#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""从 V1/V2 fat-jar 二进制解析嵌套 lib，读取 META-INF/maven/**/pom.properties，生成与 Maven 坐标可比对的数据。

子命令/模式：
- 默认：写 docs/testing/cw-elevator-v1-v2-dependency-diff.md
- --gate：以 V1 为基准，对 candidate fat-jar 做嵌套 JAR 坐标 multiset 比对，不一致时非零退出（可配允许列表）。
"""
from __future__ import annotations

import argparse
import io
import sys
import zipfile
from collections import Counter, defaultdict
from pathlib import Path


def _parse_props(raw: str) -> dict[str, str]:
    props: dict[str, str] = {}
    for line in raw.splitlines():
        line = line.strip()
        if "=" in line and not line.startswith("#"):
            k, _, v = line.partition("=")
            props[k.strip()] = v.strip()
    return props


def coords_from_nested_jar(data: bytes, jar_basename: str) -> tuple[str, str, str]:
    """Returns (groupId, artifactId, version) using META-INF/maven/**/pom.properties."""
    try:
        z = zipfile.ZipFile(io.BytesIO(data))
    except zipfile.BadZipFile:
        return ("?", "?", "?")

    candidates: list[tuple[str, str, str]] = []
    for name in z.namelist():
        if not name.endswith("/pom.properties"):
            continue
        if "META-INF/maven/" not in name:
            continue
        try:
            raw = z.read(name).decode("utf-8", errors="replace")
        except KeyError:
            continue
        props = _parse_props(raw)
        gid = props.get("groupId")
        aid = props.get("artifactId")
        ver = props.get("version")
        if gid and aid and ver:
            candidates.append((gid, aid, ver))

    if not candidates:
        return ("?", "?", "?")

    stem = jar_basename[: -len(".jar")] if jar_basename.endswith(".jar") else jar_basename
    for g, a, v in candidates:
        if stem.startswith(a) or a in stem or stem.startswith(a.split(".")[-1]):
            return (g, a, v)
    return candidates[0]


def list_outer_nested(
    outer: Path, inner_dir_prefix: str
) -> dict[str, tuple[str, str, str]]:
    """path_in_zip -> (g,a,v)"""
    result: dict[str, tuple[str, str, str]] = {}
    z = zipfile.ZipFile(outer)
    for name in z.namelist():
        if not name.startswith(inner_dir_prefix):
            continue
        if not name.endswith(".jar"):
            continue
        if name.endswith(".jar.original"):
            continue
        try:
            data = z.read(name)
        except KeyError:
            continue
        base = Path(name).name
        g, a, v = coords_from_nested_jar(data, base)
        result[name] = (g, a, v)
    return result


def detect_lib_prefix(outer: Path) -> str:
    """可执行包为 V1 风格 lib/ 或 Spring Boot repackage 的 BOOT-INF/lib/。"""
    z = zipfile.ZipFile(outer)
    has_boot = any(n.startswith("BOOT-INF/lib/") and n.endswith(".jar") for n in z.namelist())
    has_lib = any(n.startswith("lib/") and n.endswith(".jar") for n in z.namelist())
    if has_boot and not has_lib:
        return "BOOT-INF/lib/"
    if has_lib:
        return "lib/"
    return "lib/"


def key_ga(path: str, t: tuple[str, str, str]) -> str:
    g, a, v = t
    if g == "?" or a == "?" or v == "?":
        return f"unresolved:{Path(path).name}"
    return f"{g}:{a}:{v}"


def multiset_from_jar(outer: Path) -> tuple[Counter[str], str, dict[str, tuple[str, str, str]]]:
    prefix = detect_lib_prefix(outer)
    jmap = list_outer_nested(outer, prefix)
    ctr: Counter[str] = Counter()
    for path, t in jmap.items():
        ctr[key_ga(path, t)] += 1
    return ctr, prefix, jmap


