generated from john/python-template
304 lines
10 KiB
Python
304 lines
10 KiB
Python
"""Tests for transcription.models — Document, Job, Transcript, TranscriptRevision models."""
|
|
|
|
from uuid import UUID
|
|
|
|
import pytest
|
|
from sqlalchemy.exc import IntegrityError
|
|
|
|
from transcription.models import Document, Job, JobStatus, Transcript, TranscriptRevision
|
|
|
|
|
|
def _make_document(**overrides) -> Document:
|
|
"""Create a Document with sensible defaults."""
|
|
defaults = {"filename": "letter.jpg", "file_path": "/uploads/letter.jpg"}
|
|
defaults.update(overrides)
|
|
return Document(**defaults)
|
|
|
|
|
|
def _persist_document(session) -> Document:
|
|
"""Create, persist, and return a Document."""
|
|
doc = _make_document()
|
|
session.add(doc)
|
|
session.commit()
|
|
session.refresh(doc)
|
|
return doc
|
|
|
|
|
|
def _persist_job(session, document: Document) -> Job:
|
|
"""Create, persist, and return a Job linked to a Document."""
|
|
job = Job(document_id=document.id)
|
|
session.add(job)
|
|
session.commit()
|
|
session.refresh(job)
|
|
return job
|
|
|
|
|
|
class TestDocumentModel:
|
|
"""Verify Document creation and default field population."""
|
|
|
|
def test_can_be_persisted(self, session):
|
|
"""A Document round-trips through the database with correct fields."""
|
|
doc = _persist_document(session)
|
|
fetched = session.get(Document, doc.id)
|
|
assert fetched is not None
|
|
assert fetched.filename == "letter.jpg"
|
|
assert fetched.file_path == "/uploads/letter.jpg"
|
|
|
|
def test_defaults_are_populated(self, session):
|
|
"""id is a UUID and uploaded_at is populated on creation."""
|
|
doc = _persist_document(session)
|
|
assert isinstance(doc.id, UUID)
|
|
assert doc.uploaded_at is not None
|
|
|
|
|
|
class TestJobModel:
|
|
"""Verify Job creation, defaults, and status transitions."""
|
|
|
|
def test_can_be_created_for_document(self, session):
|
|
"""A Job linked to a Document via FK persists correctly."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
fetched = session.get(Job, job.id)
|
|
assert fetched is not None
|
|
assert fetched.document_id == doc.id
|
|
|
|
def test_defaults_are_populated(self, session):
|
|
"""Default status is queued; created_at and updated_at are populated."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
assert job.status == JobStatus.QUEUED
|
|
assert job.retry_count == 0
|
|
assert job.created_at is not None
|
|
assert job.updated_at is not None
|
|
|
|
def test_transitions_to_transcribed(self, session):
|
|
"""Status updates from queued to processing to transcribed."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
assert job.status == JobStatus.QUEUED
|
|
|
|
job.status = JobStatus.PROCESSING
|
|
session.add(job)
|
|
session.commit()
|
|
session.refresh(job)
|
|
assert job.status == JobStatus.PROCESSING
|
|
|
|
job.status = JobStatus.TRANSCRIBED
|
|
session.add(job)
|
|
session.commit()
|
|
session.refresh(job)
|
|
assert job.status == JobStatus.TRANSCRIBED
|
|
|
|
def test_transitions_to_failed(self, session):
|
|
"""Status updates from processing to failed."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
|
|
job.status = JobStatus.PROCESSING
|
|
session.add(job)
|
|
session.commit()
|
|
session.refresh(job)
|
|
|
|
job.status = JobStatus.FAILED
|
|
session.add(job)
|
|
session.commit()
|
|
session.refresh(job)
|
|
assert job.status == JobStatus.FAILED
|
|
|
|
|
|
class TestTranscriptModel:
|
|
"""Verify Transcript persistence for success and failure cases."""
|
|
|
|
def test_success_record_persists(self, session):
|
|
"""A Transcript with text set and error_detail None persists correctly."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
transcript = Transcript(job_id=job.id, provider="openrouter", prompt_name="transcribe_document.md", text="Dear Sir, ...")
|
|
session.add(transcript)
|
|
session.commit()
|
|
session.refresh(transcript)
|
|
|
|
fetched = session.get(Transcript, transcript.id)
|
|
assert fetched is not None
|
|
assert fetched.text == "Dear Sir, ..."
