#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""ru-4 源代码鉴别材料 重生成器（确定性、字节忠实）。
复刻 ru-4 既有格式:7 模块按目录字典序(ad→biz→community→compliance→feed→telemetry→trade),
模块内按相对 game-cloud 路径字典序;仅 src/main 下 .java;
分隔「// ===== 文件: <相对game-cloud路径> =====」;每文件 = 分隔行 + 源码行(splitlines,文件间无空行);
总行 = 文件数 + 物理行;取前 1500 + 后 1500,中间省略。
只读源码,产物写 /tmp/ru4_new.txt;打印逐文件/分模块/总计数,并落各模块清单表行 /tmp/ru4_tbl_*.md。"""
import os, glob
from collections import OrderedDict

REPO = "/Users/lili/Project/games-development-ai"
CLOUD = os.path.join(REPO, "game-cloud")
# 目录字典序(与既有 .txt 起始 game-module-ad 一致)
MODULES = ["game-module-ad", "game-module-biz", "game-module-community", "game-module-compliance",
           "game-module-feed", "game-module-telemetry", "game-module-trade"]
SEP = "// ===== 文件: {rel} ====="

files = []  # (module, rel_from_cloud, wcl, splitlines)
no_trailing_nl = []
for m in MODULES:
    found = []
    for j in glob.glob(os.path.join(CLOUD, m, "game-module-*", "src", "main", "**", "*.java"), recursive=True):
        found.append((os.path.relpath(j, CLOUD), j))
    found.sort(key=lambda x: x[0])
    for rel, j in found:
        raw = open(j, encoding="utf-8").read()
        wcl = raw.count("\n")                       # wc -l 口径
        if raw and not raw.endswith("\n"):
            no_trailing_nl.append(rel)
        files.append((m, rel, wcl, raw.splitlines()))  # splitlines:无文件间空行,总行=文件数+物理行

# 拼接(字节忠实:每文件=分隔行 + 源码各行)
concat = []
for m, rel, wcl, sl in files:
    concat.append(SEP.format(rel=rel))
    concat.extend(sl)
total = len(concat)
front = concat[:1500]
back = concat[-1500:]
omit_from, omit_to, omit_cnt = 1501, total - 1500, total - 3000

nfiles = len(files)
nsrc = sum(f[2] for f in files)
nonblank = sum(sum(1 for x in sl if x.strip() != "") for (m, rel, wcl, sl) in files)
blank = nsrc - nonblank
mod = OrderedDict()
for m, rel, wcl, sl in files:
    e = mod.setdefault(m, [0, 0]); e[0] += 1; e[1] += wcl

RULE = "=" * 80
header = [
    "软件名称：绘境AI游戏分发变现与合规云服务系统  版本：V1.0",
    "说明：本鉴别材料为本软件原创业务源代码（Java，包名 com.wanxiang.huijing.game.module.*），按纳入文件目录字典序连续拼接。",
    "统计：纳入原创源文件 %d 个；连续有效代码（含文件分隔标记行）共 %d 行（其中物理代码行 %d 行、空白行 %d 行、非空白有效行 %d 行）。" % (nfiles, total, nsrc, blank, nonblank),
    "因总行数超过 3000 行，依软著登记规范取【前 1500 行 + 后 1500 行】，中间部分以分隔标识省略。完整文件清单见《源文件清单.md》。",
    "（已严格排除第三方与生成物：芋道(Yudao Cloud，MIT)基座 huijing-framework/gateway/module-system/module-infra/module-bpm/module-pay/server/dependencies、node_modules、target、各模块 src/test 测试代码、littlejsengine、spikes、*.min.js 等。）",
]
out = header + [RULE, "【前 1500 行】"] + front \
    + [RULE, "【中间部分省略：第 %d 行 至 第 %d 行，共省略 %d 行】" % (omit_from, omit_to, omit_cnt), RULE, "【后 1500 行】"] + back
open("/tmp/ru4_new.txt", "w", encoding="utf-8").write("\n".join(out) + "\n")

print("FILES=%d  SRC_LINES=%d  CONCAT=%d  NONBLANK=%d  BLANK=%d  OMIT=%d (%d-%d)  TXT_LINES=%d" % (
    nfiles, nsrc, total, nonblank, blank, omit_cnt, omit_from, omit_to, len(out)))
for m, (c, l) in mod.items():
    print("  %-26s %d files / %d lines" % (m, c, l))
if no_trailing_nl:
    print("  [WARN 无尾换行]", no_trailing_nl)
print("  BOUNDARY concat[1500]=%r" % (front[1499],))
# 各模块清单表行(刷新《源文件清单.md》用)
for mm in MODULES:
    rows = ["| %s | %d |" % (rel, wcl) for (m, rel, wcl, sl) in files if m == mm]
    open("/tmp/ru4_tbl_%s.md" % mm, "w", encoding="utf-8").write("\n".join(rows) + "\n")