P5b: self-contained HTML QC report and QC table (simurg qc report)
- pipeline/qc_report: dataset discovery/labels/types (QC-041..043), robust TSV/CSV reading (QC-061), join + unmatched warning (QC-040), privacy options (QC-046), xlsx (4 sheets, conditional fills) / csv table (QC-045), byte-deterministic HTML with gzip+base64 embedded data (QC-044, QC-047) - qc/report: template, CSS and JS app — Control Dashboard, Automated QC (table with sort/filter/search/ paging/column menu/export, metric scatters with threshold lines), Sample Heatmaps (chip/plate, continuous/binary), Trend Analysis (grouped summary, Tukey box plots, counts, sex, PCA), thresholds modal (apply / download / upload YAML), linked selection across plots and table (QC-048..060) - bundled plotly.js 2.35.2 (MIT); openpyxl dependency (MIT) - tests: discovery/labels/errors/determinism/xlsx + headless-browser self-test - known_unknowns K-020: remaining report deviations
This commit is contained in:
kemal committed
1 parent
7904399161
commit
78e107aaa4
13 files changed
+1601
No files matched your search
@@ -0,0 +1,167 @@
|
||||
"""QC report (spec QC-040 … QC-063): dataset discovery, labels, types, table, HTML determinism and error codes."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import datetime as dt
|
||||
import json
|
||||
import shutil
|
||||
import subprocess
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
from typer.testing import CliRunner
|
||||
|
||||
from simurg.cli.main import app
|
||||
from simurg.core.errors import SimurgError
|
||||
from simurg.core.log import Reporter
|
||||
from simurg.pipeline.qc_report import QcReportOptions, assign_labels, resolve_datasets, run_qc_report
|
||||
|
||||
RAW_HDR = "sample_id\tcategory\tbead_address\tcontrol_name\tcolor_label\tgreen_intensity\tred_intensity\n"
|
||||
PROBES = [("Staining", "1", "Biotin (High)", "Green"), ("Staining", "2", "Biotin (Bgnd)", "Blue"),
|
||||
("Non-Polymorphic", "3", "NP (A)", "Red"), ("NEGATIVE", "4", "neg", "Red")] # fmt: skip
|
||||
|
||||
|
||||
def make_dataset(folder: Path, n: int = 4, array_type: str = "genotyping", summary: bool = True) -> Path:
|
||||
folder.mkdir(parents=True, exist_ok=True)
|
||||
ids = [f"2000000000{k}_R0{i + 1}C01" for k in (1,) for i in range(n)]
|
||||
raw = RAW_HDR + "".join(
|
||||
f"{s}\t{c}\t{a}\t{nm}\t{col}\t{1000 * (j + 1) + i}\t{500 + i}\n"
|
||||
for i, s in enumerate(ids)
|
||||
for j, (c, a, nm, col) in enumerate(PROBES)
|
||||
)
|
||||
(folder / "controls_raw.tsv").write_text(raw)
|
||||
qc = "sample_id\tarray_type\timaging_date\tsample_plate\tsample_well\tscanner_id\tscanner_version\tbg_offset\tstaining_green\tnp_red\tna_metrics\n"
|
||||
qc += "".join(
|
||||
f"{s}\t{array_type}\t2026-01-0{i + 1}T10:00:00\tP1\tA0{i + 1}\tN1\t1.0\t3000\t{6 - i}\t{10 + i}\t\n"
|
||||
for i, s in enumerate(ids)
|
||||
)
|
||||
(folder / "controls_qc.tsv").write_text(qc)
|
||||
if summary:
|
||||
cols = ["sample_id", "sample_name", "sample_folder", "status", "fail_reason", "autosomal_call_rate",
|
||||
"call_rate", "logr_dev", "sex_est", "sex_reported", "sex_check", "het_rate", "contamination",
|
||||
"functional_status", "control_status", "control_flag_codes", "flags", "Sample_Group"] # fmt: skip
|
||||
lines = ["\t".join(cols)]