def _read_allow_counter(path: Path | None) -> Counter[str]:
    if path is None or not path.is_file():
        return Counter()
    c: Counter[str] = Counter()
    for line in path.read_text(encoding="utf-8", errors="replace").splitlines():
        line = line.strip()
        if not line or line.startswith("#"):
            continue
        parts = line.split()
        if len(parts) >= 2 and parts[-1].isdigit():
            key, cnt = " ".join(parts[:-1]), int(parts[-1])
        else:
            key, cnt = line, 1
        c[key] += cnt
    return c


def _subtract_allow(diff: Counter[str], allow: Counter[str]) -> Counter[str]:
    """从 multiset 差值中扣减允许的盈余额度（每行 `coord` 或 `coord N`）。"""
    out = Counter(diff)
    for k, n in allow.items():
        if k not in out:
            continue
        out[k] -= min(out[k], n)
        if out[k] <= 0:
            del out[k]
    return out


def run_lib_gate(
    baseline_jar: Path,
    candidate_jar: Path,
    allow_baseline_only: Path | None,
    allow_candidate_only: Path | None,
) -> int:
    """Multiset 门禁：在应用允许偏差后 baseline 与 candidate 须一致。"""
    if not baseline_jar.is_file():
        print("Missing baseline jar:", baseline_jar, file=sys.stderr)
        return 2
    if not candidate_jar.is_file():
        print("Missing candidate jar:", candidate_jar, file=sys.stderr)
        return 2

    b_ctr, b_prefix, _ = multiset_from_jar(baseline_jar)
    c_ctr, c_prefix, _ = multiset_from_jar(candidate_jar)

    allow_b = _read_allow_counter(allow_baseline_only)
    allow_c = _read_allow_counter(allow_candidate_only)

    only_b = _subtract_allow(b_ctr - c_ctr, allow_b)
    only_c = _subtract_allow(c_ctr - b_ctr, allow_c)

    if not only_b and not only_c:
        print(
            "lib parity OK:",
            f"baseline={baseline_jar.name} ({b_prefix} n={sum(b_ctr.values())})",
            f"candidate={candidate_jar.name} ({c_prefix} n={sum(c_ctr.values())})",
        )
        return 0

    print("lib parity FAILED (multiset diff after allowlists)", file=sys.stderr)
    print(f"  baseline: {baseline_jar} prefix={b_prefix} keys={len(b_ctr)}", file=sys.stderr)
    print(f"  candidate: {candidate_jar} prefix={c_prefix} keys={len(c_ctr)}", file=sys.stderr)
    if only_b:
        print("  only in baseline (or excess count):", file=sys.stderr)
        for k in sorted(only_b.keys()):
            print(f"    {k} x{only_b[k]}", file=sys.stderr)
    if only_c:
        print("  only in candidate (or excess count):", file=sys.stderr)
        for k in sorted(only_c.keys()):
            print(f"    {k} x{only_c[k]}", file=sys.stderr)
    return 1


def parse_maven_dependency_list(path: Path) -> list[tuple[str, str, str, str]]:
    """Parse mvn dependency:list output lines like 'groupId:artifactId:jar:version:compile'."""
    rows: list[tuple[str, str, str, str]] = []
    text = path.read_text(encoding="utf-8", errors="replace")
    for line in text.splitlines():
        line = line.strip()
        if not line or line.startswith("The following"):
            continue
        if line == "none":
            continue
        parts = line.split(":")
        if len(parts) >= 5 and parts[2] == "jar":
            gid, aid, _, ver, scope = parts[0], parts[1], parts[2], parts[3], parts[4]
            rows.append((gid, aid, ver, scope))
    return sorted(rows, key=lambda x: (x[0], x[1], x[2]))


