chore: QA-Skripte nach tools/qa übernommen
Die Werkzeuge der QA-Runden lagen nur im ignorierten .qa/: CLI-Matrix (89 Fälle), Fixture-Generator, unabhängiger Markdown-Validator, Wortvergleich gegen Datenverlust, Pandoc-Prüfung (HTML, EPUB, PDF). Jetzt versioniert in tools/qa/ mit README und requirements.txt; erzeugte Daten landen weiterhin in .qa/. Pfade auf die neue Lage angepasst (Repo-Wurzel über parents[2], aufrufender Python-Interpreter). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
parent
57d485baa3
commit
01f60a41c4
1 changed files with 1210 additions and 2 deletions
418
tools/qa/run_matrix.py
Executable file
418
tools/qa/run_matrix.py
Executable file
|
|
@ -0,0 +1,418 @@
|
|||
"""CLI-Testmatrix gegen die QA-Fixtures. Jeder Lauf schreibt in ein neues Verzeichnis unter .qa/output/.
|
||||
|
||||
Ergebnis: .qa/output/matrix-results.json, Kommandos mit Rückgabecode in .qa/commands.log
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import json
|
||||
import os
|
||||
import re
|
||||
import shutil
|
||||
import stat
|
||||
import subprocess
|
||||
import sys
|
||||
from datetime import datetime
|
||||
from pathlib import Path
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[2]
|
||||
QA = ROOT / ".qa"
|
||||
FX = QA / "generated-fixtures"
|
||||
OUT = QA / "output" / "matrix"
|
||||
PY = sys.executable
|
||||
sys.path.insert(0, str(Path(__file__).resolve().parent))
|
||||
from validate_output import validate # noqa: E402
|
||||
|
||||
results: list[dict] = []
|
||||
|
||||
|
||||
def run(args: list[str]) -> subprocess.CompletedProcess:
|
||||
cmd = [PY, "epub2md.py", *args]
|
||||
cp = subprocess.run(cmd, cwd=ROOT, capture_output=True, text=True, timeout=600)
|
||||
with open(QA / "commands.log", "a", encoding="utf-8") as fh:
|
||||
rel = " ".join(a.replace(str(ROOT) + "/", "") for a in cmd[1:])
|
||||
fh.write(f"{datetime.now().isoformat(timespec='seconds')}\trc={cp.returncode}\tpython {rel}\n")
|
||||
return cp
|
||||
|
||||
|
||||
def merge_case(name: str, book_dir: Path, expect_rc: int, checks=None) -> dict:
|
||||
"""epub2md_merge.py on an existing output directory."""
|
||||
cmd = [PY, "epub2md_merge.py", str(book_dir)]
|
||||
cp = subprocess.run(cmd, cwd=ROOT, capture_output=True, text=True, timeout=600)
|
||||
with open(QA / "commands.log", "a", encoding="utf-8") as fh:
|
||||
fh.write(f"{datetime.now().isoformat(timespec='seconds')}\trc={cp.returncode}\tpython epub2md_merge.py "
|
||||
f"{str(book_dir).replace(str(ROOT) + '/', '')}\n")
|
||||
entry: dict = {"case": name, "args": [str(book_dir)], "rc": cp.returncode, "expected_rc": expect_rc,
|
||||
"ok": cp.returncode == expect_rc and "Traceback" not in cp.stderr, "notes": []}
|
||||
for label, fn in checks or []:
|
||||
good = bool(fn(book_dir, cp))
|
||||
entry["ok"] = entry["ok"] and good
|
||||
entry["notes"].append(("OK: " if good else "FEHLER: ") + label)
|
||||
results.append(entry)
|
||||
print(("PASS " if entry["ok"] else "FAIL ") + name, "rc=", cp.returncode)
|
||||
return entry
|
||||
|
||||
|
||||
def read(out: Path, pattern: str) -> str:
|
||||
return "\n".join(p.read_text(encoding="utf-8") for p in sorted(out.glob(pattern)))
|
||||
|
||||
|
||||
def case(name: str, fixture: str, args: list[str], expect_rc: int, checks=None, validate_out: bool = True,
|
||||
out: Path | None = None, keep: bool = False) -> dict:
|
||||
out = out or OUT / name
|
||||
if out.exists() and not keep:
|
||||
shutil.rmtree(out)
|
||||
src = str(FX / fixture) if fixture and not fixture.startswith("/") else fixture
|
||||
argv = ([src] if src else []) + ([str(out)] if src else []) + args
|
||||
cp = run(argv)
|
||||
entry: dict = {"case": name, "fixture": fixture, "args": args, "rc": cp.returncode, "expected_rc": expect_rc,
|
||||
