iomgaa
7e7f32a13d
- ceus：灌注指标改为对病灶去暗区（roi_perf）和紫圈（roi_pan）两组算，roi_all 指标删除； D1–D4 仍要 roi_all、roi_perf 两条都过才定位（负责人 2026-10-17：医生要正常区与病灶对比看） - build/library：tic.json 只出两组 11 个指标与 7 个槽位的视频时间；医生截图挪进 ultrasound/keyframes/key.png，UltrasoundRoute 去掉 key_image、keyframes 多一个 key - review：TIC参数.csv 取代 TIC结果.csv；问题清单按原因对照表写成大白话，对照表外的原因报错； 程序原文进 snapshot/问题清单_原文.csv；说明.pdf 整份重写为合作者口吻 - 文档：结构.md、医生核对包、装配两个库、CEUS定量 任务书 04。 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
596 lines
26 KiB
Python
596 lines
26 KiB
Python
"""原因对照表：每种程序原因 → 给医生的那句话，逐条手写预期（任务书 04「原因的写法」）。
|
||
|
||
程序原因按源码里的写法造（`_anchor_frames.py`、`_tic_fitting.py`、`ceus/` 各步的异常、
|
||
`ceus_quant.py` 的目视核对、delivery 的 C01–C13、build 的 C14、C15），数值是编的。
|
||
"""
|
||
|
||
import ast
|
||
import json
|
||
import re
|
||
from pathlib import Path
|
||
from typing import Any
|
||
|
||
import pytest
|
||
|
||
from cadx.answers import SkippedRow
|
||
from cadx.review._problems import (
|
||
UnknownReasonError,
|
||
case_label,
|
||
ceus_failure,
|
||
check_plain,
|
||
curve_problems,
|
||
group_reason,
|
||
skip_problem,
|
||
)
|
||
from fake_ceus_run import summary
|
||
|
||
FORBIDDEN = re.compile(r"C[0-9]|roi_|_|\{|\[|band|[0-9]{8}-[0-9]{6}")
|
||
BROKEN = "视频文件本身有问题（帧率、帧数或时间信息不对），程序读不了"
|
||
NO_SPLIT = "视频里找不到左右两屏的分界线"
|
||
MISALIGNED = "视频左右两屏的位置对不齐"
|
||
|
||
|
||
@pytest.mark.parametrize(
|
||
("step", "reason", "expected"),
|
||
[
|
||
(1, "配准目视核对未通过", f"{MISALIGNED}（我们人工逐例看过，这一例没对上）"),
|
||
(1, "cv2 打不开视频：/x/U001_roi2.mp4", "视频文件打不开"),
|
||
(1, "探测到非法 fps=0.0：/x/U001.mp4", BROKEN),
|
||
(1, "探测到非法元数据 frame_count=0 width=1 height=1：/x/U001.mp4", BROKEN),
|
||
(1, "视频包含非法显示时间戳 -1.0：/x/U001.mp4", BROKEN),
|
||
(1, "视频显示时间戳不严格递增：第 3 帧 0.1，前一帧 0.2：/x/U001.mp4", BROKEN),
|
||
(1, "时间戳帧数与容器记录的帧数不一致：3 / 4：/x/U001.mp4", BROKEN),
|
||
(1, "解码帧数 3 与容器记录的帧数 4 不一致：/x/U001.mp4", BROKEN),
|
||
(1, "帧尺寸不一致：首帧 (4, 6, 3)，第 2 帧 (4, 8, 3)。", "视频前后画面大小不一致"),
|
||
(1, "帧序列为空，无法求跨帧平均灰度。", "视频里一帧画面都没有"),
|
||
(1, "帧序列为空，无法找分屏线。", "视频里一帧画面都没有"),
|
||
(
|
||
1,
|
||
"中部 band 内未找到稳定暗缝（谷值 40.2 未显著低于 中位 51.0），疑为无缝或全黑画面。",
|
||
NO_SPLIT,
|
||
),
|
||
(1, "非法 search_band：要求 0<=lo<hi<=1，收到 (0.6, 0.4)。", NO_SPLIT),
|
||
(1, "search_band 在宽度 10 下退化为空区间 [4, 4)。", NO_SPLIT),
|
||
(1, "检测到的 split_x=9 落在搜索 band [3, 7) 外。", NO_SPLIT),
|
||
(1, "切分线 split_x=0 使某一侧宽度退化（左 0 / 右 5）。", NO_SPLIT),
|
