🎨 Add ruff & pyright

.pre-commit-config.yaml

@@ -15,20 +15,30 @@ repos:
       - id: trailing-whitespace
 
   # python checks
-  - repo: https://github.com/PyCQA/isort
-    rev: 5.12.0
-    hooks:
-      - id: isort
-        args: ["--profile", "black", "--filter-files"]
   - repo: https://github.com/psf/black
     rev: 23.1.0
     hooks:
       - id: black
-  - repo: https://github.com/pycqa/flake8
-    rev: 6.0.0
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.0.272
+    hooks:
+      - id: ruff
+  - repo: local
+    hooks:
+      - id: backend-pyright
+        name: Run typechecker for backend
+        entry: pdm run -p backend/ pyright backend/transcribee_backend
+        language: system
+        files: backend/.*
+        pass_filenames: false
+  - repo: local
     hooks:
-      - id: flake8
-        exclude: ^backend/.*/migrations/.*$
+      - id: worker-pyright
+        name: Run typechecker for worker
+        entry: pdm run -p worker/ pyright worker/transcribee_worker
+        language: system
+        files: worker/.*
+        pass_filenames: false
   - repo: local
     hooks:
       - id: backend-openapi

.ruff.toml

@@ -0,0 +1 @@
+line-length = 100

backend/pyproject.toml

@@ -13,7 +13,7 @@ dependencies = [
     "fastapi>=0.92.0",
     "uvicorn[standard]>=0.20.0",
     "sqlmodel @ git+https://github.com/transcribee/sqlmodel.git@transcribee_main",
-    "alembic>=1.9.4",
+    "alembic>=1.11.1",
     "transcribee-proto @ file:///${PROJECT_ROOT}/../proto",
     "python-multipart>=0.0.6",
     "filetype>=1.2.0",
@@ -40,6 +40,7 @@ dev = [
     "pytest>=7.3.1",
     "httpx>=0.24.0",
     "pytest-alembic>=0.10.4",
+    "pyright>=1.1.314",
 ]
 
 [tool.pdm.scripts]

backend/scripts/reset_task.py

@@ -13,7 +13,7 @@
     )
     args = parser.parse_args()
     with SessionContextManager() as session:
-        task = session.exec(
+        task = session.execute(
             update(Task)
             .where(
                 or_(Task.id == args.uuid, Task.document_id == args.uuid),

backend/transcribee_backend/db/migrations/env.py

@@ -4,7 +4,7 @@
 from sqlalchemy import MetaData, engine_from_config, pool
 from sqlmodel import SQLModel
 from transcribee_backend.db import DATABASE_URL
-from transcribee_backend.models import *
+from transcribee_backend.models import *  # noqa
 
 # this is the Alembic Config object, which provides
 # access to the values within the .ini file in use.

backend/transcribee_backend/db/migrations/versions/14b5324ec3e2_add_task_completion_data_document_.py

@@ -6,7 +6,6 @@
 
 """
 import sqlalchemy as sa
-import sqlmodel
 from alembic import op
 
 # revision identifiers, used by Alembic.

backend/transcribee_backend/db/migrations/versions/1ec1e251173a_init.py

@@ -7,6 +7,7 @@
 """
 import sqlalchemy as sa
 import sqlmodel
+import sqlmodel.sql.sqltypes
 from alembic import op
 
 # revision identifiers, used by Alembic.

backend/transcribee_backend/db/migrations/versions/6392770332cd_add_taskattempt.py

@@ -1,3 +1,4 @@
+# pyright: basic
 """add TaskAttempt
 
 Revision ID: 6392770332cd
@@ -8,12 +9,12 @@
 import uuid
 
 import sqlalchemy as sa
+from sqlalchemy.sql.operators import is_
+import sqlmodel.sql.sqltypes
 import sqlmodel
 from alembic import op
-from sqlalchemy.dialects import sqlite
 from transcribee_backend.config import settings
 from transcribee_backend.helpers.time import now_tz_aware
-from transcribee_backend.models.task import TaskState
 