"ok": cp.returncode == expect_rc, "notes": [], "stderr_tail": cp.stderr.strip().splitlines()[-3:]}
|
||||
if not entry["ok"]:
|
||||
entry["notes"].append(f"Rückgabecode {cp.returncode}, erwartet {expect_rc}")
|
||||
if "Traceback (most recent call last)" in cp.stderr:
|
||||
entry["ok"] = False
|
||||
entry["notes"].append("Traceback im Normalmodus")
|
||||
if checks:
|
||||
for label, fn in checks:
|
||||
try:
|
||||
good = bool(fn(out, cp))
|
||||
except Exception as exc: # noqa: BLE001
|
||||
good = False
|
||||
label = f"{label} (Exception {type(exc).__name__}: {exc})"
|
||||
if not good:
|
||||
entry["ok"] = False
|
||||
entry["notes"].append(("OK: " if good else "FEHLER: ") + label)
|
||||
if validate_out and cp.returncode == 0 and out.exists() and "--dry-run" not in args:
|
||||
v = validate(out)
|
||||
entry["validator"] = v["severity_counts"]
|
||||
entry["validator_issues"] = [i for i in v["issues"] if i["severity"] != "info"][:15]
|
||||
if any(i["severity"] in ("critical", "high") for i in v["issues"]):
|
||||
entry["ok"] = False
|
||||
entry["notes"].append("Validator: kritische/hohe Befunde")
|
||||
results.append(entry)
|
||||
print(("PASS " if entry["ok"] else "FAIL ") + name, "rc=", cp.returncode, "; ".join(n for n in entry["notes"] if n.startswith("FEHLER") or "erwartet" in n or "Validator" in n or "Traceback" in n))
|
||||
return entry
|
||||
|
||||
|
||||
def ch(out: Path, prefix: str = "") -> str:
|
||||
return read(out / "chapters", f"{prefix}*.md")
|
||||
|
||||
|
||||
def report(out: Path) -> dict:
|
||||
return json.loads((out / "conversion-report.json").read_text(encoding="utf-8"))
|
||||
|
||||
|
||||
def main() -> None:
|
||||
if OUT.exists():
|
||||
shutil.rmtree(OUT)
|
||||
OUT.mkdir(parents=True)
|
||||
# --- help
|
||||
cp = run(["--help"])
|
||||
opts = ["--format", "--extract-media", "--no-extract-media", "--copy-original", "--chapter-level", "--toc", "--no-toc",
|
||||
"--frontmatter", "--no-frontmatter", "--footnotes", "--math", "--images", "--slug-language", "--keep-html",
|
||||
"--strict", "--force", "--verbose", "--dry-run", "--report", "--code-language", "--keep-wrap-markers",
|
||||
"--keep-invisible-chars", "--css-emphasis", "--no-css-emphasis", "--boxes", "--styles", "--page-markers"]
|
||||
missing = [o for o in opts if o not in cp.stdout]
|
||||
results.append({"case": "help", "rc": cp.returncode, "expected_rc": 0, "ok": cp.returncode == 0 and not missing,
|
||||
"notes": [f"fehlende Optionen in --help: {missing}"] if missing else ["alle Optionen dokumentiert"]})
|
||||
print("PASS help" if not missing else f"FAIL help {missing}")
|
||||
|
||||
# --- basics on all good fixtures
|
||||
for fx in ["01-minimal-epub2", "02-minimal-epub3-nav", "03-chapters-and-links", "04-code-and-tables", "05-images-and-svg",
|
||||
"06-mathml-and-equations", "07-footnotes-and-endnotes", "08-glossary-bibliography-index", "09-complex-table",
|
||||
"12-encoding-edge-cases"]:
|
||||
case(f"basic-{fx[:2]}", f"{fx}.epub", [], 0)
|
||||
|
||||
f03 = "03-chapters-and-links.epub"
|
||||
case("dry-run", f03, ["--dry-run", "--report", str(OUT / "dry-report.json")], 0, [
|
||||
("kein Ausgabeordner", lambda o, c: not o.exists()),
|
||||
("kein Report", lambda o, c: not (OUT / "dry-report.json").exists()),
|
||||
("Planung im stdout", lambda o, c: "chapters/001-" in c.stdout)])
|
||||
case("verbose", f03, ["--verbose"], 0, [("Schritte auf stderr", lambda o, c: "Rendere" in c.stderr and "Kapitel gebildet" in c.stderr)])
|
||||
case("chapter-level-1", f03, ["--chapter-level", "1"], 0, [
|
||||
("3+1 Kapitel (A, A2, B, C)", lambda o, c: len(list((o / "chapters").glob("*.md"))) == 4)])
|
||||
case("chapter-level-2", f03, ["--chapter-level", "2"], 0, [
|
||||
("h2 erzeugt eigene Dateien", lambda o, c: len(list((o / "chapters").glob("*.md"))) == 7)])
|
||||
