proposa-ai/scripts/tests/test_step6_assembler.py

"""Step 6 聚合装配的快速回归测试（不调用 LLM）。"""

import os
import tempfile
import unittest
import zipfile
import xml.etree.ElementTree as ET

from docx import Document as DocxDocument

from models import BidOutline
from step6_exporting.assembler import (
    _HeadingRecord,
    _build_index_table_from_template,
    _load_table_template,
    assemble_step5_document,
)


_W_NS = "http://schemas.openxmlformats.org/wordprocessingml/2006/main"


def _w(tag: str) -> str:
    return f"{{{_W_NS}}}{tag}"


def _element_text(element) -> str:
    return "".join(node.text or "" for node in element.iter(_w("t")))


def _make_chapter(path: str, title: str, sub_headings) -> None:
    doc = DocxDocument()
    section = doc.sections[0]
    section.header.paragraphs[0].text = "章节原生页眉"
    section.footer.paragraphs[0].text = "章节原生页脚"
    doc.add_paragraph(title, style="Heading 1")
    for heading in sub_headings:
        doc.add_paragraph(heading, style="Heading 2")
        doc.add_paragraph(f"这是 {heading} 的响应正文，包含数字 123 和英文 English。")
    doc.save(path)


def _make_template(path: str) -> None:
    doc = DocxDocument()
    section = doc.sections[0]
    section.different_first_page_header_footer = True
    default_header = section.header
    default_header.paragraphs[0].text = "默认页眉"
    first_header = section.first_page_header
    first_header.paragraphs[0].text = "首页页眉"
    footer = section.footer
    footer.paragraphs[0].text = "页脚"
    doc.add_paragraph("模板占位正文")
    doc.save(path)


def _make_index_table_template(path: str) -> None:
    doc = DocxDocument()
    table = doc.add_table(rows=2, cols=4)
    header = ["序号", "主要内容概述", "章节", "详细内容所在投标文件页次"]
    for col, text in enumerate(header):
        table.cell(0, col).text = text
    for col, text in enumerate(["1", "示例", "第一章", "P1"]):
        table.cell(1, col).text = text
    doc.save(path)


class Step6AssemblerTests(unittest.TestCase):
    def test_assemble_produces_cover_toc_index_and_chapters(self):
        temp_dir = tempfile.mkdtemp()
        template_path = os.path.join(temp_dir, "template.docx")
        _make_template(template_path)

        ch1 = os.path.join(temp_dir, "1.docx")
        ch2 = os.path.join(temp_dir, "2.docx")
        ch3 = os.path.join(temp_dir, "3.docx")
        _make_chapter(ch1, "第一章 投标人资格、资信证明", ["十七、“★” 要求承诺函"])
        _make_chapter(ch2, "第二章 投标报价", ["五、报价得分"])
        _make_chapter(ch3, "第三章 需求理解", ["一、服务目标定位"])

        outline = BidOutline(project_name="测试项目")
        outline.evaluation_index_entries = [
            {
                "entry_type": "scoring",
                "source_id": "SC-01#1",
                "criterion_id": "SC-01",
                "display_name": "报价得分",
                "requirement": "报价得分＝报价分值×（评标基准价/评审价）",
                "score": 10.0,
                "final_heading_id": "2.5",
                "final_heading_title": "五、报价得分",
                "final_heading_path": "投标报价 → 五、报价得分",
            },
            {
                "entry_type": "rejection",
                "source_id": "RI-01",
                "criterion_id": "",
                "display_name": "资格条件",
                "requirement": "投标人不满足《中华人民共和国政府采购法》第二十二条规定。",
                "score": None,
                "final_heading_id": "1.17",
                "final_heading_title": "十七、“★” 要求承诺函",
                "final_heading_path": "投标人资格、资信证明 → 十七、“★” 要求承诺函",
            },
        ]

