#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""生成《2号人事部·功能仿写开发需求文档》(Markdown)。
数据来源: SQLite features(172功能页释义/开发提示) + downloads(41个真实导出Excel, 提取表头做字段字典)
        + screenshots(整页界面参考)。
输出: docs/开发需求文档.md
"""
import os, sqlite3, collections, re, datetime

BASE = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
DB = os.path.join(BASE, "db", "research.db")
DL_ROOT = os.path.join(BASE, "downloads")
OUT = os.path.join(BASE, "docs", "开发需求文档.md")
EXCEL_EXT = (".xlsx", ".xls")

def db():
    c = sqlite3.connect(DB)
    c.row_factory = sqlite3.Row
    return c


def load_tree():
    c = db()
    rows = [dict(r) for r in c.execute(
        "SELECT node_key,parent_key,level,name,full_path,url,description,understanding,"
        "screenshot_path,keep,sort_order FROM features ORDER BY sort_order,id")]
    c.close()
    by = collections.defaultdict(list)
    for r in rows:
        by[r["parent_key"] or ""].append(r)
    return by


def excel_headers(path, max_scan=6):
    """读取 Excel(.xlsx 用 openpyxl / .xls 用 xlrd), 返回识别出的表头字段列表
    (取前若干行中非空单元格最多的一行)。"""
    try:
        if path.lower().endswith(".xls"):
            import xlrd, warnings
            warnings.filterwarnings("ignore")
            wb = xlrd.open_workbook(path)
            ws = wb.sheet_by_index(0)
            best, best_cnt = [], 0
            for r in range(min(max_scan, ws.nrows)):
                vals = [str(ws.cell_value(r, cc)).strip() for cc in range(ws.ncols)
                        if str(ws.cell_value(r, cc)).strip()]
                if len(vals) > best_cnt:
                    best, best_cnt = vals, len(vals)
            return [] if best_cnt == 1 else best
        import openpyxl
        wb = openpyxl.load_workbook(path, read_only=True, data_only=True)
        ws = wb.active
        best, best_cnt = [], 0
        for row in ws.iter_rows(min_row=1, max_row=max_scan):
            vals = [str(c.value).strip() for c in row if c.value is not None and str(c.value).strip()]
            if len(vals) > best_cnt:
                best, best_cnt = vals, len(vals)
        wb.close()
        # 去掉过长(可能是合并标题)的单格行
        if best_cnt == 1:
            return []
        return best
    except Exception as e:
        return []


def scan_downloads():
    """返回 {module_name: [(dir, filename, headers[])]}。module_name 已去数字前缀。"""
    out = collections.defaultdict(list)
    files = []
    for root, _, fs in os.walk(DL_ROOT):
        for f in sorted(fs):
            if f.lower().endswith(EXCEL_EXT):
                files.append((root, f))
    for root, f in files:
        rel = os.path.relpath(root, DL_ROOT)
        segs = [] if rel == "." else rel.split(os.sep)
        seg0 = segs[0] if segs else ""
        module = "任务中心(异步导出)" if seg0.startswith("_任务") else re.sub(r"^\d+_", "", seg0)
        headers = excel_headers(os.path.join(root, f))
        out[module].append((rel.replace(os.sep, "/"), f, headers))
    return out


def slug(s):
    return re.sub(r"[^\w\u4e00-\u9fa5]+", "-", s or "").strip("-")


def load_approval_templates():
    """读取审批模板库清单(若表不存在返回空)。"""
    try:
        c = db()
        rows = [dict(r) for r in c.execute(
            "SELECT name,category,provider,description FROM approval_templates ORDER BY sort_order,id")]
        c.close()
        return rows
    except Exception:
        return []


def main():
    by = load_tree()
    dl = scan_downloads()
    total_files = sum(len(v) for v in dl.values())

    modules = by.get("", [])
    leaves_count = 0

    def count_leaves(pk):
        n = 0
        for r in by.get(pk, []):
            if by.get(r["node_key"]):
                n += count_leaves(r["node_key"])
            else:
                n += 1
        return n

    L = []
    w = L.append
    now = datetime.datetime.now().strftime("%Y-%m-%d %H:%M")
    w("# 2号人事部 · 功能仿写开发需求文档\n")
    w("> **数据来源**：https://2haohr.com 全站功能调研\n>")
    w("> **调研规模**：%d 个一级模块 / %d 个功能页\n>" % (len(modules), sum(count_leaves(m["node_key"]) for m in modules)))
    w("> **配套资产**：整页界面截图 `screenshots/`｜真实导出数据样本 `downloads/`（%d 个 Excel）｜交互式审核+样本下载界面 `http://localhost:8777`\n>" % total_files)
    w("> **生成时间**：%s\n" % now)
    w("\n本文档用于**仿写开发**：每个功能页含【功能释义】【开发提示】【界面参考截图】，"
      "并在模块末尾附**真实导出数据的字段清单**（直接来自线上导出的 Excel 表头），可据此设计数据库表结构与列表字段。\n")