|
|
assert fetched.error_detail is None
|
|
|
|
def test_failure_record_persists(self, session):
|
|
"""A Transcript with text None and error_detail set persists correctly."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
transcript = Transcript(
|
|
job_id=job.id,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
error_detail="Provider timeout",
|
|
)
|
|
session.add(transcript)
|
|
session.commit()
|
|
session.refresh(transcript)
|
|
|
|
fetched = session.get(Transcript, transcript.id)
|
|
assert fetched is not None
|
|
assert fetched.text is None
|
|
assert fetched.error_detail == "Provider timeout"
|
|
|
|
def test_job_id_is_unique(self, session):
|
|
"""Inserting two transcripts with the same job_id raises an integrity error."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
|
|
t1 = Transcript(job_id=job.id, provider="openrouter", prompt_name="transcribe_document.md", text="First")
|
|
session.add(t1)
|
|
session.commit()
|
|
|
|
t2 = Transcript(job_id=job.id, provider="openrouter", prompt_name="transcribe_document.md", text="Duplicate")
|
|
session.add(t2)
|
|
with pytest.raises(IntegrityError):
|
|
session.commit()
|
|
|
|
|
|
class TestTranscriptRevisionModel:
|
|
"""Verify TranscriptRevision persistence and version uniqueness constraints."""
|
|
|
|
def test_revision_record_persists(self, session):
|
|
"""A TranscriptRevision with version metadata persists correctly."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
revision = TranscriptRevision(
|
|
job_id=job.id,
|
|
version_number=1,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
model="google/gemini-2.5-flash",
|
|
source="ai",
|
|
text="Initial text",
|
|
)
|
|
session.add(revision)
|
|
session.commit()
|
|
session.refresh(revision)
|
|
|
|
fetched = session.get(TranscriptRevision, revision.id)
|
|
assert fetched is not None
|
|
assert fetched.version_number == 1
|
|
assert fetched.text == "Initial text"
|
|
assert fetched.source == "ai"
|
|
|
|
def test_job_version_pair_is_unique(self, session):
|
|
"""Duplicate version_number for same job raises integrity error."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
|
|
first = TranscriptRevision(
|
|
job_id=job.id,
|
|
version_number=1,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
source="ai",
|
|
text="Initial",
|
|
)
|
|
duplicate = TranscriptRevision(
|
|
job_id=job.id,
|
|
version_number=1,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
source="user",
|
|
text="Edited",
|
|
)
|
|
session.add(first)
|
|
session.commit()
|
|
|
|
session.add(duplicate)
|
|
with pytest.raises(IntegrityError):
|
|
session.commit()
|
|
|
|
def test_same_version_number_allowed_for_different_jobs(self, session):
|
|
"""Version numbers are scoped per job, not globally."""
|
|
doc1 = _persist_document(session)
|
|
job1 = _persist_job(session, doc1)
|
|
doc2 = _make_document(filename="letter2.jpg", file_path="/uploads/letter2.jpg")
|
|
session.add(doc2)
|
|
session.commit()
|
|
session.refresh(doc2)
|
|
job2 = _persist_job(session, doc2)
|
|
|
|
r1 = TranscriptRevision(
|
|
job_id=job1.id,
|
|
version_number=1,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
source="ai",
|
|
text="Job1 v1",
|
|
)
|
|
r2 = TranscriptRevision(
|
|
job_id=job2.id,
|
|
version_number=1,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
source="ai",
|
|
text="Job2 v1",
|
|
)
|
|
session.add(r1)
|
|
session.add(r2)
|
|
session.commit()
|
|
|
|
|
|
class TestRelationships:
|
|
"""Verify SQLModel relationship navigation between models."""
|
|
|
|
def test_document_exposes_jobs(self, session):
|
|
"""document.jobs returns the linked Job list."""
|
|
doc = _persist_document(session)
|
|
_persist_job(session, doc)
|
|
_persist_job(session, doc)
|
|
|
|
session.refresh(doc)
|
|
assert len(doc.jobs) == 2
|
|
assert all(isinstance(j, Job) for j in doc.jobs)
|
|
|
|
def test_job_exposes_transcript(self, session):
|
|
"""job.transcript returns the linked Transcript."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
transcript = Transcript(
|
|
job_id=job.id,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
text="Transcribed text",
|
|
)
|
|
session.add(transcript)
|
|
session.commit()
|
|
|
|
session.refresh(job)
|
|
assert job.transcript is not None
|
|
assert isinstance(job.transcript, Transcript)
|
|
assert job.transcript.text == "Transcribed text"
|
|
|
|
def test_job_exposes_transcript_revisions(self, session):
|
|
"""job.transcript_revisions returns all linked revisions."""
|
|
doc = _persist_document(session)
|
|
job = _persist_job(session, doc)
|
|
session.add(
|
|
TranscriptRevision(
|
|
job_id=job.id,
|
|
version_number=1,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
source="ai",
|
|
text="v1",
|
|
)
|
|
)
|
|
session.add(
|
|
TranscriptRevision(
|
|
job_id=job.id,
|
|
version_number=2,
|
|
provider="openrouter",
|
|
prompt_name="transcribe_document.md",
|
|
source="user",
|
|
text="v2",
|
|
)
|
|
)
|
|
session.commit()
|
|
|
|
session.refresh(job)
|
|
assert len(job.transcript_revisions) == 2
|