#!/usr/bin/env python3
"""生成样例文档: clean.docx（应返回空清单）与 drifted.docx（人为埋入 3 处漂移）。

用法: python make_samples.py [params.example.json] [输出目录]

正文与表格用 {{key}} / {{key:单位}} / {{key:单位:,}} 占位符从 params.json 渲染，
模拟「参数单一来源 + 模板渲染」的流程。drifted.docx 在渲染后手工改动 3 处:
  1. 单位漂移: 表格 1「小时流量」行的单位格 m³/h 被改成 m³/d（数字 62.5 未变）
  2. 修约漂移: 正文「吨水电耗 0.62 kWh/t」被写成 0.6（表格里仍是 0.62）
  3. 派生量漂移: 表格 2「PAC年药剂费」仍是按旧水量 1200 t/d 算的 30660 元（应为 38325 元）
"""
import re
import sys
from pathlib import Path

import docx

from check_consistency import load_params

HERE = Path(__file__).resolve().parent

TITLE = "某工业园区 {{treated_flow_tpd}} t/d 污水处理站工程方案"
BODY = [
    ("h1", "1 项目概况"),
    ("p", "本工程设计处理水量 {{treated_flow_tpd}} t/d（{{treated_flow_tpd:m³/h}} m³/h），"
          "进水COD {{cod_in_mgl}} mg/L，出水COD {{cod_out_mgl}} mg/L，COD去除率 {{cod_removal_pct}}%。"),
    ("p", "方案编制日期：2026年10月7日；出水执行 GB 18918-2002 一级A 标准。"),
    ("h1", "2 工艺设计"),
    ("p", "调节池水力停留时间 {{hrt_h}} h，共设 {{basin_count}} 座。"),
    ("p", "混凝段 PAC投加量 {{pac_dose_mgl}} mg/L，按设计水量计 PAC日耗量 {{pac_daily_kgd}} kg/d。"),
    ("p", "吨水电耗 {{power_kwh_per_t}} kWh/t，日耗电量 {{daily_power_kwhd}} kWh/d。"),
    ("p", "PAC单价 {{pac_price_yuan_kg}} 元/kg，PAC年药剂费 {{pac_annual_cost_wan}} 万元。"),
    ("h1", "3 投资估算"),
    ("p", "设备费 {{equipment_cost_wan}} 万元，安装费 {{install_cost_wan}} 万元，"
          "工程总报价 {{total_quote_wan}} 万元。主要参数与费用见表 1、表 2。"),
]
TABLE1 = [["序号", "指标", "数值", "单位"],
          ["1", "设计水量", "{{treated_flow_tpd}}", "t/d"],
          ["2", "小时流量", "{{treated_flow_tpd:m³/h}}", "m³/h"],
          ["3", "进水COD", "{{cod_in_mgl}}", "mg/L"],
          ["4", "出水COD", "{{cod_out_mgl}}", "mg/L"],
          ["5", "PAC投加量", "{{pac_dose_mgl}}", "mg/L"],
          ["6", "PAC日耗量", "{{pac_daily_kgd}}", "kg/d"],
          ["7", "吨水电耗", "{{power_kwh_per_t}}", "kWh/t"],
          ["8", "日耗电量", "{{daily_power_kwhd}}", "kWh/d"]]
TABLE2 = [["序号", "项目", "金额"],
          ["1", "设备费", "{{equipment_cost_wan:元:,}} 元"],
          ["2", "安装费", "{{install_cost_wan}} 万元"],
          ["3", "PAC年药剂费", "{{pac_annual_cost_wan:元}} 元"],
          ["4", "工程总报价", "{{total_quote_wan}} 万元"]]


def render(tpl, Q, cfg):
    def sub(m):
        key, unit, sep = (m.group(1).split(":") + ["", ""])[:3]
        q = Q[key]
        unit = unit or q["unit"]
        v = q["value"] * cfg["units"][q["unit"]][1] / cfg["units"][unit][1]
        d = q["unit_decimals"].get(unit, q["decimals"])
        return f"{v:{sep}.{d}f}"
    return re.sub(r"\{\{([^{}]+)\}\}", sub, tpl)


def build(out_path, params_path, drift=False):
    Q, cfg = load_params(params_path)
    body = [(kind, render(t, Q, cfg)) for kind, t in BODY]
    t1 = [[render(c, Q, cfg) for c in row] for row in TABLE1]
    t2 = [[render(c, Q, cfg) for c in row] for row in TABLE2]
    if drift:
        t1[2][3] = "m³/d"                                                   # 1. 单位
        body = [(k, t.replace("吨水电耗 0.62", "吨水电耗 0.6")) for k, t in body]  # 2. 修约
        t2[3][2] = "30660 元"                                                # 3. 派生量
    doc = docx.Document()
    doc.add_heading(render(TITLE, Q, cfg), level=0)
    for kind, text in body:
        doc.add_heading(text, level=1) if kind == "h1" else doc.add_paragraph(text)
    for caption, rows in (("表 1 主要设计参数", t1), ("表 2 投资与运行费用", t2)):
        doc.add_paragraph(caption)
        table = doc.add_table(rows=len(rows), cols=len(rows[0]))
        table.style = "Table Grid"
        for r, row in enumerate(rows):
            for c, val in enumerate(row):
                table.cell(r, c).text = val
    doc.save(out_path)
    return out_path


if __name__ == "__main__":
    params = sys.argv[1] if len(sys.argv) > 1 else str(HERE / "params.example.json")
    out = Path(sys.argv[2]) if len(sys.argv) > 2 else HERE
    print(build(out / "clean.docx", params))
    print(build(out / "drifted.docx", params, drift=True))