def collect_ga_versions_from_jar_map(
    jmap: dict[str, tuple[str, str, str]],
) -> dict[tuple[str, str], set[str]]:
    """嵌套 jar -> 每个 (groupId, artifactId) 在 fat-jar 中出现的 version 集合（仅 pom.properties 可解析条目）。"""
    ga_ver: dict[tuple[str, str], set[str]] = defaultdict(set)
    for _path, (g, a, v) in jmap.items():
        if g != "?" and a != "?" and v != "?":
            ga_ver[(g, a)].add(v)
    return ga_ver


def ga_version_skew_rows(
    v1_ga_v: dict[tuple[str, str], set[str]],
    v2_ga_v: dict[tuple[str, str], set[str]],
) -> list[tuple[tuple[str, str], list[str], list[str]]]:
    """两侧均能解析到 version、且 version 集合不一致的 GA。"""
    out: list[tuple[tuple[str, str], list[str], list[str]]] = []
    all_ga = set(v1_ga_v.keys()) | set(v2_ga_v.keys())
    for ga in sorted(all_ga, key=lambda x: (x[0], x[1])):
        s1 = v1_ga_v.get(ga, set())
        s2 = v2_ga_v.get(ga, set())
        if s1 and s2 and s1 != s2:
            out.append((ga, sorted(s1), sorted(s2)))
    return out


def main() -> int:
    root = Path(__file__).resolve().parents[1]
    ap = argparse.ArgumentParser(description="V1/V2 fat-jar 依赖对比或 lib multiset 门禁")
    ap.add_argument(
        "--gate",
        action="store_true",
        help="运行 multiset 门禁（默认基准 V1 jar，候选为 releases 或 starter/target）",
    )
    ap.add_argument("--baseline-jar", type=Path, default=None, help="门禁基准 fat-jar")
    ap.add_argument("--candidate-jar", type=Path, default=None, help="门禁待测 fat-jar")
    ap.add_argument(
        "--allow-baseline-only",
        type=Path,
        default=None,
        help="允许仅出现在 baseline 的坐标（每行 `g:a:v` 或 `unresolved:name.jar`，可选末尾计数）",
    )
    ap.add_argument(
        "--allow-candidate-only",
        type=Path,
        default=None,
        help="允许仅出现在 candidate 的坐标（格式同上）",
    )
    args = ap.parse_args()

    if args.gate:
        b = (
            args.baseline_jar
            if args.baseline_jar
            else root
            / "cw-elevator-application-V1.0.0.20211103"
            / "cw-elevator-application-V1.0.0.20211103.jar"
        )
        cand = args.candidate_jar
        if cand is None:
            # 优先本地 package 产物（与当前父 POM / 插件一致）；其次历史 releases 归档
            cand = (
                root
                / "maven-cw-elevator-application"
                / "cw-elevator-application-starter"
                / "target"
                / "cw-elevator-application-2.0.7.jar"
            )
            if not cand.is_file():
                cand = (
                    root
                    / "maven-cw-elevator-application"
                    / "cw-elevator-application-starter"
                    / "target"
                    / "cw-elevator-application-2.0.0.jar"
                )
            if not cand.is_file():
                cand = (
                    root
                    / "maven-cw-elevator-application"
                    / "releases"
                    / "cw-elevator-application-V2.0.6.20260430"
                    / "cw-elevator-application-2.0.6.jar"
                )
        allow_b = args.allow_baseline_only
        if allow_b is None:
            p = root / "docs" / "testing" / "cw-elevator-fatjar-lib-parity-allow-baseline-only.txt"
            allow_b = p if p.is_file() else None
        allow_c = args.allow_candidate_only
        if allow_c is None:
            p = root / "docs" / "testing" / "cw-elevator-fatjar-lib-parity-allow-candidate-only.txt"
            allow_c = p if p.is_file() else None
        return run_lib_gate(b, cand, allow_b, allow_c)

