Update job detail page to track revisions and display document image next to transcription text.

This commit is contained in:
Jim Lancaster
2026-06-28 18:02:19 -05:00
parent e2e421835f
commit 761765636a
19 changed files with 744 additions and 111 deletions
+131 -7
View File
@@ -1,11 +1,11 @@
"""Tests for transcription.models — Document, Job, Transcript persistence and relationships."""
"""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
from transcription.models import Document, Job, JobStatus, Transcript, TranscriptRevision
def _make_document(**overrides) -> Document:
@@ -113,7 +113,7 @@ class TestTranscriptModel:
"""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, text="Dear Sir, ...")
transcript = Transcript(job_id=job.id, provider="openrouter", prompt_name="transcribe_document.md", text="Dear Sir, ...")
session.add(transcript)
session.commit()
session.refresh(transcript)
@@ -127,7 +127,12 @@ class TestTranscriptModel:
"""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, error_detail="Provider timeout")
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)
@@ -142,16 +147,101 @@ class TestTranscriptModel:
doc = _persist_document(session)
job = _persist_job(session, doc)
t1 = Transcript(job_id=job.id, text="First")
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, text="Duplicate")
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."""
@@ -169,7 +259,12 @@ class TestRelationships:
"""job.transcript returns the linked Transcript."""
doc = _persist_document(session)
job = _persist_job(session, doc)
transcript = Transcript(job_id=job.id, text="Transcribed text")
transcript = Transcript(
job_id=job.id,
provider="openrouter",
prompt_name="transcribe_document.md",
text="Transcribed text",
)
session.add(transcript)
session.commit()
@@ -177,3 +272,32 @@ class TestRelationships:
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