|
||||
for i, s in enumerate(ids):
|
||||
vals = {"sample_id": s, "sample_name": f"S{i}", "sample_folder": "x", "status": "OK", "fail_reason": "",
|
||||
"autosomal_call_rate": f"{0.99 - 0.01 * i}", "call_rate": "0.99", "logr_dev": f"{0.1 + 0.05 * i}",
|
||||
"sex_est": "MF"[i % 2], "sex_reported": "U", "sex_check": "unknown", "het_rate": "0.17",
|
||||
"contamination": "0.001", "functional_status": "", "control_status": "",
|
||||
"control_flag_codes": "", "flags": "", "Sample_Group": "g" + str(i % 2)} # fmt: skip
|
||||
lines.append("\t".join(vals[c] for c in cols))
|
||||
(folder / "sample_summary.tsv").write_text("\n".join(lines) + "\n")
|
||||
return folder
|
||||
|
||||
|
||||
def run(tmp_path: Path, data: list[str], **kw: object) -> Path:
|
||||
out = tmp_path / "out"
|
||||
opts = QcReportOptions(data=data, out=out, report_name="r", now=dt.datetime(2026, 1, 2, tzinfo=dt.UTC), **kw) # type: ignore[arg-type]
|
||||
assert run_qc_report(opts, Reporter(command="t", quiet=True)) == 0
|
||||
return out
|
||||
|
||||
|
||||
def test_discovery_forms(tmp_path: Path) -> None: # QC-041
|
||||
a = make_dataset(tmp_path / "p" / "b")
|
||||
b = make_dataset(tmp_path / "p" / "a")
|
||||
assert resolve_datasets([str(tmp_path / "p")]) == [b, a] # sorted by path
|
||||
assert resolve_datasets([f"{a},{b}"]) == [a, b]
|
||||
with pytest.raises(SimurgError) as e:
|
||||
resolve_datasets([str(tmp_path / "nope")])
|
||||
assert e.value.code == "E_QC_NO_DATASETS"
|
||||
(tmp_path / "empty").mkdir()
|
||||
with pytest.raises(SimurgError) as e:
|
||||
resolve_datasets([str(tmp_path / "empty")])
|
||||
assert e.value.code == "E_QC_NO_DATASETS"
|
||||
(b / "controls_raw.tsv").unlink()
|
||||
with pytest.raises(SimurgError) as e:
|
||||
resolve_datasets([f"{a},{b}"])
|
||||
assert e.value.code == "E_QC_DATASET_INCOMPLETE"
|
||||
|
||||
|
||||
def test_labels(tmp_path: Path) -> None: # QC-042
|
||||
rep = Reporter(command="t", quiet=True)
|
||||
f = [tmp_path / "x" / "run", tmp_path / "y" / "run"]
|
||||
assert assign_labels(f, None, rep) == ["run", "run_2"]
|
||||
assert assign_labels(f, ["My Label<1>", "b"], rep) == ["My_Label_1_", "b"]
|
||||
assert rep.warning_counts().get("W_QC_LABEL_SANITIZED") == 1
|
||||
for labels, code in ((["a"], "E_QC_LABEL_COUNT"), (["a b", "a_b"], "E_QC_LABEL_DUPLICATE")):
|
||||
with pytest.raises(SimurgError) as e:
|
||||
assign_labels(f, labels, rep)
|
||||
assert e.value.code == code
|
||||
|
||||
|
||||
def test_report_outputs_and_determinism(tmp_path: Path) -> None: # QC-044, QC-045, QC-047
|
||||
d = make_dataset(tmp_path / "ds")
|
||||
out = run(tmp_path, [str(d)], output_format="both")
|
||||
html = (out / "r.html").read_text()
|
||||
assert "SimurgArray QC Report" in html and "https://" not in html.split("plotly.js v")[0]
|
||||
first = html
|
||||
out2 = run(tmp_path, [str(d)], output_format="both")
|
||||
assert (out2 / "r.html").read_text() == first # byte-deterministic with a fixed timestamp
|
||||
rows = (out / "r_table.csv").read_text().splitlines()
|
||||
head = rows[0].split(",")
|
||||
assert head[:7] == ["dataset", "sample_id", "sample_name", "functional_status", "control_status",
|
||||
"control_flag_codes", "na_metrics"] # fmt: skip
|
||||
