Step 1 complete!

This commit is contained in:
Jim Lancaster
2026-06-24 11:49:00 -05:00
parent 5165fa64bc
commit cba4d890a1
4 changed files with 221 additions and 5 deletions
+2 -1
View File
@@ -6,6 +6,7 @@ Three models capture the MVP lifecycle:
from datetime import datetime, timezone
from enum import StrEnum
from typing import Optional
from uuid import UUID, uuid4
from sqlmodel import Field, Relationship, SQLModel
@@ -47,7 +48,7 @@ class Job(SQLModel, table=True):
# --- relationships ---
document: Document = Relationship(back_populates="jobs")
transcript: "Transcript | None" = Relationship(back_populates="job")
transcript: Optional["Transcript"] = Relationship(back_populates="job")
class Transcript(SQLModel, table=True):
+32 -2
View File
@@ -1,16 +1,34 @@
"""Tests for transcription.config — settings loading, provider defaults, and paths."""
from pathlib import Path
import pytest
from pydantic import ValidationError
from transcription.config import Provider, Settings
def _make_settings(**overrides) -> Settings:
"""Build a Settings instance with a dummy API key unless overridden."""
defaults = {"openrouter_api_key": "test-key-abc123"}
defaults.update(overrides)
return Settings(**defaults)
class TestSettingsLoading:
"""Verify Settings construction and required-field validation."""
def test_loads_from_env(self):
def test_loads_from_env(self, monkeypatch):
"""Settings constructs when OPENROUTER_API_KEY is provided."""
monkeypatch.setenv("OPENROUTER_API_KEY", "test-key-xyz")
settings = Settings()
assert settings.openrouter_api_key == "test-key-xyz"
def test_requires_api_key(self):
def test_requires_api_key(self, monkeypatch):
"""Settings raises ValidationError when OPENROUTER_API_KEY is missing."""
monkeypatch.delenv("OPENROUTER_API_KEY", raising=False)
with pytest.raises(ValidationError):
Settings(_env_file=None)
class TestProviderSettings:
@@ -18,12 +36,21 @@ class TestProviderSettings:
def test_defaults_to_openrouter(self):
"""Default provider is openrouter when not explicitly set."""
settings = _make_settings()
assert settings.provider == Provider.OPENROUTER
assert settings.provider == "openrouter"
def test_rejects_invalid_value(self):
"""Setting PROVIDER to an invalid value raises ValidationError."""
with pytest.raises(ValidationError):
_make_settings(provider="not-a-provider")
def test_optional_fields_default_to_none(self):
"""provider_model, openrouter_http_referer, and openrouter_app_title are None when unset."""
settings = _make_settings()
assert settings.provider_model is None
assert settings.openrouter_http_referer is None
assert settings.openrouter_app_title is None
class TestPathSettings:
@@ -31,3 +58,6 @@ class TestPathSettings:
def test_path_fields_are_path_objects(self):
"""upload_dir and prompt_dir are Path instances."""
settings = _make_settings()
assert isinstance(settings.upload_dir, Path)
assert isinstance(settings.prompt_dir, Path)
+61 -2
View File
@@ -1,18 +1,77 @@
"""Tests for transcription.db — schema bootstrap and session factory."""
from unittest.mock import patch
from sqlalchemy import inspect, text
from sqlmodel import Session, SQLModel, create_engine
from sqlmodel.pool import StaticPool
def _in_memory_engine():
"""Create a fresh in-memory SQLite engine for isolated db tests."""
return create_engine(
"sqlite://",
connect_args={"check_same_thread": False},
poolclass=StaticPool,
)
class TestSchemaBootstrap:
"""Verify create_all produces the expected table set."""
def test_create_all_creates_expected_tables(self):
"""After create_all(), document, job, and transcript tables exist."""
engine = _in_memory_engine()
# Ensure models are imported so metadata is populated
from transcription.models import Document, Job, Transcript # noqa: F401
SQLModel.metadata.create_all(engine)
inspector = inspect(engine)
table_names = set(inspector.get_table_names())
assert "document" in table_names
assert "job" in table_names
assert "transcript" in table_names
class TestSessionFactory:
"""Verify get_session yields and cleans up sessions."""
def test_get_session_yields_session(self):
def test_get_session_yields_session(self, monkeypatch):
"""get_session() yields a usable Session object."""
monkeypatch.setenv("OPENROUTER_API_KEY", "test-key-for-db")
# Clear the lru_cache so Settings is re-created with our env var
from transcription.config import get_settings
get_settings.cache_clear()
def test_session_is_closed_after_generator_exit(self):
engine = _in_memory_engine()
SQLModel.metadata.create_all(engine)
import transcription.db as db_module
with patch.object(db_module, "engine", engine):
with db_module.get_session() as session:
assert isinstance(session, Session)
get_settings.cache_clear()
def test_session_is_closed_after_generator_exit(self, monkeypatch):
"""After the context manager exits, the session is closed."""
monkeypatch.setenv("OPENROUTER_API_KEY", "test-key-for-db")
from transcription.config import get_settings
get_settings.cache_clear()
engine = _in_memory_engine()
SQLModel.metadata.create_all(engine)
import transcription.db as db_module
with patch.object(db_module, "engine", engine):
with db_module.get_session() as session:
# Session is usable inside the context
session.execute(text("SELECT 1"))
captured = session
# After exiting, the session's internal connection is released
# (no active transaction bound to the session)
assert captured._transaction is None
get_settings.cache_clear()
+126
View File
@@ -1,16 +1,54 @@
"""Tests for transcription.models — Document, Job, Transcript persistence and relationships."""
from uuid import UUID
import pytest
from sqlalchemy.exc import IntegrityError
from transcription.models import Document, Job, JobStatus, Transcript
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:
@@ -18,15 +56,53 @@ class TestJobModel:
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.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:
@@ -34,12 +110,45 @@ class TestTranscriptModel:
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, 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, 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, text="First")
session.add(t1)
session.commit()
t2 = Transcript(job_id=job.id, text="Duplicate")
session.add(t2)
with pytest.raises(IntegrityError):
session.commit()
class TestRelationships:
@@ -47,6 +156,23 @@ class TestRelationships:
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, 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"