    # 概述表
    w("\n## 一、模块总览\n")
    w("| # | 模块 | 功能页数 | 抓取的数据样本 |")
    w("|---|------|:---:|:---:|")
    for i, m in enumerate(modules, 1):
        nm = m["name"]
        cnt = count_leaves(m["node_key"])
        samp = len(dl.get(nm, []))
        w("| %d | %s | %d | %s |" % (i, nm, cnt, (str(samp) + " 个" if samp else "—")))

    # 明细
    w("\n## 二、模块功能明细\n")
    for mi, m in enumerate(modules, 1):
        w("\n### %d. %s\n" % (mi, m["name"]))

        def walk(pk, depth):
            for r in by.get(pk, []):
                children = by.get(r["node_key"])
                if children:
                    if r["level"] >= 1:
                        w("\n#### ▸ %s\n" % r["name"])
                    walk(r["node_key"], depth + 1)
                else:
                    nonlocal leaves_count
                    leaves_count += 1
                    keep = "" if r["keep"] in (None, 1) else " ~~(审核标记删除)~~"
                    w("\n**%s**%s" % (r["name"], keep))
                    if r["url"]:
                        w("- `URL` %s" % r["url"])
                    if r["description"]:
                        w("- **功能释义**：%s" % r["description"])
                    if r["understanding"] and "待补充细化" not in r["understanding"]:
                        w("- **开发提示**：%s" % r["understanding"])
                    if r["screenshot_path"]:
                        w("- **界面参考**：`%s`" % r["screenshot_path"])
                    w("")

        walk(m["node_key"], 0)

        # 模块导出样本字段
        samples = dl.get(m["name"], [])
        if samples:
            w("\n> **📎 本模块真实导出数据样本（%d 个）与字段：**\n" % len(samples))
            for rel, fn, headers in samples:
                name = os.path.splitext(fn)[0]
                if headers:
                    fields = "、".join(headers[:40])
                    more = " …" if len(headers) > 40 else ""
                    w("> - **%s**（%d 字段）：%s%s" % (name, len(headers), fields, more))
                else:
                    w("> - **%s**（字段未能解析，见文件）" % name)
            w("")

    # 附: 任务中心异步导出
    if dl.get("任务中心(异步导出)"):
        w("\n## 三、异步导出样本（任务中心生成）\n")
        w("> 以下报表在页面点击导出后由服务端异步生成，经“任务中心”收割：\n")
        for rel, fn, headers in dl["任务中心(异步导出)"]:
            name = os.path.splitext(fn)[0]
            fields = ("：" + "、".join(headers[:40])) if headers else ""
            w("> - **%s**%s" % (name, fields))
        w("")

    # 附: 审批模板库清单
    tpls = load_approval_templates()
    if tpls:
        w("\n## 四、审批模板库清单（预置审批流程）\n")
        w("> 来自 `审批 / 常规 / 审批模板库`（%d 个预置流程模板）。这些是**审批流程模板**"
          "（点“使用”实例化为审批表单），非数据表，无导出文件；仿写时可据此设计审批流程配置。\n" % len(tpls))
        bycat = collections.OrderedDict()
        for t in tpls:
            bycat.setdefault(t["category"] or "其他", []).append(t)
        for cat, items in bycat.items():
            w("\n#### ▸ %s（%d）\n" % (cat, len(items)))
            w("| 模板 | 提供方 | 说明 |")
            w("|------|------|------|")
            for t in items:
                w("| %s | %s | %s |" % (t["name"], t["provider"] or "—", t["description"] or ""))
        w("")


    w("\n---\n")
    w("*本文档由调研数据自动汇编。功能释义/开发提示可在 `http://localhost:8777` 审核界面逐条修订后重新生成；"
      "字段清单来自线上真实导出 Excel 的表头，为一手数据。*\n")

    os.makedirs(os.path.dirname(OUT), exist_ok=True)
    with open(OUT, "w", encoding="utf-8") as f:
        f.write("\n".join(L))
    print("written:", OUT)
    print("modules:", len(modules), "leaves:", leaves_count, "sample files:", total_files)


if __name__ == "__main__":
    main()