||
(
|
||
1,
|
||
"左右宽度比 0.30 超出 [0.5, 2.0]（左 100 / 右 330），切分线不可信。",
|
||
"找到的左右两屏分界线不可信（两屏宽窄差太多）",
|
||
),
|
||
(
|
||
1,
|
||
"这一半画面里找不到扇形区（没有随时间变化的区域）",
|
||
"视频有一屏里找不到超声扇形画面",
|
||
),
|
||
(
|
||
1,
|
||
"左半扇形内可用的模板块只有 2 块，少于 4 块",
|
||
f"{MISALIGNED}（画面里能拿来比对的细节太少）",
|
||
),
|
||
(
|
||
1,
|
||
"两半配准不一致：互相关估计 (dx, dy) = (3, 4)，模板匹配估计 (9, 4)，"
|
||
"相差 6 像素，超过门槛 4",
|
||
MISALIGNED,
|
||
),
|
||
(
|
||
2,
|
||
"左右两半都找到绿框（左半 3 个提示帧，右半 2 个），定不出造影画面在哪一半",
|
||
"视频左右两屏都有绿框，分不出哪一屏是造影画面",
|
||
),
|
||
(2, "左右两半都没有找到绿框，选不出提示帧", "视频里找不到画好的绿框"),
|
||
(
|
||
4,
|
||
"逐帧 mask 未产出任何有效 ROI 轮廓，无从时间插值。",
|
||
"整段视频里程序都没能画出病灶轮廓",
|
||
),
|
||
],
|
||
)
|
||
def test_ceus_failure(step: int, reason: str, expected: str) -> None:
|
||
assert ceus_failure(step, reason) == f"{expected}，程序没法处理这段视频"
|
||
|
||
|
||
REPO = Path(__file__).resolve().parents[1]
|
||
# ceus 各步一例失败时抛的 CeusCaseError 子类 → 第几步（各类的 *_STEP 常量）。
|
||
_CASE_ERRORS = {
|
||
"VideoDecodeError": 1,
|
||
"SplitDetectionError": 1,
|
||
"RegistrationError": 1,
|
||
"ContrastSideError": 2,
|
||
"NoValidContourError": 4,
|
||
}
|
||
|
||
|
||
def _raised_messages() -> list[tuple[str, int, str]]:
|
||
"""源码里每处 raise 一例失败的地方：（位置，第几步，消息，格式化部分代成 1）。"""
|
||
found: list[tuple[str, int, str]] = []
|
||
files = [
|
||
*sorted((REPO / "src" / "cadx" / "ceus").glob("*.py")),
|
||
REPO / "scripts" / "ceus_quant.py",
|
||
]
|
||
for path in files:
|
||
for node in ast.walk(ast.parse(path.read_text(encoding="utf-8"))):
|
||
if not (isinstance(node, ast.Raise) and isinstance(node.exc, ast.Call)):
|
||
continue
|
||
func = node.exc.func
|
||
name = func.attr if isinstance(func, ast.Attribute) else getattr(func, "id", "")
|
||
args = node.exc.args
|
||
if name in _CASE_ERRORS:
|
||
step, message = _CASE_ERRORS[name], args[0]
|
||
elif name == "CeusCaseError":
|
||
first = args[0]
|
||
assert isinstance(first, ast.Constant) and isinstance(first.value, int)
|
||
step, message = first.value, args[1]
|
||
else:
|
||
continue
|
||
if isinstance(message, ast.Constant):
|
||
text = str(message.value)
|
||
else:
|
||
assert isinstance(message, ast.JoinedStr), f"{path}:{node.lineno}"
|
||
text = "".join(
|
||
str(v.value) if isinstance(v, ast.Constant) else "1" for v in message.values
|
||
)
|
||
found.append((f"{path.name}:{node.lineno}", step, text))
|
||
return found
|
||
|
||
|
||