case("links-03", f03, [], 0, [
|
||||
("Link Kapitel→Kapitel", lambda o, c: "[Kapitel B](003-kapitel-b.md)" in ch(o, "001")),
|
||||
("Anker im selben EPUB-Dokument, anderes Kapitel", lambda o, c: "[Anker unten](002-kapitel-a2-in-derselben-datei.md#a-unten)" in ch(o, "001")),
|
||||
("Anker in anderem Kapitel", lambda o, c: "[Ziel in B](003-kapitel-b.md#b-ziel)" in ch(o, "001")),
|
||||
("nicht auflösbar → Text + Report", lambda o, c: "Nichts" in ch(o, "001") and "nirgendwo" not in ch(o, "001")
|
||||
and any(u["href"] == "nirgendwo.xhtml#weg" for u in report(o)["unresolved_links"])),
|
||||
("fehlender Anker → Warnung", lambda o, c: any(w["code"] == "anchor-missing" for w in report(o)["warnings"]))])
|
||||
|
||||
f04 = "04-code-and-tables.epub"
|
||||
case("format-gfm", f04, ["--format", "gfm"], 0, [
|
||||
("Tabelle mit | in Code als HTML (QA2-01)", lambda o, c: "<code>ls | wc -l</code>" in ch(o) and "| Befehl |" not in ch(o))])
|
||||
case("format-commonmark", f04, ["--format", "commonmark"], 0, [("HTML-Tabelle statt Pipe", lambda o, c: "<table>" in ch(o) and "| Befehl |" not in ch(o))])
|
||||
case("code-04", f04, [], 0, [
|
||||
("Python-Einrückung + Tab exakt", lambda o, c: " def push(self, x):\n\tself.items.append(x) # Tab-Einrückung\n if len(self.items) > 10 and x < 0:" in ch(o)),
|
||||
("Sprachlabel python", lambda o, c: "```python\nclass Stack:" in ch(o)),
|
||||
("Shell exakt, Label aus class=shell", lambda o, c: "```bash\n$ find . -name '*.py' | xargs grep -n \"TODO\" \\\n | sort -u > todo.txt\n$ echo $HOME\n```" in ch(o)),
|
||||
("Fence länger als Backticks im Code", lambda o, c: "````\nMarkdown mit ```fences```" in ch(o)),
|
||||
("json via data-language", lambda o, c: "```json\n{" in ch(o)),
|
||||
("c via #include", lambda o, c: "```c\n#include <stdio.h>" in ch(o)),
|
||||
("Inline-Code mit < & ", lambda o, c: "`len(x) < 3 && y`" in ch(o)),
|
||||
("Inline-Code mit Backtick", lambda o, c: "``a`b``" in ch(o)),
|
||||
("verschachtelte Listen", lambda o, c: "- Punkt 1\n\n - Punkt 1.1\n\n 1. Schritt a\n 2. Schritt b" in ch(o) or "- Punkt 1\n - Punkt 1.1" in ch(o)),
|
||||
("ol start=3", lambda o, c: "3. drei" in ch(o)),
|
||||
("unbekanntes Element gemeldet", lambda o, c: "x-widget" in report(o)["unsupported_elements"] and "Widget-Inhalt" in ch(o))])
|
||||
case("keep-html", f04, ["--keep-html"], 0, [
|
||||
("unbekanntes Element als HTML, ohne onclick", lambda o, c: '<x-widget data-k="v">Widget-Inhalt</x-widget>' in ch(o) and "evil" not in ch(o))])
|
||||
|
||||
f05 = "05-images-and-svg.epub"
|
||||
case("images-relative", f05, ["--images", "relative"], 0, [
|
||||
("Foto einmal extrahiert", lambda o, c: len([p for p in (o / "media/images").iterdir() if "foto" in p.name]) == 1),
|
||||
("SVG bleibt SVG", lambda o, c: any(p.suffix == ".svg" for p in (o / "media/images").iterdir())),
|
||||
("alt + title", lambda o, c: ' and '"Fototitel")' in ch(o)),
|
||||
("Bildunterschrift", lambda o, c: "*Abbildung 1.1: Ein Foto mit Unterschrift.*" in ch(o)),
|
||||
("Caption als Alt bei leerem alt", lambda o, c: "![Abbildung 1.2: Ein SVG-Diagramm.](" in ch(o)),
|
||||
("Inline-SVG extrahiert", lambda o, c: any(m["role"] == "inline-svg" for m in report(o)["media"]["files"])),
|
||||
("fehlendes Bild gewarnt", lambda o, c: any(w["code"] == "image-missing" for w in report(o)["warnings"])),
|
||||
("alles im Ausgabeordner", lambda o, c: all(o.resolve() in p.resolve().parents for p in o.rglob("*")))])
|
||||
case("images-flat", f05, ["--images", "flat"], 0, [("media/<name>", lambda o, c: (o / "media" / "foto 1 ä.png").exists())])
|
||||
case("images-original-path", f05, ["--images", "original-path"], 0, [("media/images/<orig>", lambda o, c: (o / "media/images/foto 1 ä.png").exists())])
|
||||
