feat(series-studio): initial implementation of unified series studio engine (PDF & DOCX)
This commit is contained in:
commit
03d9922066
20 files changed
+7082
No files matched your search
@@ -0,0 +1,176 @@
|
||||
#!/usr/bin/env python3
|
||||
# -*- coding: utf-8 -*-
|
||||
"""
|
||||
test_series_studio.py
|
||||
Suite de 10 tests unitaires et d'intégration obligatoires pour series-studio.
|
||||
Exécutée par le vérificateur : pytest -v -p no:cacheprovider /app/tests
|
||||
"""
|
||||
|
||||
import io
|
||||
import json
|
||||
import os
|
||||
import re
|
||||
import zipfile
|
||||
import fitz
|
||||
import pytest
|
||||
from starlette.testclient import TestClient
|
||||
|
||||
from main import app, JOBS_DIR
|
||||
|
||||
client = TestClient(app)
|
||||
|
||||
EXEMPLE_SERIE06_PATH = os.path.join(os.path.dirname(__file__), "..", "exemples", "serie06.json")
|
||||
if not os.path.exists(EXEMPLE_SERIE06_PATH):
|
||||
EXEMPLE_SERIE06_PATH = "/app/exemples/serie06.json"
|
||||
|
||||
def load_serie06():
|
||||
with open(EXEMPLE_SERIE06_PATH, "r", encoding="utf-8") as f:
|
||||
return json.load(f)
|
||||
|
||||
def test_json_invalide_refuse_avec_chemin():
|
||||
"""Un JSON invalide est rejeté avec indication du chemin exact du champ fautif."""
|
||||
src = load_serie06()
|
||||
mauvais = dict(src)
|
||||
mauvais["cartouche"] = "invalide_string"
|
||||
|
||||
res = client.post("/validate", json={"source": mauvais})
|
||||
assert res.status_code == 200
|
||||
data = res.json()
|
||||
assert data.get("valid") is False
|
||||
assert "errors" in data
|
||||
assert any("cartouche" in err.get("path", "") for err in data["errors"])
|
||||
|
||||
def test_pdf_deux_pages():
|
||||
"""Le PDF compilé par Typst pour la série 06 fait exactement 2 pages calibrées."""
|
||||
src = load_serie06()
|
||||
res = client.post("/render", json={"source": src, "theme": "classique"})
|
||||
assert res.status_code == 200
|
||||
job_id = res.json()["job_id"]
|
||||
|
||||
res_pdf = client.get(f"/jobs/{job_id}/pdf")
|
||||
assert res_pdf.status_code == 200
|
||||
doc = fitz.open(stream=res_pdf.content, filetype="pdf")
|
||||
assert len(doc) == 2
|
||||
|
||||
def test_docx_sans_dollar():
|
||||
"""Le document Word ne contient aucun signe dollar ni syntaxe LaTeX résiduelle dans son XML."""
|
||||
src = load_serie06()
|
||||
res = client.post("/render", json={"source": src, "theme": "classique"})
|
||||
assert res.status_code == 200
|
||||
job_id = res.json()["job_id"]
|
||||
|
||||
res_docx = client.get(f"/jobs/{job_id}/docx")
|
||||
assert res_docx.status_code == 200
|
||||
z = zipfile.ZipFile(io.BytesIO(res_docx.content))
|
||||
doc_xml = z.read("word/document.xml").decode("utf-8")
|
||||
assert doc_xml.count("$") == 0
|
||||
|
||||
def test_omml_natif():
|
||||
"""Toutes les formules mathématiques dans Word sont converties en éléments OMML natifs."""
|
||||
src = load_serie06()
|
||||
res = client.post("/render", json={"source": src, "theme": "classique"})
|
||||
assert res.status_code == 200
|
||||
job_id = res.json()["job_id"]
|
||||
|
||||
res_docx = client.get(f"/jobs/{job_id}/docx")
|
||||
z = zipfile.ZipFile(io.BytesIO(res_docx.content))
|
||||
doc_xml = z.read("word/document.xml").decode("utf-8")
|
||||
omath_count = len(re.findall(r"<m:oMath[ >]", doc_xml))
|
||||
assert omath_count == 84
|
||||
|
||||
def test_cartouche_identique():
|
||||
"""Le cartouche officiel (titre, discipline, enseignant, classe, année) est identique dans PDF et DOCX."""