def test_every_raised_ceus_failure_is_in_the_table() -> None:
|
||
"""ceus 和 ceus_quant 里每一处 raise 一例失败的消息，对照表都认得（新加一处就会在这里报出）。"""
|
||
messages = _raised_messages()
|
||
assert len(messages) >= 20
|
||
for where, step, text in messages:
|
||
assert not FORBIDDEN.search(ceus_failure(step, text)), where
|
||
|
||
|
||
@pytest.mark.parametrize(
|
||
("step", "reason"),
|
||
[(3, "分割模型报错"), (2, "配准目视核对未通过"), (1, "编造的原因")],
|
||
)
|
||
def test_unknown_ceus_failure_stops(step: int, reason: str) -> None:
|
||
with pytest.raises(UnknownReasonError):
|
||
ceus_failure(step, reason)
|
||
|
||
|
||
FIT_CASES = {
|
||
"insufficient_valid_samples": "能用的画面太少（很多帧认不出这块区域，或是重复帧），连不成曲线",
|
||
"candidate_grid_unavailable": "能用的画面时间点太少，连不成曲线",
|
||
"credible_peak_not_found": "曲线太平，没有明显的峰",
|
||
"candidate_arrival_not_found": "峰之前找不到亮度明确、持续升起来的时刻",
|
||
"insufficient_bolus_samples": "造影剂第一次流过的那一段能用的画面太少",
|
||
"long_gap_in_wash_in": "变亮的那一段中间缺了一大段画面",
|
||
"arrival_bounds_invalid": "峰离视频开头太近，或造影剂第一次流过的那一段太短",
|
||
"optimizer_failed": "曲线形状不规则，平滑曲线怎么也贴不上",
|
||
"parameter_on_boundary": (
|
||
"试着贴出来的曲线在峰高、峰出现的时间、曲线宽窄、开始变亮的时刻或增强前水平上，"
|
||
"有一项卡在了程序允许范围的边上"
|
||
),
|
||
"model_peak_on_boundary": "试着贴出来的曲线，峰落在可用那一段的最开头或最末尾，峰的位置不可信",
|
||
"fit_quality_failed": "曲线起伏杂乱，平滑曲线和它差得太多",
|
||
}
|
||
|
||
|
||
@pytest.mark.parametrize(("code", "phrase"), FIT_CASES.items())
|
||
def test_fit_reason_codes(code: str, phrase: str) -> None:
|
||
content = summary("U001", ())
|
||
content["reference_valid"] = True
|
||
content["fits"]["roi_pan"].update(success=False, failure_reason=code)
|
||
assert group_reason(content, "roi_pan") == f"{phrase}，没能拟合"
|
||
content["fits"]["roi_perf"].update(success=False, failure_reason=code)
|
||
assert group_reason(content, "roi_perf") == f"{phrase}，没能拟合"
|
||
|
||
|
||
def test_every_fit_reject_code_is_in_the_table() -> None:
|
||
source = (REPO / "src" / "cadx" / "ceus" / "_tic_fitting.py").read_text(encoding="utf-8")
|
||
codes = set(re.findall(r'reject\("([a-z_]+)"\)', source))
|
||
assert codes == set(FIT_CASES)
|
||
|
||
|
||
def test_group_reason_pan_unusable_and_unknown_code() -> None:
|
||
content = summary("U001", ())
|
||
assert group_reason(content, "roi_pan") == "视频里大部分时间认不出紫圈"
|
||
content["fits"]["roi_perf"].update(success=False, failure_reason="new_code")
|
||
with pytest.raises(UnknownReasonError):