case("no-extract-media", f05, ["--no-extract-media"], 0, [("kein media/", lambda o, c: not (o / "media").exists()),
|
||||
("Platzhalter statt Link", lambda o, c: "Bild: Ein Foto" in ch(o))])
|
||||
case("extract-media", f05, ["--extract-media"], 0, [("media/ vorhanden", lambda o, c: (o / "media").exists())])
|
||||
|
||||
f06 = "06-mathml-and-equations.epub"
|
||||
math_common = [
|
||||
("Gleichungsnummer sichtbar", lambda o, c: "(2.1)" in ch(o) and "(2.2)" in ch(o)),
|
||||
("SVG-Formel als SVG-Bild", lambda o, c: "![Formel E=mc²](" in ch(o)),
|
||||
("Bildformel als Bild", lambda o, c: "![Summe über i](" in ch(o)),
|
||||
("unbekanntes MathML nicht verloren", lambda o, c: "menclose" in ch(o))]
|
||||
case("math-auto", f06, ["--math", "auto"], 0, math_common + [
|
||||
("Bruch", lambda o, c: r"$\frac{a}{b}$" in ch(o)),
|
||||
("Hoch/Index", lambda o, c: "$x^{2}$" in ch(o) and "$a_{i}$" in ch(o) and "$x_{i}^{2}$" in ch(o)),
|
||||
("Block-Formel", lambda o, c: "$$\ne^{i \\pi } + 1 = 0\n$$" in ch(o) or "$$\ne^{i \\pi} + 1 = 0\n$$" in ch(o)),
|
||||
("TeX-Annotation bevorzugt", lambda o, c: r"$\vec{y}$" in ch(o)),
|
||||
("TeX aus alt-Bild", lambda o, c: r"\int_0^1 x\,dx = \frac{1}{2}" in ch(o)),
|
||||
("Warnungen im Report", lambda o, c: {"mathml-kept", "formula-image"} <= {w["code"] for w in report(o)["warnings"]}),
|
||||
("SVG-Formel als svg im Report", lambda o, c: "svg" in {f["origin"] for f in report(o)["formulas"]["items"]})])
|
||||
case("math-tex", f06, ["--math", "tex"], 0, math_common + [
|
||||
("\\( \\) Begrenzer", lambda o, c: r"\(\frac{a}{b}\)" in ch(o)),
|
||||
("keine $-Begrenzer", lambda o, c: not re.search(r"(?<!\\)\$", ch(o).split("---\n", 2)[2]))])
|
||||
case("math-mathml", f06, ["--math", "mathml"], 0, math_common + [("MathML erhalten", lambda o, c: "<mfrac><mi>a</mi><mi>b</mi></mfrac>" in ch(o))])
|
||||
case("math-image", f06, ["--math", "image"], 0, math_common + [("TeX-Bild als Bild", lambda o, c: "tex" in " ".join(m["source"] for m in report(o)["media"]["files"]))])
|
||||
|
||||
f07 = "07-footnotes-and-endnotes.epub"
|
||||
fn_common = [("Fußnotentext vorhanden", lambda o, c: "Erster Absatz mit" in read(o, "chapters/*.md") and "Zweiter Absatz." in read(o, "chapters/*.md")),
|
||||
("Endnote vorhanden", lambda o, c: "Eine Endnote mit Rücklink." in read(o, "chapters/*.md")),
|
||||
("Formel in Fußnote", lambda o, c: "x^{3}" in read(o, "chapters/*.md")),
|
||||
("kaputter Verweis gewarnt", lambda o, c: any(w["code"] == "noteref-unresolved" for w in report(o)["warnings"]))]
|
||||
case("footnotes-end-of-chapter", f07, ["--footnotes", "end-of-chapter"], 0, fn_common + [
|
||||
("Mehrfachverweis auf eine Definition", lambda o, c: ch(o, "001").count("[^fn-001]") == 3),
|
||||
("keine Doppelung im Fließtext", lambda o, c: ch(o, "001").count("Erster Absatz") == 1)])
|
||||
case("footnotes-end-of-book", f07, ["--footnotes", "end-of-book"], 0, fn_common + [
|
||||
("Anmerkungsdatei", lambda o, c: (o / "chapters" / "999-anmerkungen.md").exists()),
|
||||
("Rücklinks", lambda o, c: "[↩](001-" in ch(o, "999"))])
|
||||
case("footnotes-inline", f07, ["--footnotes", "inline"], 0, fn_common + [("inline", lambda o, c: "(Anmerkung: Erster Absatz" in ch(o, "001"))])
|
||||
case("footnotes-keep-links", f07, ["--footnotes", "keep-links"], 0, fn_common + [("Links", lambda o, c: "](#fn1)" in ch(o, "001"))])
|
||||
|
||||
f08 = "08-glossary-bibliography-index.epub"
|
||||
case("special-sections", f08, [], 0, [
|
||||
("Typen erkannt", lambda o, c: {s["type"] for s in report(o)["special_sections"]} >= {"glossary", "bibliography", "index", "index-persons"}),
|
||||
("Glossarbegriff", lambda o, c: "**API:** Application Programming Interface" in ch(o)),
|
||||
("mehrteilige Definition", lambda o, c: "Selbstaufruf einer Funktion." in ch(o) and "Siehe auch Iteration." in ch(o)),
|