assert "neg_red" not in head and "Biotin (High)_green" in head # QC-010 exclusion
|
||||
assert head[-1] == "sample_well" or "Sample_Group" in head
|
||||
from openpyxl import load_workbook
|
||||
|
||||
wb = load_workbook(out / "r_table.xlsx")
|
||||
assert wb.sheetnames == ["Samples", "Thresholds", "Datasets", "About"]
|
||||
ws = wb["Samples"]
|
||||
status = [c.value for c in ws["D"]][1:]
|
||||
assert status == ["PASS", "PASS", "FAIL", "FAIL"] # call rate 0.97 < 0.98; LRR SD 0.25 > 0.20
|
||||
|
||||
|
||||
def test_privacy_and_errors(tmp_path: Path) -> None: # QC-043, QC-046, QC-063
|
||||
d = make_dataset(tmp_path / "ds")
|
||||
out = run(tmp_path, [str(d)], output_format="csv", no_metadata=True)
|
||||
head = (out / "r_table.csv").read_text().splitlines()[0]
|
||||
assert "Sample_Group" not in head and "sample_plate" not in head
|
||||
runner = CliRunner()
|
||||
res = runner.invoke(app, ["qc", "report", "--data", str(d), "--out", str(tmp_path / "o"), "--output-format", "pdf"])
|
||||
assert res.exit_code == 2 and "E_QC_FORMAT" in res.output
|
||||
m = make_dataset(tmp_path / "meth", array_type="methylation")
|
||||
res = runner.invoke(app, ["qc", "report", "--data", f"{d},{m}", "--out", str(tmp_path / "o")])
|
||||
assert res.exit_code == 2 and "E_QC_MIXED_TYPES" in res.output
|
||||
bad = tmp_path / "bad.yaml"
|
||||
bad.write_text("qc: {report: {callRate: 2}}")
|
||||
res = runner.invoke(app, ["qc", "report", "--data", str(d), "--out", str(tmp_path / "o"), "--config", str(bad)])
|
||||
assert res.exit_code == 2 and "E_QC_THRESHOLD_RANGE" in res.output
|
||||
|
||||
|
||||
def test_unmatched_samples_warn(tmp_path: Path) -> None: # QC-040
|
||||
d = make_dataset(tmp_path / "ds")
|
||||
lines = (d / "sample_summary.tsv").read_text().splitlines()
|
||||
(d / "sample_summary.tsv").write_text("\n".join(lines[:-1]) + "\n")
|
||||
rep = Reporter(command="t", quiet=True)
|
||||
run_qc_report(QcReportOptions(data=[str(d)], out=tmp_path / "o", output_format="csv"), rep)
|
||||
assert rep.warning_counts().get("W_QC_UNMATCHED_SAMPLES") == 1
|
||||
|
||||
|
||||
@pytest.mark.slow
|
||||
def test_report_selftest_in_browser(tmp_path: Path) -> None: # QC-051, QC-055 (client-side behaviour)
|
||||
browser = shutil.which("chromium") or shutil.which("chromium-browser") or shutil.which("google-chrome")
|
||||
if not browser:
|
||||
pytest.skip("no Chromium for the headless report test")
|
||||
d = make_dataset(tmp_path / "ds", n=6)
|
||||
out = run(tmp_path, [str(d)], output_format="csv")
|
||||
dom = subprocess.run(
|
||||
[
|
||||
browser,
|
||||
"--headless=new",
|
||||
"--no-sandbox",
|
||||
"--disable-gpu",
|
||||
"--virtual-time-budget=20000",
|
||||
"--dump-dom",
|
||||
f"file://{out / 'r.html'}#selftest",
|
||||
],
|
||||
capture_output=True,
|
||||
text=True,
|
||||
timeout=180,
|
||||
).stdout
|
||||
marker = 'data-selftest="'
|
||||
assert marker in dom, dom[-2000:]
|
||||
res = json.loads(dom.split(marker, 1)[1].split('"', 1)[0].replace(""", '"'))
|
||||
assert res["yamlUnknown"] == "rejected" and res["yaml"] is True
|
||||
assert res["failAfterApply"] == 6 # callRate 0.9985 fails every synthetic sample (max 0.99)
|
||||
assert res["rowsWithSelection"] == 3 and res["linkedPlots"] > 0
|
||||
Reference in new issue
Block a user