|
||
group_reason(content, "roi_perf")
|
||
|
||
|
||
BOUNDARY = FIT_CASES["parameter_on_boundary"]
|
||
LESION = "；B0、D1–D4 要病灶整体和病灶去暗区两条曲线都拟合上才取，所以这 5 个关键帧没有取"
|
||
FIT_FAILED = "病灶两条曲线（roi_all、roi_perf）的 Lognormal 拟合没有都过质量门"
|
||
|
||
|
||
def _curves(located: tuple[str, ...], edit: dict[str, Any]) -> list[tuple[str, str]]:
|
||
content = summary("U001", located)
|
||
content["reference_valid"] = True
|
||
content["fits"]["roi_pan"].update(success=True, failure_reason=None)
|
||
for slot, reason in edit.items():
|
||
content["anchor_frames"][slot]["reason"] = reason
|
||
return curve_problems(content)
|
||
|
||
|
||
ALL = ("B0", "D1", "D2", "D3", "D4", "F1", "F2")
|
||
|
||
|
||
@pytest.mark.parametrize(
|
||
("slot", "reason", "expected"),
|
||
[
|
||
(
|
||
"B0",
|
||
"B0 锚点 3.70 秒的容差窗口 [3.20, 4.20] 秒 内，"
|
||
"观测到的基线段里、到达时刻之前没有可用帧",
|
||
"B0 没有取：开始增强之前的那一小段里没有能用的画面",
|
||
),
|
||
(
|
||
"B0",
|
||
"拟合曲线升不过到达门槛，没有到达时刻",
|
||
"B0 没有取：曲线上升得太少，定不出造影剂到达的时刻",
|
||
),
|
||
(
|
||
"D1",
|
||
"拟合曲线上没有 D1 时刻（升不过到达门槛）",
|
||
"D1 没有取：曲线上升得太少，定不出造影剂到达的时刻",
|
||
),
|
||
(
|
||
"B0",
|
||
"视频开头没有观测到稳定的增强前基线（基线是估计的），B0 不定义",
|
||
"B0 没有取：视频一开始病灶就已经在变亮，没录到增强之前的那一段",
|
||
),
|
||
(
|
||
"D4",
|
||
"峰后观测不够长，或洗出拐点落在观测范围之外，拟合曲线上没有 D4",
|
||
"D4 没有取：峰值之后录得不够长，找不到变暗最快的时刻",
|
||
),
|
||
(
|
||
"D3",
|
||
"D3 时刻 -0.40 秒落在视频有效帧范围 [0.04, 40.20] 秒 之外",
|
||
"D3 没有取：算出的时刻不在视频能用的画面范围里"
|
||
"（D3 在视频第 -0.4 秒，能用的画面是第 0.0–40.2 秒）",
|
||
),
|
||
(
|
||
"D2",
|
||
"D2 容差窗口 [4.70, 5.30] 秒 内的候选帧都被时间更早的槽位占用，无法保持先后顺序",
|
||
"D2 没有取：这个时刻附近能用的画面都不晚于前一个关键帧取的那一帧"
|
||
"（D1–D4 要按时间先后各取一帧，后一个必须比前一个晚）",
|
||
),
|
||
(
|
||
"D2",
|
||
"D2 容差窗口 [4.70, 5.30] 秒 内没有可用帧（拟合区间内、不重复）",
|
||
"D2 没有取：这个时刻附近没有能用的画面"
|
||
"（附近都是重复帧，或已过了造影剂第一次流过的那一段）",
|
||
),
|
||
(
|
||
"F1",
|
||
"推注后第 20 秒在视频第 -5.00 秒（推注时间 X=25 秒），"
|
||
"落在视频有效帧范围 [0.00, 40.20] 秒之外",
|
||
"F1 没有取：换算到视频里的时刻不在录到的范围内"
|
||
"（F1 是推注后第 20 秒，应在视频第 -5.0 秒，那时还没开始录）",
|
||
),
|
||
(
|
||
"F2",
|
||
"推注后第 60 秒在视频第 55.00 秒（推注时间 X=5 秒），"
|
||
"落在视频有效帧范围 [0.00, 40.20] 秒之外",
|
||
"F2 没有取：换算到视频里的时刻不在录到的范围内"
|
||
"（F2 是推注后第 60 秒，应在视频第 55.0 秒，视频只录到第 40.2 秒）",
|
||
),
|
||
(
|
||
"F2",
|
||
"推注后第 60 秒在视频第 55.00 秒（推注时间 X=5 秒），落在没有不重复的帧之外",
|
||
"F2 没有取：换算到视频里的时刻不在录到的范围内"
|
||
"（F2 是推注后第 60 秒，应在视频第 55.0 秒，视频里没有能用的画面）",
|
||
),
|
||
(
|
||
"F2",
|
||
"视频第 55.00 秒 ±2 秒内没有不重复的帧",
|
||
"F2 没有取：这个时刻前后没有能用的画面（都是重复帧）",