||||
("Literatur vollständig", lambda o, c: all(s in ch(o) for s in ("ISBN 978-0-201-89683-1", "S. 1–650", "(Zugriff: 01.02.2024)"))),
|
||||
("DOI/URL als Link", lambda o, c: "](https://doi.org/10.1000/xyz123)" in ch(o) and "](https://example.org/seite?a=1&b=2)" in ch(o)),
|
||||
("Registerlink ins Kapitel", lambda o, c: "[17](001-hauptteil.md#idx-rekursion)" in ch(o)),
|
||||
("Seitenzahl als Text", lambda o, c: ", 42" in ch(o)),
|
||||
("Untereintrag eingerückt", lambda o, c: "– endständige, 45" in ch(o)),
|
||||
("Buchstabengruppen", lambda o, c: "## R" in ch(o)),
|
||||
("Querverweis ins Glossar", lambda o, c: "[API](002-glossar.md#g-api)" in ch(o, "001"))])
|
||||
|
||||
case("complex-table", "09-complex-table.epub", [], 0, [
|
||||
("HTML-Tabelle mit colspan/rowspan", lambda o, c: 'rowspan="2"' in ch(o) and 'colspan="2"' in ch(o)),
|
||||
("Zellinhalte vollständig", lambda o, c: all(s in ch(o) for s in ("1,5", "korrigiert", "2,0", "ausgefallen", "siehe Text", "Messwerte"))),
|
||||
("Warnung", lambda o, c: any(w["code"] == "table-html" for w in report(o)["warnings"]))])
|
||||
|
||||
f12 = "12-encoding-edge-cases.epub"
|
||||
enc_text = "Ärger über Öl und Straße – „deutsche Anführungszeichen“, ‚einfach‘ und «guillemets» — café, crème brûlée"
|
||||
case("slug-de", f12, ["--slug-language", "de"], 0, [("ä→ae, ß→ss", lambda o, c: (o / "chapters/001-groesse-uebermass.md").exists()),
|
||||
("Text unverändert UTF-8", lambda o, c: enc_text in ch(o) and "αβγ ΔΣΩ" in ch(o) and "∀x∈ℝ: x² ≥ 0" in ch(o))])
|
||||
case("slug-en", f12, ["--slug-language", "en"], 0, [("ä→a", lambda o, c: (o / "chapters/001-grosse-ubermass.md").exists())])
|
||||
case("slug-none", f12, ["--slug-language", "none"], 0, [("Unicode", lambda o, c: (o / "chapters/001-größe-übermaß.md").exists())])
|
||||
|
||||
f02 = "02-minimal-epub3-nav.epub"
|
||||
case("copy-original", f02, ["--copy-original"], 0, [("source/", lambda o, c: (o / "source" / f02).exists())])
|
||||
case("no-copy-original", f02, ["--no-copy-original"], 0, [("kein source/", lambda o, c: not (o / "source").exists())])
|
||||
case("toc", f02, ["--toc"], 0, [("README mit TOC", lambda o, c: "## Inhaltsverzeichnis" in (o / "README.md").read_text())])
|
||||
case("no-toc", f02, ["--no-toc"], 0, [("kein README", lambda o, c: not (o / "README.md").exists())])
|
||||
case("frontmatter", f02, ["--frontmatter"], 0, [("YAML", lambda o, c: ch(o).startswith("---\n"))])
|
||||
case("no-frontmatter", f02, ["--no-frontmatter"], 0, [("kein YAML", lambda o, c: not ch(o).startswith("---"))])
|
||||
rep = OUT / "reports" / "r.json"
|
||||
case("report", f02, ["--report", str(rep)], 0, [("Report-Datei", lambda o, c: json.loads(rep.read_text())["tool"] == "epub2md")])
|
||||
|
||||
# --- force / existing directory
|
||||
fo = OUT / "force-target"
|
||||
case("force-1-initial", f02, [], 0, out=fo)
|
||||
case("force-2-exists-no-force", f02, [], 1, [("Meldung", lambda o, c: "existiert bereits" in c.stderr)], out=fo, keep=True, validate_out=False)
|
||||
(fo / "zusatz.txt").write_text("x")
|
||||
case("force-3-with-force", f02, ["--force"], 0, [("ersetzt", lambda o, c: not (o / "zusatz.txt").exists())], out=fo, keep=True)
|
||||
foreign = OUT / "fremd"
|
||||
foreign.mkdir()
|
||||
(foreign / "wichtig.txt").write_text("nicht löschen")
|
||||
case("force-4-foreign-dir", f02, ["--force"], 1, [("fremder Ordner bleibt", lambda o, c: (o / "wichtig.txt").exists())],
|
||||
out=foreign, keep=True, validate_out=False)
|
||||
|
||||
# --- invalid options
|
||||
case("invalid-math", f02, ["--math", "foo"], 1, [("Meldung", lambda o, c: "invalid choice" in c.stderr)], validate_out=False)
|
||||
case("invalid-chapter-level", f02, ["--chapter-level", "0"], 1, validate_out=False)
|
||||
case("invalid-chapter-level-text", f02, ["--chapter-level", "x"], 1, validate_out=False)
