Files
aimsight-libredesk/tests/test_pomper.py
T

199 lines
10 KiB
Python
Raw Normal View History

import json
import shutil
import tempfile
import unittest
import urllib.parse
from pathlib import Path
from conversations.pomper import Lecteur, pomper
from faux_libredesk import FauxLibreDesk
LISTE = ("GET", "/api/v1/conversations/all")
def conv(ref, uuid, updated_at):
return {"uuid": uuid, "reference_number": ref, "subject": f"Sujet {ref}", "status": "Open", "inbox_name": "Support",
"created_at": "2026-11-01T08:00:00Z", "updated_at": updated_at, "previous_conversations": [{"uuid": "x"}],
"contact": {"first_name": "Jean", "last_name": "Dupont", "email": "j@x.ch"}}
def piece(nom, url, uuid="a1", taille=3, type_="image/png"):
return {"name": nom, "url": url, "uuid": uuid, "size": taille, "content_type": type_}
def message(heure, pieces=(), contenu="<p>Bonjour</p>"):
return {"uuid": "m" + heure, "created_at": f"2026-11-18T{heure}:00Z", "type": "incoming", "sender_type": "contact",
"private": False, "content": contenu, "text_content": "", "attachments": list(pieces), "author": None}
def pages(elements, taille=2):
"""Une route paginée comme LibreDesk : ?page=N, data = {results, total, per_page, total_pages, page}."""
def reponse(chemin):
page = int(urllib.parse.parse_qs(urllib.parse.urlsplit(chemin).query).get("page", ["1"])[0])
total_pages = max(1, -(-len(elements) // taille))
return {"results": elements[(page - 1) * taille:page * taille], "total": len(elements), "per_page": taille,
"total_pages": total_pages, "page": page}
return reponse
class TestPomper(unittest.TestCase):
def setUp(self):
self.faux, self.s3 = FauxLibreDesk(), FauxLibreDesk()
self.addCleanup(self.faux.arreter)
self.addCleanup(self.s3.arreter)
self.sortie = Path(tempfile.mkdtemp())
self.addCleanup(shutil.rmtree, self.sortie)
self.lignes = []
def lecteur(self):
return Lecteur(self.faux.url, "k", "s")
def pomper(self, **kw):
return pomper(self.lecteur(), self.sortie, journal=self.lignes.append, **kw)
def servir(self, convs, messages=None):
self.faux.reponses[LISTE] = pages(convs)
for c in convs:
self.faux.reponses[("GET", f"/api/v1/conversations/{c['uuid']}/messages")] = pages(
(messages or {}).get(c["uuid"], [message("08:00")]))
def demandes(self, prefixe):
return [c for m, c, _ in self.faux.requetes if urllib.parse.urlsplit(c).path.startswith(prefixe)]
def repere(self):
return json.loads((self.sortie / "pumped.json").read_text())["until"]
def test_tout_au_premier_pompage(self):
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z"), conv("2", "u2", "2026-11-17T12:00:00Z"),
conv("1", "u1", "2026-11-16T12:00:00Z")])
self.assertEqual(self.pomper(), 3)
self.assertEqual(sorted(p.name for p in self.sortie.iterdir() if p.is_dir()), ["1", "2", "3"])
c = json.loads((self.sortie / "3" / "conversation.json").read_text())
self.assertEqual(c["subject"], "Sujet 3")
self.assertNotIn("previous_conversations", c)
self.assertIn("Bonjour", (self.sortie / "3" / "transcript.md").read_text())
self.assertTrue(self.repere())
self.assertIn("order_by=updated_at", self.demandes("/api/v1/conversations/all")[0])
def test_incremental_s_arrete_sous_le_repere(self):
(self.sortie / "pumped.json").write_text(json.dumps({"until": "2026-11-18T10:00:00+00:00"}))
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z"), conv("2", "u2", "2026-11-18T09:30:00Z"),
conv("1", "u1", "2026-11-18T08:00:00Z")])
self.assertEqual(self.pomper(), 2)
self.assertFalse((self.sortie / "1").exists())
self.assertEqual(self.demandes("/api/v1/conversations/u1"), [])
self.assertNotEqual(self.repere(), "2026-11-18T10:00:00+00:00")
def test_interrompu_le_repere_ne_bouge_pas(self):
(self.sortie / "pumped.json").write_text(json.dumps({"until": "2026-11-01T00:00:00+00:00"}))
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z"), conv("2", "u2", "2026-11-17T12:00:00Z")])
self.faux.reponses[("GET", "/api/v1/conversations/u2/messages")] = 500
with self.assertRaises(Exception) as erreur:
self.pomper()
erreur.exception.close()
self.assertTrue((self.sortie / "3" / "transcript.md").exists())
self.assertEqual(self.repere(), "2026-11-01T00:00:00+00:00")
def test_plage_ne_touche_pas_le_repere(self):
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z"), conv("2", "u2", "2026-11-17T12:00:00Z"),
conv("1", "u1", "2026-11-16T12:00:00Z")])
self.assertEqual(self.pomper(depuis="2026-11-17", jusqu_a="2026-11-18"), 1)
self.assertTrue((self.sortie / "2").exists())
self.assertFalse((self.sortie / "3").exists() or (self.sortie / "1").exists())
self.assertFalse((self.sortie / "pumped.json").exists())
def test_pieces_jointes_locales(self):
url = f"{self.faux.url}/uploads/f1?signature=abc&expires=1"