|
||
),
|
||
(
|
||
"F1",
|
||
"没有推注时间，推注后时刻换算不到视频时间",
|
||
"F1 没有取：没有推注时间，换算不出视频里的时刻",
|
||
),
|
||
],
|
||
)
|
||
def test_keyframe_reasons(slot: str, reason: str, expected: str) -> None:
|
||
located = tuple(s for s in ALL if s != slot)
|
||
assert _curves(located, {slot: reason}) == [("关键帧", expected)]
|
||
|
||
|
||
def test_same_reason_merges_into_one_row() -> None:
|
||
reason = "视频第 {} 秒 ±2 秒内没有不重复的帧"
|
||
rows = _curves(("B0", "D1", "D2", "D3", "D4"), {"F1": reason.format(1), "F2": reason.format(2)})
|
||
assert rows == [("关键帧", "F1、F2 没有取：这个时刻前后没有能用的画面（都是重复帧）")]
|
||
|
||
|
||
@pytest.mark.parametrize(
|
||
("failed", "expected"),
|
||
[
|
||
(
|
||
{"roi_all": "parameter_on_boundary"},
|
||
f"病灶整体那条曲线没能拟合（{BOUNDARY}）{LESION}",
|
||
),
|
||
(
|
||
{"roi_all": "parameter_on_boundary", "roi_perf": "parameter_on_boundary"},
|
||
f"病灶整体、病灶去暗区两条曲线都没能拟合（都是{BOUNDARY}）{LESION}"
|
||
"，TIC 参数里病灶一组也空着",
|
||
),
|
||
(
|
||
{"roi_all": "credible_peak_not_found", "roi_perf": "fit_quality_failed"},
|
||
"病灶整体、病灶去暗区两条曲线都没能拟合（病灶整体：曲线太平，没有明显的峰；"
|
||
f"病灶去暗区：曲线起伏杂乱，平滑曲线和它差得太多）{LESION}"
|
||
"，TIC 参数里病灶一组也空着",
|
||
),
|
||
],
|
||
)
|
||
def test_lesion_fit_failure_is_one_row(failed: dict[str, str], expected: str) -> None:
|
||
content = summary("U001", ("F1", "F2"))
|
||
content["reference_valid"] = True
|
||
content["fits"]["roi_pan"].update(success=True, failure_reason=None)
|
||
for curve, code in failed.items():
|
||
content["fits"][curve].update(success=False, failure_reason=code)
|
||
for slot in ("B0", "D1", "D2", "D3", "D4"):
|
||
content["anchor_frames"][slot]["reason"] = FIT_FAILED
|
||
assert curve_problems(content) == [("TIC 曲线", expected)]
|
||
|
||
|
||
def test_roi_perf_fit_reason_counts_as_fit_failure() -> None:
|
||
content = summary("U001", ("F1", "F2"))
|
||
content["reference_valid"] = True
|
||
content["fits"]["roi_pan"].update(success=True, failure_reason=None)
|
||
content["fits"]["roi_perf"].update(success=False, failure_reason="optimizer_failed")
|
||
for slot in ("B0", "D1", "D2", "D3", "D4"):
|
||
content["anchor_frames"][slot]["reason"] = (
|
||
"roi_perf 曲线的 Lognormal 拟合没过质量门（optimizer_failed）"
|
||
)
|
||
assert [stage for stage, _ in curve_problems(content)] == ["TIC 曲线"]
|
||
|
||
|
||
def test_fit_failure_without_its_keyframes_stops() -> None:
|
||
content = summary("U001", ALL)
