    回退到正则启发式提取。
+
+    Args:
+        max_commits: 扫描的最大提交数量
+        known_prefixes: 已知的空间 prefixCode 列表（来自 API）。
+                        提供时优先用精确匹配；未提供时用正则启发式。
+
+    Returns:
+        去重后的空间前缀列表（保留原始大小写），按提交次数降序排列。
+        非 git 仓库或无匹配时返回空列表。
+    """
+    try:
+        result = subprocess.run(
+            ["git", "log", "--oneline", "-n", str(max_commits)],
+            capture_output=True, text=True, encoding="utf-8", errors="replace",
+            timeout=10
+        )
+        if result.returncode != 0:
+            return []
+    except (subprocess.TimeoutExpired, FileNotFoundError):
+        return []
+
+    log_text = result.stdout.strip()
+    if not log_text:
+        return []
+
+    # 策略 1：用已知前缀精确匹配 "前缀-数字"
+    if known_prefixes:
+        counter: Dict[str, int] = {}
+        for prefix in known_prefixes:
+            pat = re.compile(re.escape(prefix) + r'-\d+', re.IGNORECASE)
+            count = len(pat.findall(log_text))
+            if count > 0:
+                counter[prefix] = count
+        if counter:
+            sorted_prefixes = sorted(counter, key=counter.get, reverse=True)
+            logger.info("git log 提取前缀: %s, 策略=精确匹配", sorted_prefixes)
+            return sorted_prefixes
+
+    # 策略 2：正则启发式兜底（无已知前缀或精确匹配无结果）
+    # 支持含连字符的复合前缀，如 DevOps-iScan-36261、TPUE-EE-2-22
+    # 贪婪匹配：最后一个 -数字 是卡片 ID，前面所有部分都是空间前缀
+    pattern = re.compile(r'\b([a-zA-Z][a-zA-Z0-9_]*(?:-[a-zA-Z0-9_]+)*)-(\d+)\b')
+    counter_fallback: Dict[str, int] = {}
+
+    for line in log_text.splitlines():
+        for match in pattern.finditer(line):
+            raw_prefix = match.group(1)
+            prefix_lower = raw_prefix.lower()
+            # 检查前缀首段是否是常见误报（如 feat-xxx、fix-xxx）
+            first_segment = raw_prefix.split('-', 1)[0].lower()
+            if first_segment in GIT_LOG_FALSE_POSITIVE_PREFIXES:
+                continue
+            if len(raw_prefix) <= 1:
+                continue
+            counter_fallback[prefix_lower] = counter_fallback.get(prefix_lower, 0) + 1
+
+    sorted_prefixes = sorted(counter_fallback, key=counter_fallback.get, reverse=True)
+    logger.info("git log 提取前缀: %s, 策略=正则启发式", sorted_prefixes)
+    return sorted_prefixes

package/comate-engine/assets/skills/auto-commit/scripts/icafe/__init__.py

@@ -0,0 +1,66 @@
+"""iCafe Card Query Client - 卡片条件查询助手
+
+提供 iCafe 平台卡片条件查询相关的 API 接口封装，
+支持通过 IQL 表达式、卡片 ID、空间等维度灵活查询卡片信息。
+"""
+
+from .client import (
+    ICafeQueryConfig,
+    ICafeQueryClient,
+    # 常量
+    ICAFE_BASE_URL,
+    FARSEER_BASE_URL,
+    DEFAULT_TIMEOUT,
+    DEFAULT_LOOKBACK_DAYS,
+    DEFAULT_MAX_RECORDS,
+    DEFAULT_MAX_COMMITS,
+    DEFAULT_MAX_DIFF_LINES,
+    GIT_LOG_FALSE_POSITIVE_PREFIXES,
+    ISSUE_TYPE_MAP,
+    ISSUE_TYPE_NAME_MAP,
+)
+from git_utils import (
+    get_current_user,
+    get_git_diff_summary,
+    extract_space_ids_from_git_log,
+)
+from .matching import (
+    auto_detect_space,
+    find_matching_card,
+    match_diff_to_cards,
+    build_type_filter_iql,
+    format_card_summary,
+    format_query_result,
+)
+from .farseer import get_binding_card_types
+
+__all__ = [
+    # 配置和客户端
+    'ICafeQueryConfig',
+    'ICafeQueryClient',
+    # 常量
+    'ICAFE_BASE_URL',
+    'FARSEER_BASE_URL',
+    'DEFAULT_TIMEOUT',
+    'DEFAULT_LOOKBACK_DAYS',
+    'DEFAULT_MAX_RECORDS',
+    'DEFAULT_MAX_COMMITS',
+    'DEFAULT_MAX_DIFF_LINES',
+    'GIT_LOG_FALSE_POSITIVE_PREFIXES',
+    'ISSUE_TYPE_MAP',
+    'ISSUE_TYPE_NAME_MAP',
+    # Git 工具
+    'get_current_user',
+    'get_git_diff_summary',
+    'extract_space_ids_from_git_log',
+    # 自动检测和匹配
+    'auto_detect_space',
+    'find_matching_card',
+    'match_diff_to_cards',
+    'build_type_filter_iql',
+    # 格式化
+    'format_card_summary',
+    'format_query_result',
+    # Farseer API
+    'get_binding_card_types',
+]