"""xlsx 版式元数据的流式扫描:openpyxl read_only 拿不到的,这里要拿全、拿准。 导入的第二遍按这里算出的边界读值(tabular.load_into),所以边界错了就是数据错了: `` 陈旧时少读一截、离群格把区域撑成几千万格。夹具都用 openpyxl 现造, 需要 openpyxl 写不出来的形态(缓存值、无坐标的单元格、DTD)时直接改压缩包里的 XML。 """ from __future__ import annotations import io import re import zipfile from collections.abc import Callable from dataclasses import asdict import pytest from openpyxl import Workbook from app.data import xlsx_scan from app.data.xlsx_scan import UnsupportedTable, scan def save(book: Workbook) -> bytes: buf = io.BytesIO() book.save(buf) return buf.getvalue() def patch(raw: bytes, edits: dict[str, Callable[[str], str]]) -> bytes: """改压缩包里的部件:{路径: 文本 → 文本}。""" src = zipfile.ZipFile(io.BytesIO(raw)) out = io.BytesIO() with zipfile.ZipFile(out, "w", zipfile.ZIP_DEFLATED) as dst: for info in src.infolist(): data = src.read(info.filename) if info.filename in edits: data = edits[info.filename](data.decode()).encode() dst.writestr(info, data) return out.getvalue() def cache(values: dict[str, str]) -> Callable[[str], str]: """给 openpyxl 写出的公式补上缓存值(模拟 Excel 保存过)。""" def edit(text: str) -> str: def fill(m: re.Match[str]) -> str: ref = m.group(1) return f'{m.group(2)}{values[ref]}' if ref in values else m.group(0) return re.sub(r'([^<]*)', fill, text) return edit SHEET1 = "xl/worksheets/sheet1.xml" # -------------------------------------------------------------------------- # 边界 # -------------------------------------------------------------------------- def test_bounds_come_from_cells_not_from_dimension(): """ 是写文件的程序自己填的,可能是陈旧的:真实边界只能从非空格算。""" book = Workbook() ws = book.active ws["C5"] = "地区" ws["D5"] = "金额" for r in range(6, 26): ws.cell(r, 3, f"分区{r}") ws.cell(r, 4, r) raw = patch(save(book), {SHEET1: lambda t: re.sub(r'', '', t)}) sheet = scan(raw).sheets[0] assert sheet.dimension == "A1:B2" # 记下来,但不信 assert sheet.bounds is not None assert sheet.bounds.a1() == "C5:D25" assert sheet.nonempty == 42 def test_missing_dimension_is_fine(): """openpyxl 的 write_only 根本不写 。""" book = Workbook(write_only=True) ws = book.create_sheet("明细") ws.append(["a", "b"]) for i in range(30): ws.append([i, i * 2]) sheet = scan(save(book)).sheets[0] assert sheet.dimension is None assert sheet.bounds.a1() == "A1:B31" def test_blank_strings_do_not_count_but_formulas_do(): """只有空白的文本算空(导入时也按空值存);有公式的格不算空,哪怕没有缓存值。""" book = Workbook() ws = book.active ws["A1"] = "x" ws["B1"] = " " # 共享字符串里的空白 ws["E9"] = "=A1" # 没有缓存值的公式 sheet = scan(save(book)).sheets[0] assert sheet.nonempty == 2 assert sheet.bounds.a1() == "A1:E9" def test_inline_strings_and_cells_without_coordinates(): """有的程序写内联字符串、省略行号和单元格坐标:位置按顺序推出来。""" book = Workbook() book.active["A1"] = "占位" raw = save(book) data = ('甲1' ' 乙' '2') raw = patch(raw, {SHEET1: lambda t: re.sub(r".*", data, t, flags=re.S)}) sheet = scan(raw).sheets[0] assert sheet.bounds.a1() == "A1:C2" assert sheet.nonempty == 4 # 「 」是空白,不算 def test_far_cells_name_the_outliers(): """稀疏判定拒收时要说出是哪几个远处的格子:A 列一百行数据、XFD 列冒出来一个。""" book = Workbook() ws = book.active for r in range(1, 101): ws.cell(r, 1, r) ws["XFD5"] = "误输入" ws["A90000"] = 1 sheet = scan(save(book)).sheets[0] assert sheet.bounds.a1() == "A1:XFD90000" assert sheet.far_cells == ["XFD5", "A90000"] def test_a_far_cell_jumping_both_ways_is_named_once(): """右下角一个格子(XFD1048576)同时撑出了右边界和下边界:示例里只出现一次。""" book = Workbook() ws = book.active for r in range(1, 201): ws.cell(r, 1, r) ws.cell(r, 2, r) ws["XFD1048576"] = "x" assert scan(save(book)).sheets[0].far_cells == ["XFD1048576"] # -------------------------------------------------------------------------- # 版式 # -------------------------------------------------------------------------- def test_merged_hidden_and_autofilter(): book = Workbook() ws = book.active for r in range(1, 12): ws.append([f"r{r}", r, r * 10, r * 100]) ws.merge_cells("A1:B1") ws.merge_cells("C3:D4") for r in (3, 4, 5, 9): ws.row_dimensions[r].hidden = True ws.column_dimensions["C"].hidden = True ws.auto_filter.ref = "A1:D11" ws.auto_filter.add_filter_column(0, ["r2"]) sheet = scan(save(book)).sheets[0] assert sorted(sheet.merged) == ["A1:B1", "C3:D4"] assert sheet.merged_total == 2 assert sheet.hidden_rows == [(3, 5), (9, 9)] # 连续的合成区间 assert sheet.hidden_cols == [(3, 3)] assert sheet.autofilter == "A1:D11" # 不是公式:字节级查找 b"