|
||
content["fits"]["roi_all"].update(success=False, failure_reason="optimizer_failed")
|
||
with pytest.raises(ValueError, match="对不上"):
|
||
curve_problems(content)
|
||
|
||
|
||
def test_pan_rows() -> None:
|
||
content = summary("U001", ALL)
|
||
assert curve_problems(content) == [
|
||
("紫圈", "视频里大部分时间认不出紫圈，紫圈那条曲线不能用，TIC 参数里胰腺实质一组空着")
|
||
]
|
||
content["reference_valid"] = True
|
||
content["fits"]["roi_pan"]["failure_reason"] = "long_gap_in_wash_in"
|
||
assert curve_problems(content) == [
|
||
(
|
||
"TIC 曲线",
|
||
"紫圈那条曲线没能拟合（变亮的那一段中间缺了一大段画面），TIC 参数里胰腺实质一组空着",
|
||
)
|
||
]
|
||
|
||
|
||
def test_unknown_keyframe_reason_stops() -> None:
|
||
with pytest.raises(UnknownReasonError):
|
||
_curves(("B0",), {s: "编造的原因" for s in ALL if s != "B0"})
|
||
|
||
|
||
CLINICAL = "Clinical_agent/临床工作簿_40例.xlsx"
|
||
ULTRASOUND = "CEUS_agent/超声报告文本_40例.xlsx"
|
||
MR = "MR_agent/MR报告文本_40例.xlsx"
|
||
DIAGNOSIS = "病理类型（1 IPMN，2 SPN，3 MCN，4 SCN）"
|
||
|
||
|
||
def _row(case: str, check: str, files: tuple[str, ...], value: object) -> SkippedRow:
|
||
return SkippedRow(
|
||
case=case,
|
||
check=check,
|
||
description="（检查说明不用）",
|
||
files=files,
|
||
columns=(),
|
||
value_json=json.dumps(value, ensure_ascii=False),
|
||
)
|
||
|
||
|
||
NOT_IN = "，这一例没进资料库"
|
||
|
||
|
||
@pytest.mark.parametrize(
|
||
("row", "expected"),
|
||
[
|
||
(
|
||
_row(f"{CLINICAL} 第 5 行", "C01", (CLINICAL,), {"原始写法": "P1", "行号": 5}),
|
||
"临床表第 5 行的病例编号认不出（「P1」）",
|
||
),
|
||
(
|
||
_row("病例目录 1000", "C01", ("CEUS_agent/1000",), {"目录名": "1000"}),
|
||
"病例文件夹名「1000」认不出是哪个病例",
|
||
),
|
||
(
|
||
_row("003", "C02", (MR,), {"行号": [4, 9]}),
|
||
"MR 表里这个病例号出现了 2 次（第 4、9 行）",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C03",
|
||
(ULTRASOUND, "CEUS_agent/003/"),
|
||
{"缺在": [ULTRASOUND, "CEUS_agent/003/"]},
|
||
),
|
||
"在超声表、病例文件夹里找不到这一例",
|
||
),
|
||
(
|
||
_row("003", "C04", ("CEUS_agent/003",), {"文件": ["U003_key.png", "U003.mp4"]}),
|
||
"病例文件夹里应当恰好有截图、造影视频、单例超声报告表三个文件，实际有 2 个",
|
||
),
|
||
(
|
||
_row("003", "C05", (ULTRASOUND,), {"CEUS视频ID": "U030", "病例号": "003"}),
|
||
"超声表里这一例的 CEUS视频ID 写的是「U030」，和病例号对不上",
|
||
),
|
||
(
|
||
_row("003", "C06", ("CEUS_agent/003/U003_report.xlsx",), {"工作表": ["A", "B"]}),
|
||
"单例超声报告表的工作表不对（应只有一张）",
|
||
),
|
||
(
|
||
_row(
|
||
"003", "C06", ("CEUS_agent/003/U003_report.xlsx",), {"表头": "与超声汇总表不一致"}