|
||||
src = load_serie06()
|
||||
c = src["cartouche"]
|
||||
res = client.post("/render", json={"source": src, "theme": "classique"})
|
||||
assert res.status_code == 200
|
||||
job_id = res.json()["job_id"]
|
||||
|
||||
pdf_bytes = client.get(f"/jobs/{job_id}/pdf").content
|
||||
doc_pdf = fitz.open(stream=pdf_bytes, filetype="pdf")
|
||||
pdf_txt = "\n".join(p.get_text() for p in doc_pdf)
|
||||
|
||||
docx_bytes = client.get(f"/jobs/{job_id}/docx").content
|
||||
z = zipfile.ZipFile(io.BytesIO(docx_bytes))
|
||||
docx_xml = z.read("word/document.xml").decode("utf-8")
|
||||
|
||||
for champ in [c["titre"], c["enseignant"], c["discipline"], c["annee"]]:
|
||||
assert champ in pdf_txt, f"Champ {champ} manquant dans PDF"
|
||||
assert champ in docx_xml, f"Champ {champ} manquant dans DOCX"
|
||||
|
||||
def test_auteur_enseignant():
|
||||
"""L'auteur des métadonnées du DOCX est l'enseignant déclaré et jamais 'python-docx'."""
|
||||
src = load_serie06()
|
||||
expected_author = src["cartouche"]["enseignant"].strip().lower()
|
||||
res = client.post("/render", json={"source": src, "theme": "classique"})
|
||||
assert res.status_code == 200
|
||||
job_id = res.json()["job_id"]
|
||||
|
||||
docx_bytes = client.get(f"/jobs/{job_id}/docx").content
|
||||
z = zipfile.ZipFile(io.BytesIO(docx_bytes))
|
||||
core_xml = z.read("docProps/core.xml").decode("utf-8")
|
||||
m = re.search(r"<dc:creator>([^<]*)</dc:creator>", core_xml)
|
||||
assert m is not None
|
||||
author = m.group(1).strip().lower()
|
||||
assert author != "python-docx"
|
||||
assert author == expected_author
|
||||
|
||||
def test_theme_inconnu_refuse():
|
||||
"""Une demande de rendu avec un thème inexistant est rejetée avec un code HTTP 400."""
|
||||
src = load_serie06()
|
||||
res = client.post("/render", json={"source": src, "theme": "theme_totalement_inconnu_xyz"})
|
||||
assert res.status_code == 400
|
||||
|
||||
def test_rerendu_autre_theme_sans_modele():
|
||||
"""Le ré-rendu d'un job dans un autre thème s'exécute de façon déterministe sans appel LLM."""
|
||||
src = load_serie06()
|
||||
res1 = client.post("/render", json={"source": src, "theme": "classique"})
|
||||
assert res1.status_code == 200
|
||||
j1 = res1.json()["job_id"]
|
||||
|
||||
res2 = client.post(f"/jobs/{j1}/rerender", json={"theme": "sobre"})
|
||||
assert res2.status_code == 200
|
||||
j2 = res2.json()["job_id"]
|
||||
assert j2 != j1
|
||||
|
||||
# Vérifier que les fichiers du nouveau job existent
|
||||
pdf_res = client.get(f"/jobs/{j2}/pdf")
|
||||
docx_res = client.get(f"/jobs/{j2}/docx")
|
||||
assert pdf_res.status_code == 200
|
||||
assert docx_res.status_code == 200
|
||||
|
||||
def test_figure_declarative():
|
||||
"""Les figures déclaratives sont tracées avec matplotlib à 300 dpi et intégrées au rendu."""
|
||||
import figures
|
||||
out_png = os.path.join(JOBS_DIR, "test_fig_decl.png")
|
||||
spec = {
|
||||
"figsize": [4.0, 4.0],
|
||||
"xlim": [-2, 2],
|
||||
"ylim": [-2, 2],
|
||||
"aspect": "equal",
|
||||
"grid": {"x_step": 1.0, "y_step": 1.0},
|
||||
"axes": True,
|
||||
"elements": [
|
||||
{"kind": "circle", "center": [0, 0], "radius": 1.0, "stroke": "#1e3a8a", "lw": 1.8},
|
||||
{"kind": "point", "x": 1.0, "y": 0, "color": "#b91c1c", "label": "$A$"}
|
||||
]
|
||||
}
|
||||
rendered_path = figures.render_declarative_figure(spec, out_png)
|
||||
assert os.path.exists(rendered_path)
|
||||
assert os.path.getsize(rendered_path) > 1000
|
||||
|
||||
def test_aucun_secret():
|
||||
"""Aucune clé API, jeton d'accès ou mot de passe n'est présent dans le code source de /app."""
|
||||
base_dir = os.path.abspath(os.path.join(os.path.dirname(__file__), ".."))
|
||||
sensitive_pattern = re.compile(r"sk-[A-Za-z0-9]{20,}|Bearer [A-Za-z0-9._~+/=-]{20,}")
|
||||
|
||||
for root, dirs, files in os.walk(base_dir):
|
||||
# Ignorer le dossier de référence et jobs
|
||||
if "ref" in root or "jobs" in root or ".git" in root:
|
||||
continue
|
||||
for fname in files:
|
||||
if fname.endswith((".py", ".json", ".md")):
|
||||
fpath = os.path.join(root, fname)
|
||||
with open(fpath, "r", encoding="utf-8", errors="ignore") as f:
|
||||
content = f.read()
|
||||
assert not sensitive_pattern.search(content), f"Secret potentiel dans {fpath}"
|
||||
Reference in new issue
Block a user