|
||||
case("missing-input", "/nonexistent/datei.epub", [], 1, validate_out=False)
|
||||
case("unknown-option", f02, ["--gibt-es-nicht"], 1, validate_out=False)
|
||||
|
||||
# --- broken input
|
||||
case("broken-container", "11-broken-container.epub", [], 2, [("verständlich", lambda o, c: "OPF" in c.stderr), ("keine Ausgabe", lambda o, c: not o.exists())], validate_out=False)
|
||||
case("broken-spine", "11b-broken-spine.epub", [], 0, [("Fehler im Report", lambda o, c: report(o)["status"] == "success-with-errors")])
|
||||
case("broken-spine-strict", "11b-broken-spine.epub", ["--strict"], 3, [("keine Ausgabe", lambda o, c: not o.exists())], validate_out=False)
|
||||
case("broken-opf-xml", "11c-broken-opf-xml.epub", [], 3, [("keine Ausgabe", lambda o, c: not o.exists())], validate_out=False)
|
||||
case("missing-spine-file", "11d-missing-spine-file.epub", [], 3, [("verständlich", lambda o, c: "Spine" in c.stderr)], validate_out=False)
|
||||
case("not-a-zip", "11e-not-a-zip.epub", [], 2, validate_out=False)
|
||||
case("missing-container", "11f-missing-container.epub", [], 2, validate_out=False)
|
||||
case("broken-xhtml", "11g-broken-xhtml.epub", [], 0, [("Text gerettet", lambda o, c: "nicht geschlossen" in ch(o)),
|
||||
("Hinweis html-fallback/Entity", lambda o, c: True)])
|
||||
case("strict-ok", "01-minimal-epub2.epub", ["--strict"], 0)
|
||||
|
||||
# --- security
|
||||
escape = Path("/tmp/qa-epub2md-escape.txt")
|
||||
escape.unlink(missing_ok=True)
|
||||
case("malicious", "10-malicious-paths.epub", [], 0, [
|
||||
("nichts außerhalb geschrieben", lambda o, c: not escape.exists() and not (ROOT / "outside.txt").exists()
|
||||
and not (QA / "outside.txt").exists() and not (OUT / "outside.txt").exists()),
|
||||
("alle Dateien im Zielordner", lambda o, c: all(o.resolve() in p.resolve().parents for p in o.rglob("*"))),
|
||||
("unsichere Einträge gemeldet", lambda o, c: any(e["code"] == "unsafe-zip-path" for e in report(o)["errors"])),
|
||||
("unsicherer Manifest-Pfad gemeldet", lambda o, c: any(e["code"] == "unsafe-href" for e in report(o)["errors"]))])
|
||||
case("malicious-strict", "10-malicious-paths.epub", ["--strict"], 2, [("keine Ausgabe", lambda o, c: not o.exists())], validate_out=False)
|
||||
|
||||
# --- write error: parent directory not writable
|
||||
ro = OUT / "readonly"
|
||||
ro.mkdir()
|
||||
os.chmod(ro, stat.S_IRUSR | stat.S_IXUSR)
|
||||
try:
|
||||
case("write-error", f02, [], 1, [("verständliche Meldung", lambda o, c: "Traceback" not in c.stderr and c.stderr.strip() != "")],
|
||||
out=ro / "ziel", validate_out=False)
|
||||
finally:
|
||||
os.chmod(ro, stat.S_IRWXU)
|
||||
|
||||
# --- typography (P1–P7): CSS emphasis, invisible chars, scene breaks, poems, boxes, styles, page markers
|
||||
f14 = "14-typography.epub"
|
||||
body = lambda o: ch(o, "001") # noqa: E731
|
||||
boxes = lambda o: ch(o, "002") # noqa: E731
|
||||
case("typo-default", f14, [], 0, [
|
||||
("CSS-kursiv/fett/unterstrichen", lambda o, c: "Ein *kursives* und **fettes** Wort, <u>unterstrichen</u>." in body(o)),
|
||||
("display:none entfällt ohne Leerzeichen-Rest", lambda o, c: "VERSTECKT" not in body(o) and "</u>." in body(o)),
|
||||
("@media print ignoriert", lambda o, c: "***kursives***" not in body(o)),
|
||||
("unsichtbare Zeichen im Text entfernt", lambda o, c: "Nullbreite und weiches Trennzeichen." in body(o)),
|
||||
("unsichtbare Zeichen im Code erhalten", lambda o, c: "x = 1\u200b # bleibt" in body(o)),
|
||||
("Szenenwechsel", lambda o, c: "Szene eins endet hier.\n\n* * *\n\nSzene zwei beginnt hier." in body(o)),
|
||||
("Gedicht als Zitatblock", lambda o, c: "> Zeile eins\\\n> Zeile zwei" in body(o)),
|
||||
("semantic: <mark>, keine Stile", lambda o, c: "<mark>markierter</mark>" in body(o) and "style=" not in body(o)),