        output_path = os.path.join(temp_dir, "aggregated.docx")
        records = [
            {"id": "1", "title": "投标人资格、资信证明", "artifact_path": ch1, "status": "complete"},
            {"id": "2", "title": "投标报价", "artifact_path": ch2, "status": "complete"},
            {"id": "3", "title": "需求理解", "artifact_path": ch3, "status": "complete"},
        ]

        report = assemble_step5_document(
            records,
            outline,
            output_path,
            template_path=template_path,
            project_name="测试项目",
        )

        self.assertTrue(os.path.isfile(output_path))
        self.assertEqual(report.index_row_count, 1)
        self.assertTrue(report.toc_inserted)
        self.assertEqual(report.chapter_count, 3)

        with zipfile.ZipFile(output_path, "r") as output_zip:
            names = set(output_zip.namelist())
            self.assertIn("word/document.xml", names)
            document_root = ET.fromstring(output_zip.read("word/document.xml"))
            body = document_root.find(_w("body"))
            self.assertIsNotNone(body)
            body_text = _element_text(body)
            self.assertIn("测试项目", body_text)
            self.assertIn("目  录", body_text)
            self.assertIn("第一章 投标人资格、资信证明", body_text)
            self.assertIn("第二章 投标报价", body_text)
            self.assertIn("第三章 需求理解", body_text)
            self.assertIn("商务部分", body_text)
            self.assertIn("技术部分", body_text)
            self.assertIn("主要内容概述", body_text)

            instr_texts = [node.text or "" for node in body.iter(_w("instrText"))]
            combined = " ".join(instr_texts)
            self.assertIn("TOC", combined)
            self.assertIn("PAGEREF", combined)
            self.assertNotIn("第二十二条", body_text)

            header1 = output_zip.read("word/header1.xml")
            header2 = output_zip.read("word/header2.xml")
            footer1 = output_zip.read("word/footer1.xml")
            self.assertIn("章节原生页眉".encode("utf-8"), header1)
            self.assertIn("章节原生页脚".encode("utf-8"), footer1)

            # 正文 run 应统一为宋体 + Times New Roman。
            for run in body.iter(_w("r")):
                if _element_text(run).startswith("这是"):
                    r_pr = run.find(_w("rPr"))
                    self.assertIsNotNone(r_pr)
                    fonts = r_pr.find(_w("rFonts"))
                    self.assertIsNotNone(fonts)
                    self.assertEqual(fonts.get(_w("eastAsia")), "宋体")
                    self.assertEqual(fonts.get(_w("ascii")), "Times New Roman")
                    break
            else:
                self.fail("未找到正文 run 以验证字体")

    def test_index_table_reuses_extracted_template(self):
        temp_dir = tempfile.mkdtemp()
        table_template_path = os.path.join(temp_dir, "index_table.docx")
        _make_index_table_template(table_template_path)

        template_tbl = _load_table_template(table_template_path)
        self.assertIsNotNone(template_tbl)

        records = [
            _HeadingRecord(chapter_title="投标报价", text="第二章 投标报价", level=1, bookmark="CodexIdx1"),
            _HeadingRecord(chapter_title="投标报价", text="五、报价得分", level=2, bookmark="CodexIdx2"),
        ]
        entries = [
            {
                "entry_type": "scoring",
                "display_name": "报价得分",
                "score": 10.0,
                "final_heading_path": "投标报价 → 五、报价得分",
            }
        ]

        table, row_count = _build_index_table_from_template(template_tbl, entries, records)
        self.assertEqual(row_count, 1)
        rows = table.findall(_w("tr"))
        self.assertEqual(len(rows), 2)
        header_cells = [_element_text(tc).strip() for tc in rows[0].findall(_w("tc"))]
        self.assertEqual(header_cells, ["序号", "主要内容概述", "章节", "详细内容所在投标文件页次"])
        data_cells = [_element_text(tc).strip() for tc in rows[1].findall(_w("tc"))]
        self.assertEqual(data_cells[0], "1")
        self.assertIn("报价得分", data_cells[1])
        self.assertEqual(data_cells[2], "第二章 投标报价 → 五、报价得分")


if __name__ == "__main__":
    unittest.main()