|
||
),
|
||
"单例超声报告表的表头和超声汇总表不一样",
|
||
),
|
||
(
|
||
_row("003", "C06", ("CEUS_agent/003/U003_report.xlsx",), {"数据行数": 2}),
|
||
"单例超声报告表应只有一行数据，实际有 2 行",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C06",
|
||
("CEUS_agent/003/U003_report.xlsx", ULTRASOUND),
|
||
{
|
||
"年龄": {"单例报告表": "40岁", "超声汇总表": "41岁"},
|
||
"超声描述": {"单例报告表": "甲" * 31, "超声汇总表": "乙"},
|
||
},
|
||
),
|
||
"单例超声报告表和超声汇总表对不上："
|
||
"年龄（单例表 「40岁」，汇总表 「41岁」）；超声描述（内容不一样）",
|
||
),
|
||
(
|
||
_row("003", "C07", (CLINICAL,), {"Sex": 3, "CA199_Level": None}),
|
||
"临床表这几项的取值不合规：性别（3）、CA19-9 是否异常（空着）",
|
||
),
|
||
(
|
||
_row("003", "C08", (ULTRASOUND,), {"年龄": "四十", "推注开始时间s": -1}),
|
||
"超声表这几项的取值不合规：年龄（「四十」）、推注开始时间（-1）",
|
||
),
|
||
(
|
||
_row("003", "C09", (MR,), {DIAGNOSIS: 5, "MR描述": None}),
|
||
"MR 表这几项的取值不合规：病理诊断（5）、MR描述（空着）",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C10",
|
||
(CLINICAL, ULTRASOUND, MR),
|
||
{"年龄": {"临床表": 28, "超声表": "63岁", "MR 表": "63岁"}},
|
||
),
|
||
"三张表的年龄对不上：临床表 28 岁，超声表 63 岁，MR 表 63 岁",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C10",
|
||
(CLINICAL, ULTRASOUND, MR),
|
||
{"性别": {"超声表": "男", "MR 表": "男", "临床表（解码后）": "女"}},
|
||
),
|
||
"三张表的性别对不上：超声表 男，MR 表 男，临床表 女",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C10",
|
||
(CLINICAL, ULTRASOUND, MR),
|
||
{"性别": {"超声表": "女", "MR 表": "男"}},
|
||
),
|
||
"超声表和 MR 表的性别对不上：超声表 女，MR 表 男"
|
||
"（临床表的性别是数字编码，还没有编码对照表，这次没有拿来比）",
|
||
),
|
||
(
|
||
_row("003", "C08", (ULTRASOUND,), {"年龄": "datetime.datetime(2026, 1, 1, 0, 0)"}),
|
||
"超声表这几项的取值不合规：年龄（这一格是日期或时间格式）",
|
||
),
|
||
(
|
||
_row("003", "C07", (CLINICAL,), {"Symptom_Fever": True, "Hx_Alcohol": False}),
|
||
"临床表这几项的取值不合规：发热（这一格填的是 TRUE）、饮酒史（这一格填的是 FALSE）",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C06",
|
||
("CEUS_agent/003/U003_report.xlsx", ULTRASOUND),
|
||
{"推注开始时间s": {"单例报告表": "datetime.time(0, 0, 5)", "超声汇总表": 5}},
|
||
),
|
||
"单例超声报告表和超声汇总表对不上："
|
||
"推注开始时间（单例表 这一格是日期或时间格式，汇总表 5）",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C11",
|
||
(ULTRASOUND,),
|
||
{"超声描述": [{"关键词": "超声提示", "句子": "超声提示:囊性灶"}]},
|
||
),
|
||
"超声报告的描述里有像结论段标题的写法「超声提示：」（所在句子：超声提示:囊性灶）",
|
||
),
|
||
(
|
||
_row("003", "C12", (MR,), {"MR描述": ["SCN", "浆液性囊"]}),
|
||
"MR报告的描述里出现了亚型名「SCN」、「浆液性囊」",
|
||
),
|
||
(
|
||