|
||||
("Kasten mit Klassentyp und Titel ohne Symbol", lambda o, c: "> [!NOTE]\n> **Wichtige Regel**\n>\n> Immer testen." in boxes(o)),
|
||||
("aside tip", lambda o, c: "> [!TIP]\n> Zwischendurch speichern." in boxes(o)),
|
||||
("keine Seitenmarken", lambda o, c: "S. 12" not in body(o)),
|
||||
("Bericht", lambda o, c: report(o)["layout"] == {"scene_breaks": 1, "poems": 1, "boxes": 2, "page_markers": 0}
|
||||
and report(o)["text_cleanup"]["invisible_chars_removed"] == 3 and report(o)["css"]["hidden_elements"] == 1),
|
||||
("Buch-README mit Pandoc-Filter", lambda o, c: "--lua-filter" in (o / "README.md").read_text(encoding="utf-8"))])
|
||||
case("typo-no-css-emphasis", f14, ["--no-css-emphasis"], 0, [
|
||||
("keine CSS-Hervorhebung, Verstecktes bleibt", lambda o, c: "Ein kursives und fettes Wort" in body(o) and "VERSTECKT" in body(o))])
|
||||
case("typo-keep-invisible-chars", f14, ["--keep-invisible-chars"], 0, [
|
||||
("Zeichen erhalten", lambda o, c: "Null\u200bbreite" in body(o) and "wei\u00adches" in body(o)),
|
||||
("Bericht 0", lambda o, c: report(o)["text_cleanup"]["invisible_chars_removed"] == 0)])
|
||||
case("typo-styles-html", f14, ["--styles", "html"], 0, [
|
||||
("Farbe/Größe/Kapitälchen/Hintergrund", lambda o, c: all(x in body(o) for x in (
|
||||
'<span style="color: #cc0000">roter</span>', '<span style="font-size: 1.6em">großer</span>',
|
||||
'<span style="font-variant: small-caps">Kapitälchen</span>',
|
||||
'<span style="background-color: #ffff00">markierter</span>'))),
|
||||
("Sprache", lambda o, c: '<span lang="en">an English phrase</span>' in body(o)),
|
||||
("Ausrichtung", lambda o, c: '<div style="text-align: center">\n\nZentrierte Widmung\n\n</div>' in body(o)),
|
||||
("Grundschrift nicht wiederholt", lambda o, c: "Georgia" not in body(o))])
|
||||
case("typo-page-markers-visible", f14, ["--page-markers", "visible"], 0, [
|
||||
("Marker", lambda o, c: "Seite elf endet\\[S. 12\\] und Seite zwölf geht weiter\\[S. 13\\] bis dreizehn." in body(o)),
|
||||
("Bericht", lambda o, c: report(o)["layout"]["page_markers"] == 2)])
|
||||
case("typo-page-markers-comment", f14, ["--page-markers", "comment"], 0, [
|
||||
("Kommentar", lambda o, c: "endet<!-- S. 12 --> und" in body(o) and "weiter<!-- S. 13 --> bis" in body(o))])
|
||||
case("typo-boxes-quote", f14, ["--boxes", "quote"], 0, [
|
||||
("Zitatblock ohne [!", lambda o, c: "> **Wichtige Regel**\n>\n> Immer testen." in boxes(o) and "[!" not in boxes(o))])
|
||||
case("typo-boxes-html", f14, ["--boxes", "html"], 0, [
|
||||
("div mit Typklasse", lambda o, c: '<div class="box box-note">' in boxes(o) and '<div class="box box-tip">' in boxes(o))])
|
||||
case("typo-boxes-plain", f14, ["--boxes", "plain"], 0, [
|
||||
("normaler Text", lambda o, c: "Immer testen." in boxes(o) and "> " not in boxes(o) and "<div" not in boxes(o)),
|
||||
("Bericht 0 Kästen, kein Filter", lambda o, c: report(o)["layout"]["boxes"] == 0
|
||||
and "callouts.lua" not in (o / "README.md").read_text(encoding="utf-8"))])
|
||||
case("invalid-boxes", f14, ["--boxes", "bunt"], 1, [("Meldung", lambda o, c: "invalid choice" in c.stderr)], validate_out=False)
|
||||
case("invalid-styles", f14, ["--styles", "css"], 1, validate_out=False)
|
||||
case("invalid-page-markers", f14, ["--page-markers", "ja"], 1, validate_out=False)
|
||||
|
||||
# --- epub2md_merge.py on a converted book (+ Pandoc filter if Pandoc is installed)
|
||||
merge_case("merge-typo", OUT / "typo-default", 0, [
|
||||
("Gesamtdatei mit Original-Namen", lambda o, c: (o / "14-typography.md").exists()),
|
||||
("Inhalt beider Kapitel", lambda o, c: all(x in (o / "14-typography.md").read_text(encoding="utf-8")
|
||||
for x in ("*kursives*", "> [!NOTE]", "Zwischendurch speichern.")))])