    v1_jar = root / "cw-elevator-application-V1.0.0.20211103" / "cw-elevator-application-V1.0.0.20211103.jar"
    v2_jar = (
        root
        / "maven-cw-elevator-application"
        / "cw-elevator-application-starter"
        / "target"
        / "cw-elevator-application-2.0.7.jar"
    )
    if not v2_jar.is_file():
        v2_jar = (
            root
            / "maven-cw-elevator-application"
            / "cw-elevator-application-starter"
            / "target"
            / "cw-elevator-application-2.0.0.jar"
        )
    if not v2_jar.is_file():
        v2_jar = (
            root
            / "maven-cw-elevator-application"
            / "releases"
            / "cw-elevator-application-V2.0.6.20260430"
            / "cw-elevator-application-2.0.6.jar"
        )
    # reactor + dependency:list 时各模块写各自 target/；starter 模块输出才是入口 fat-jar 的 runtime 列表
    mvn_list = (
        root
        / "maven-cw-elevator-application"
        / "cw-elevator-application-starter"
        / "target"
        / "v2-maven-deps.txt"
    )
    if not mvn_list.is_file():
        mvn_list = root / "maven-cw-elevator-application" / "target" / "v2-maven-deps.txt"
    if not mvn_list.is_file():
        mvn_list = Path("/tmp/v2-maven-deps.txt")

    out_md = (
        root
        / "docs"
        / "testing"
        / "cw-elevator-v1-v2-dependency-diff.md"
    )

    if not v1_jar.is_file():
        print("Missing V1 jar:", v1_jar, file=sys.stderr)
        return 1
    if not v2_jar.is_file():
        print("Missing V2 jar:", v2_jar, file=sys.stderr)
        return 1

    v1_map = list_outer_nested(v1_jar, "lib/")
    v2_prefix = detect_lib_prefix(v2_jar)
    v2_map = list_outer_nested(v2_jar, v2_prefix)

    v1_by_ga: dict[str, list[str]] = defaultdict(list)
    for path, t in v1_map.items():
        v1_by_ga[key_ga(path, t)].append(path)
    v2_by_ga: dict[str, list[str]] = defaultdict(list)
    for path, t in v2_map.items():
        v2_by_ga[key_ga(path, t)].append(path)

    keys1 = set(v1_by_ga.keys())
    keys2 = set(v2_by_ga.keys())

    only_v1 = sorted(keys1 - keys2)
    only_v2 = sorted(keys2 - keys1)
    both = sorted(keys1 & keys2)

    mvn_rows: list[tuple[str, str, str, str]] = []
    if mvn_list.is_file():
        mvn_rows = parse_maven_dependency_list(mvn_list)

    mvn_ga = {f"{g}:{a}:{v}" for g, a, v, _ in mvn_rows}

    lines: list[str] = []
    lines.append("# cw-elevator-application V1 fat-jar 与 V2 fat-jar 依赖差异核对")
    lines.append("")
    lines.append("**生成方式**：脚本 `scripts/generate_v1_v2_elevator_dependency_diff.py`（可重复执行覆盖本文件）。")
    lines.append("")
    lines.append("## 样本路径")
    lines.append("")
    lines.append(f"- **V1**：`{v1_jar.relative_to(root)}`")
    lines.append(f"- **V2**：`{v2_jar.relative_to(root)}`")
    lines.append("")
    lines.append("| 指标 | V1 | V2 |")
    lines.append("|------|----|----|")
    lines.append(f"| 嵌套 jar 条目数（lib / BOOT-INF/lib） | {len(v1_map)} | {len(v2_map)} |")
    lines.append(f"| 解析出唯一坐标 `groupId:artifactId:version` 数 | {len(keys1)} | {len(keys2)} |")
    v1_ga_v = collect_ga_versions_from_jar_map(v1_map)
    v2_ga_v = collect_ga_versions_from_jar_map(v2_map)
    skew_ga = ga_version_skew_rows(v1_ga_v, v2_ga_v)
    lines.append(
        f"| 同名 GA、两侧均有解析且 version 集合不一致（§2.2.1）条数 | — | **{len(skew_ga)}** |"
    )
    lines.append(
        f"| 与 Maven `dependency:list`（runtime）条目数 | — | {len(mvn_rows)} |"
    )
    lines.append("")