_row("003", "C13", (ULTRASOUND, MR), {"超声描述": ["女"], "MR描述": ["45岁"]}),
|
||
"超声报告的描述里出现了性别或年龄「女」；MR报告的描述里出现了性别或年龄「45岁」",
|
||
),
|
||
(
|
||
_row("003", "C14", ("CEUS_agent/003/U003_key.png",), "不是 PNG"),
|
||
"医生截图不是 PNG 格式",
|
||
),
|
||
(
|
||
_row("003", "C14", ("CEUS_agent/003/U003_key.png",), "带 tRNS 透明色，重编码会丢透明"),
|
||
"医生截图带透明色，另存会丢掉透明信息",
|
||
),
|
||
(
|
||
_row(
|
||
"003",
|
||
"C14",
|
||
("20261017-080000-abcdef0/U003/keyframes/D2.png",),
|
||
"位深 16、颜色类型 2 不能逐像素保真重编码",
|
||
),
|
||
"程序从视频里取的 D2 关键帧的颜色格式特殊，程序没法一点不差地另存",
|
||
),
|
||
],
|
||
)
|
||
def test_skip_reasons(row: SkippedRow, expected: str) -> None:
|
||
assert skip_problem(row, {}) == ("资料核对", expected, expected + NOT_IN)
|
||
|
||
|
||
def test_c15_merges_with_failed_csv() -> None:
|
||
failed = _row(
|
||
"002", "C15", ("RUN/failed.csv",), "第 2 步失败：左右两半都没有找到绿框，选不出提示帧"
|
||
)
|
||
reason = "视频里找不到画好的绿框，程序没法处理这段视频"
|
||
steps = {"002": (2, "左右两半都没有找到绿框，选不出提示帧")}
|
||
assert skip_problem(failed, steps) == ("造影视频处理", reason, reason + NOT_IN)
|
||
with pytest.raises(ValueError, match="对不上"):
|
||
skip_problem(failed, {"002": (1, "两半配准不一致：")})
|
||
missing = _row("003", "C15", ("RUN/U003",), "CEUS 运行目录里没有这一例的结果目录")
|
||
assert skip_problem(missing, {}) == (
|
||
"造影视频处理",
|
||
"造影视频没有处理结果",
|
||
"造影视频没有处理结果" + NOT_IN,
|
||
)
|
||
|
||
|
||
@pytest.mark.parametrize(
|
||
"row",
|
||
[
|
||
_row("003", "C99", (), {}),
|
||
_row("003", "C07", (CLINICAL,), {"NewColumn": 1}),
|
||
_row("003", "C14", ("CEUS_agent/003/U003_key.png",), "编造的原因"),
|
||
_row("003", "C15", ("RUN/U003",), "编造的原因"),
|
||
_row("003", "C10", (CLINICAL,), {"身高": {"临床表": 1}}),
|
||
_row("003", "C03", ("x/y.xlsx",), {"缺在": ["x/y.xlsx"]}),
|
||
],
|
||
)
|
||
def test_unknown_skip_reason_stops(row: SkippedRow) -> None:
|
||
with pytest.raises(UnknownReasonError):
|
||
skip_problem(row, {})
|
||
|
||
|
||
def test_case_labels() -> None:
|
||
assert case_label("007") == "P007"
|
||
assert case_label(f"{CLINICAL} 第 5 行") == "临床表第 5 行"
|
||
assert case_label("病例目录 1000") == "病例文件夹「1000」"
|
||
|
||
|
||
def test_program_words_stop() -> None:
|
||
for text in ("C15 失败", "roi_pan", "a_b", "{", "[x]", "band", "20261017-080000"):
|
||
assert FORBIDDEN.search(text) is not None
|
||
with pytest.raises(ValueError, match="程序用语"):
|
||
check_plain(text)
|
||
assert check_plain("视频里找不到画好的绿框") == "视频里找不到画好的绿框"
|