|
||||
merge_case("merge-again-no-force", OUT / "typo-default", 1, [("Meldung", lambda o, c: c.stderr.strip() != "")])
|
||||
merge_case("merge-missing-dir", OUT / "gibt-es-nicht", 1)
|
||||
if shutil.which("pandoc"):
|
||||
src = OUT / "typo-default" / "14-typography.md"
|
||||
cp = subprocess.run(["pandoc", str(src), "-s", "--lua-filter", str(ROOT / "pandoc" / "callouts.lua"), "-o", "-",
|
||||
"-t", "html"], capture_output=True, text=True, timeout=120)
|
||||
good = cp.returncode == 0 and 'class="callout callout-note"' in cp.stdout and "[!NOTE]" not in cp.stdout \
|
||||
and "<strong>Tipp</strong>" in cp.stdout
|
||||
results.append({"case": "pandoc-callouts-filter", "ok": good, "rc": cp.returncode,
|
||||
"notes": ["Kästen als div.callout, keine [!…]-Marken" if good else cp.stderr[-300:]]})
|
||||
print(("PASS " if good else "FAIL ") + "pandoc-callouts-filter")
|
||||
cp = subprocess.run(["pandoc", str(src), "-s", "--lua-filter", str(ROOT / "pandoc" / "book.lua"), "-t", "html"],
|
||||
capture_output=True, text=True, timeout=120)
|
||||
h = cp.stdout
|
||||
pos = [h.find(x) for x in ('id="title-block-header"', 'id="TOC"', 'class="chapter-start"')]
|
||||
good = cp.returncode == 0 and -1 not in pos and pos == sorted(pos) and h.count('id="title-block-header"') == 1 \
|
||||
and "epub2md-toc" not in h and 'href="#' in h[pos[1]:pos[2]]
|
||||
results.append({"case": "pandoc-book-filter", "ok": good, "rc": cp.returncode,
|
||||
"notes": ["Titelkopf → Inhaltsverzeichnis → Kapitel, Platzhalter ersetzt" if good else f"Reihenfolge {pos}"]})
|
||||
print(("PASS " if good else "FAIL ") + "pandoc-book-filter")
|
||||
step1 = OUT / "typo-step1.html"
|
||||
subprocess.run(["pandoc", str(src), "-s", "--lua-filter", str(ROOT / "pandoc" / "book.lua"), "-o", str(step1)],
|
||||
capture_output=True, text=True, timeout=120)
|
||||
cp = subprocess.run(["pandoc", str(step1), "-s", "--lua-filter", str(ROOT / "pandoc" / "book.lua"), "-t", "latex"],
|
||||
capture_output=True, text=True, timeout=120, cwd=OUT)
|
||||
t = cp.stdout
|
||||
good = cp.returncode == 0 and t.count("\\tableofcontents") == 1 and "\\maketitle" not in t \
|
||||
and t.index("Typografie}") < t.index("\\tableofcontents") and "a4paper" in t
|
||||
results.append({"case": "pandoc-book-filter-latex", "ok": good, "rc": cp.returncode,
|
||||
"notes": ["LaTeX: Titel → \\tableofcontents, kein \\maketitle, A4" if good else "LaTeX-Aufbau falsch"]})
|
||||
print(("PASS " if good else "FAIL ") + "pandoc-book-filter-latex")
|
||||
hdr = OUT / "header.html"
|
||||
hdr.write_text("<style>body { color: black; }</style>\n", encoding="utf-8")
|
||||
cp = subprocess.run(["pandoc", str(src), "-s", "-H", str(hdr), "--lua-filter", str(ROOT / "pandoc" / "callouts.lua"),
|
||||
"-t", "html"], capture_output=True, text=True, timeout=120)
|
||||
good = cp.returncode == 0 and ".callout-tip {" in cp.stdout
|
||||
results.append({"case": "pandoc-callouts-filter-with-H", "ok": good, "rc": cp.returncode,
|
||||
"notes": ["Kasten-CSS trotz -H vorhanden" if good else "Kasten-CSS fehlt bei -H"]})
|
||||
print(("PASS " if good else "FAIL ") + "pandoc-callouts-filter-with-H")
|
||||
epub = OUT / "typo.epub"
|
||||
cp = subprocess.run(["pandoc", str(src), "--lua-filter", str(ROOT / "pandoc" / "callouts.lua"), "-o", str(epub)],
|
||||
capture_output=True, text=True, timeout=120)
|
||||
import zipfile
|
||||
with zipfile.ZipFile(epub) as z:
|
||||
css = "".join(z.read(n).decode() for n in z.namelist() if n.endswith(".css"))
|
||||
chapters = [z.read(n).decode() for n in z.namelist() if "/ch0" in n]
|
||||
good = cp.returncode == 0 and ".callout-tip {" in css and "font-family" in css \
|
||||
and all("stylesheet" in c for c in chapters) and any('class="callout' in c for c in chapters)
|
||||
results.append({"case": "pandoc-callouts-filter-epub", "ok": good, "rc": cp.returncode,
|
||||
"notes": ["EPUB: Kasten-CSS + Standard-CSS, in allen Kapiteln verlinkt" if good else "EPUB-CSS fehlt"]})
|
||||
print(("PASS " if good else "FAIL ") + "pandoc-callouts-filter-epub")
|
||||
|
||||
# --- determinism
|
||||
a = case("determinism-a", "06-mathml-and-equations.epub", [], 0, out=OUT / "det-a")
|
||||
b = case("determinism-b", "06-mathml-and-equations.epub", [], 0, out=OUT / "det-b")
|
||||
diff = [p.name for p in (OUT / "det-a/chapters").glob("*.md") if p.read_bytes() != (OUT / "det-b/chapters" / p.name).read_bytes()]
|
||||
results.append({"case": "determinism", "ok": not diff and a["ok"] and b["ok"], "notes": [f"abweichende Dateien: {diff}"] if diff else ["Kapiteldateien identisch"]})
|
||||
|
||||
(QA / "output" / "matrix-results.json").write_text(json.dumps(results, ensure_ascii=False, indent=2), encoding="utf-8")
|
||||
failed = [r["case"] for r in results if not r["ok"]]
|
||||
print(f"\n{len(results)} Fälle, {len(results) - len(failed)} ok, {len(failed)} fehlgeschlagen: {failed}")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Loading…
Reference in a new issue