self.faux.reponses[("GET", "/uploads/f1")] = b"PNG"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")],
{"u3": [message("08:00", [piece("photo.png", url)], contenu=f'<img src="{url.replace("&", "&amp;")}">')]})
self.pomper()
m = json.loads((self.sortie / "3" / "messages.json").read_text())[0]
self.assertEqual(m["attachments"][0]["url"], "attachments/photo.png")
self.assertEqual(m["content"], '<img src="attachments/photo.png">')
self.assertEqual((self.sortie / "3" / "attachments" / "photo.png").read_bytes(), b"PNG")
def test_url_relative(self):
self.faux.reponses[("GET", "/uploads/f1")] = b"PNG"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")], {"u3": [message("08:00", [piece("p.png", "/uploads/f1?s=1")])]})
self.pomper()
self.assertEqual((self.sortie / "3" / "attachments" / "p.png").read_bytes(), b"PNG")
def test_nom_de_piece_jointe_assaini(self):
self.faux.reponses[("GET", "/uploads/f1")] = b"abc"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")],
{"u3": [message("08:00", [piece("../../x.sh", "/uploads/f1"), piece("..", "/uploads/f1", uuid="a2")])]})
self.pomper()
self.assertEqual(sorted(p.name for p in (self.sortie / "3" / "attachments").iterdir()), ["a2", "x.sh"])
def test_noms_en_double(self):
self.faux.reponses[("GET", "/uploads/f1")] = b"abc"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")],
{"u3": [message("08:00", [piece("a.png", "/uploads/f1")]),
message("09:00", [piece("a.png", "/uploads/f1", uuid="a2")])]})
self.pomper()
self.assertEqual(sorted(p.name for p in (self.sortie / "3" / "attachments").iterdir()), ["a.png", "a2_a.png"])
def test_deja_telechargee_pas_de_requete(self):
self.faux.reponses[("GET", "/uploads/f1")] = b"abc"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")], {"u3": [message("08:00", [piece("a.png", "/uploads/f1")])]})
self.pomper(depuis="2026-11-18", jusqu_a="2026-11-19")
self.pomper(depuis="2026-11-18", jusqu_a="2026-11-19")
self.assertEqual(len(self.demandes("/uploads/")), 1)
def test_reference_bizarre_dossier_par_uuid(self):
self.servir([conv("../42", "u42", "2026-11-18T12:00:00Z")])
self.pomper()
self.assertTrue((self.sortie / "u42" / "conversation.json").exists())
def test_cle_seulement_vers_libredesk(self):
self.s3.reponses[("GET", "/bucket/f1")] = b"abc"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")],
{"u3": [message("08:00", [piece("a.png", f"{self.s3.url}/bucket/f1?X-Amz-Signature=z")])]})
self.pomper()
self.assertEqual((self.sortie / "3" / "attachments" / "a.png").read_bytes(), b"abc")
self.assertEqual(len(self.s3.entetes), 1)
self.assertNotIn("Authorization", self.s3.entetes[0])
self.assertEqual(self.faux.entetes[0]["Authorization"], "token k:s")
def test_aucune_ecriture(self):
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")])
self.pomper()
self.assertEqual(self.faux.ecritures(), [])
def test_piece_jointe_perdue_sautee(self): # un fichier perdu par LibreDesk ne bloque pas les pompages suivants
self.faux.reponses[("GET", "/uploads/perdu")] = 404
self.faux.reponses[("GET", "/uploads/f1")] = b"abc"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z"), conv("2", "u2", "2026-11-17T12:00:00Z")],
{"u3": [message("08:00", [piece("perdu.png", "/uploads/perdu"), piece("b.png", "/uploads/f1", uuid="a2")])]})
self.assertEqual(self.pomper(), 2)
self.assertTrue(self.repere())
self.assertEqual(sorted(p.name for p in (self.sortie / "3" / "attachments").iterdir()), ["b.png"])
m = json.loads((self.sortie / "3" / "messages.json").read_text())[0]
self.assertEqual(m["attachments"][0]["url"], "")
self.assertTrue(any("perdu.png" in l and "404" in l for l in self.lignes), self.lignes)
def test_nom_trop_long_tronque(self):
self.faux.reponses[("GET", "/uploads/f1")] = b"abc"
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")], {"u3": [message("08:00", [piece("é" * 150 + ".png", "/uploads/f1")])]})
self.pomper()
(f,) = (self.sortie / "3" / "attachments").iterdir()
self.assertLessEqual(len(f.name.encode()), 200)
self.assertTrue(f.name.endswith(".png"))
def test_schema_autre_que_http_refuse(self): # jamais un fichier local recopié dans le dossier commité
self.servir([conv("3", "u3", "2026-11-18T12:00:00Z")],
{"u3": [message("08:00", [piece("e.txt", "file:///etc/hostname")])]})
self.pomper()
self.assertFalse((self.sortie / "3" / "attachments").exists())
def test_ordre_non_respecte(self): # LibreDesk ne trie pas par updated_at : on s'arrête, le repère ne bouge pas
(self.sortie / "pumped.json").write_text(json.dumps({"until": "2026-11-01T00:00:00+00:00"}))
self.servir([conv("2", "u2", "2026-11-17T12:00:00Z"), conv("3", "u3", "2026-11-18T12:00:00Z")])
with self.assertRaisesRegex(ValueError, "ordre"):
self.pomper()
self.assertEqual(self.repere(), "2026-11-01T00:00:00+00:00")
if __name__ == "__main__":
unittest.main()