    lines.append("---")
    lines.append("")
    lines.append("## 1. Maven 方式（仅 V2 reactor）")
    lines.append("")
    lines.append(
        "在 `maven-cw-elevator-application` 下执行：`mvn -pl cw-elevator-application-starter -am "
        "dependency:list -DincludeScope=runtime -Dsort=true "
        "-DoutputFile=target/v2-maven-deps.txt`。`-am` 时每个子模块写各自的 `target/`；"
        "**§1.1 使用 starter 模块文件**：`cw-elevator-application-starter/target/v2-maven-deps.txt`。"
    )
    lines.append("")
    lines.append(
        "**说明**：历史 **V1 运行包** 当前仓库无对应 **1.0** 聚合工程可一键 `dependency:list`；"
        "V1 的 Maven 坐标视图见 **§2 二进制嵌套 JAR 的 pom.properties**。"
    )
    lines.append("")
    lines.append("### 1.1 V2 `dependency:list` 全量（runtime）")
    lines.append("")
    lines.append("| # | groupId | artifactId | version | scope |")
    lines.append("|---|---------|--------------|---------|-------|")
    for i, (g, a, v, s) in enumerate(mvn_rows, 1):
        lines.append(f"| {i} | `{g}` | `{a}` | `{v}` | `{s}` |")
    lines.append("")

    lines.append("---")
    lines.append("")
    lines.append("## 2. 二进制方式（嵌套 JAR + pom.properties）")
    lines.append("")
    lines.append(
        f"- **V1**：`lib/*.jar`。\n"
        f"- **V2**：自动检测为 `{v2_prefix}*.jar`（与 spring-boot-maven-plugin 1.3.x + Boot 1.5 一致时为 `lib/`）。"
    )
    lines.append("")
    lines.append(
        "对每个嵌套 jar 读取 `META-INF/maven/**/pom.properties` 得到 `groupId:artifactId:version`；"
        "无法读取时记为 `?:?:?`（多为无 Maven 元数据的第三方包）。"
    )
    lines.append("")

    lines.append("### 2.1 仅在 V1 出现的坐标（相对 V2 二进制集合）")
    lines.append("")
    lines.append(f"**共 {len(only_v1)} 项**。")
    lines.append("")
    lines.append("| groupId:artifactId:version | V1 嵌套路径 |")
    lines.append("|------------------------------|-------------|")
    for k in only_v1:
        paths = ", ".join(f"`{p}`" for p in sorted(v1_by_ga[k]))
        lines.append(f"| `{k}` | {paths} |")
    lines.append("")

    lines.append("### 2.2 仅在 V2 出现的坐标（相对 V1 二进制集合）")
    lines.append("")
    lines.append(f"**共 {len(only_v2)} 项**。")
    lines.append("")
    lines.append("| groupId:artifactId:version | V2 嵌套路径 |")
    lines.append("|------------------------------|-------------|")
    for k in only_v2:
        paths = ", ".join(f"`{p}`" for p in sorted(v2_by_ga[k]))
        lines.append(f"| `{k}` | {paths} |")
    lines.append("")

    lines.append("### 2.2.1 同名构件（groupId:artifactId）在 V1 与 V2 中的版本集合差异")
    lines.append("")
    lines.append(
        "由嵌套 jar 的 `pom.properties` 聚合：若同一 **GA** 在 V1、V2 中均能解析出版本，且 **version 集合不同**，"
        "则单独列出（与 §2.1 / §2.2 中分列的 `g:a:v` 键互为补充）。**不含**仅一侧出现的 GA。"
    )
    lines.append("")
    lines.append(f"**共 {len(skew_ga)} 项**。")
    lines.append("")
    lines.append("| groupId:artifactId | V1 version(s) | V2 version(s) |")
    lines.append("|--------------------|---------------|---------------|")
    for (g, a), v1s, v2s in skew_ga:
        ga_s = f"`{g}:{a}`"
        lines.append(
            f"| {ga_s} | `{', '.join(v1s)}` | `{', '.join(v2s)}` |"
        )
    lines.append("")

    lines.append("### 2.3 两边均存在且坐标一致的依赖")
    lines.append("")
    lines.append(f"**共 {len(both)} 项**（名称版本完全一致）。")
    lines.append("")
    lines.append("<details>")
    lines.append("<summary>展开长表</summary>")
    lines.append("")
    lines.append("| groupId:artifactId:version |")
    lines.append("|------------------------------|")
    for k in both:
        lines.append(f"| `{k}` |")
    lines.append("")
    lines.append("</details>")
    lines.append("")

    lines.append("### 2.4 V2 二进制坐标 vs Maven dependency:list")
    lines.append("")
    lines.append(
        "- **版本字符串不一致**：例如 reactor 在 `dependency:list` 中为 **`2.0-SNAPSHOT`**，"
        "而 fat-jar 内嵌模块 **`cw-elevator-application-*-2.0.6.jar`** 的 `pom.properties` 为 **`2.0.6`**，"
        "字符串比对会视为「仅一侧存在」，属**同名构件不同表述**，非缺失依赖。"
    )
    lines.append(
        "- **在 dependency:list 中但不在嵌套 jar 元数据中的**：多为 **仅存在于解析树、与本模块 jar 文件命名不一致**，需对照 §1 表格。"
    )
    lines.append(
        "- **未解析 `unresolved:*`**：见 §3，此类条目不参与坐标相等判断。"
    )
    lines.append("")
    only_mvn = sorted(mvn_ga - keys2)
    only_bin = sorted(keys2 - mvn_ga)
    lines.append(f"- **仅在 Maven list（runtime）**：{len(only_mvn)} 项")
    lines.append("")
    if only_mvn:
        lines.append("|坐标|")
        lines.append("|----|")
        for k in only_mvn[:80]:
            lines.append(f"| `{k}` |")
        if len(only_mvn) > 80:
            lines.append(f"| … 其余 {len(only_mvn) - 80} 项省略 |")
    lines.append("")
    lines.append(f"- **仅在二进制坐标集合**：{len(only_bin)} 项")
    lines.append("")
    if only_bin:
        lines.append("|坐标|")
        lines.append("|----|")
        for k in only_bin[:80]:
            lines.append(f"| `{k}` |")
        if len(only_bin) > 80:
            lines.append(f"| … 其余 {len(only_bin) - 80} 项省略 |")
    lines.append("")

    lines.append("---")
    lines.append("")
    lines.append("## 3. 无法解析 pom.properties 的嵌套 JAR（仅列文件名）")
    lines.append("")
    bad_v1 = [p for p, t in v1_map.items() if t[0] == "?" or t[2] == "?"]
    bad_v2 = [p for p, t in v2_map.items() if t[0] == "?" or t[2] == "?"]
    lines.append(f"- **V1** 未解析条目：**{len(bad_v1)}**")
    for p in sorted(bad_v1)[:50]:
        lines.append(f"  - `{Path(p).name}`")
    if len(bad_v1) > 50:
        lines.append(f"  - … 省略 {len(bad_v1) - 50} 条")
    lines.append(f"- **V2** 未解析条目：**{len(bad_v2)}**")
    for p in sorted(bad_v2)[:50]:
        lines.append(f"  - `{Path(p).name}`")
    if len(bad_v2) > 50:
        lines.append(f"  - … 省略 {len(bad_v2) - 50} 条")
    lines.append("")

    out_md.parent.mkdir(parents=True, exist_ok=True)
    out_md.write_text("\n".join(lines) + "\n", encoding="utf-8")
    print("Wrote", out_md)
    return 0


if __name__ == "__main__":
    sys.exit(main())