 # revision identifiers, used by Alembic.
 revision = "6392770332cd"
@@ -76,7 +77,7 @@ def upgrade_with_autocommit() -> None:
             "fk_task_assigned_worker_id_worker", type_="foreignkey"
         )
         batch_op.create_foreign_key(
-            None,
+            None,  # type: ignore
             "taskattempt",
             ["current_attempt_id"],
             ["id"],
@@ -107,11 +108,11 @@ def upgrade_with_autocommit() -> None:
     )
     bind = op.get_bind()
     session = sqlmodel.Session(bind)
-    for task in session.exec(
-        Task.select().where(~sa.sql.operators.is_(Task.c.assigned_worker_id, None))
+    for task in session.execute(
+        Task.select().where(~is_(Task.c.assigned_worker_id, None))
     ).all():
         attempt_id = uuid.uuid4()
-        session.exec(
+        session.execute(
             sqlmodel.insert(TaskAttempt).values(
                 id=attempt_id,
                 task_id=task.id,
@@ -124,7 +125,7 @@ def upgrade_with_autocommit() -> None:
                 attempt_number=1,
             )
         )
-        session.exec(
+        session.execute(
             sqlmodel.update(Task)
             .where(Task.c.id == task.id)
             .values(
@@ -139,7 +140,7 @@ def upgrade_with_autocommit() -> None:
         now = now_tz_aware()
         batch_op.execute(
             sqlmodel.update(Task)
-            .where(Task.c.is_completed == True)
+            .where(Task.c.is_completed.is_(True))
             .values(state="COMPLETED", state_changed_at=Task.c.completed_at)
         )
         batch_op.execute(

backend/transcribee_backend/db/migrations/versions/fc7c9d045329_cascade_document_task_deletion.py

@@ -6,7 +6,6 @@
 
 """
 import sqlalchemy as sa
-import sqlmodel
 from alembic import op
 
 # revision identifiers, used by Alembic.
@@ -24,7 +23,11 @@ def upgrade() -> None:
         )
         batch_op.drop_constraint("fk_task_document_id_document", type_="foreignkey")
         batch_op.create_foreign_key(
-            None, "document", ["document_id"], ["id"], ondelete="CASCADE"
+            None,  # type: ignore
+            "document",
+            ["document_id"],
+            ["id"],
+            ondelete="CASCADE",
         )
 
     with op.batch_alter_table("taskdependency", schema=None) as batch_op:
@@ -41,10 +44,18 @@ def upgrade() -> None:
             "fk_taskdependency_dependant_on_id_task", type_="foreignkey"
         )
         batch_op.create_foreign_key(
-            None, "task", ["dependent_task_id"], ["id"], ondelete="CASCADE"
+            None,  # type: ignore
+            "task",
+            ["dependent_task_id"],
+            ["id"],
+            ondelete="CASCADE",
         )
         batch_op.create_foreign_key(
-            None, "task", ["dependant_on_id"], ["id"], ondelete="CASCADE"
+            None,  # type: ignore
+            "task",
+            ["dependant_on_id"],
+            ["id"],
+            ondelete="CASCADE",
         )
 
     # ### end Alembic commands ###
@@ -53,8 +64,8 @@ def upgrade() -> None:
 def downgrade() -> None:
     # ### commands auto generated by Alembic - please adjust! ###
     with op.batch_alter_table("taskdependency", schema=None) as batch_op:
-        batch_op.drop_constraint(None, type_="foreignkey")
-        batch_op.drop_constraint(None, type_="foreignkey")
+        batch_op.drop_constraint(None, type_="foreignkey")  # type: ignore
+        batch_op.drop_constraint(None, type_="foreignkey")  # type: ignore
         batch_op.create_foreign_key(
             "fk_taskdependency_dependant_on_id_task",
             "task",
@@ -75,7 +86,7 @@ def downgrade() -> None:
         )
 
     with op.batch_alter_table("task", schema=None) as batch_op:
-        batch_op.drop_constraint(None, type_="foreignkey")
+        batch_op.drop_constraint(None, type_="foreignkey")  # type: ignore
         batch_op.create_foreign_key(
             "fk_task_document_id_document", "document", ["document_id"], ["id"]
         )

backend/transcribee_backend/helpers/tasks.py

@@ -1,7 +1,7 @@
 import datetime
 from typing import Iterable, Optional
 
-from sqlmodel import Session, select
+from sqlmodel import Session, select, col
 from transcribee_backend.config import settings
 from transcribee_backend.db import SessionContextManager
 from transcribee_backend.helpers.time import now_tz_aware
@@ -18,6 +18,9 @@ def finish_current_attempt(
     if now is None:
         now = now_tz_aware()
 
+    if task.current_attempt is None:
+        return
+
     task.current_attempt.ended_at = now
     task.current_attempt.last_keepalive = now
     task.current_attempt.extra_data = extra_data
@@ -43,8 +46,8 @@ def timeouted_tasks(session: Session) -> Iterable[Task]:
     statement = (
         select(Task)
         .where(
-            Task.current_attempt.has(
-                TaskAttempt.last_keepalive
+            col(Task.current_attempt).has(
+                col(TaskAttempt.last_keepalive)
                 < now_tz_aware() - datetime.timedelta(seconds=settings.worker_timeout)
             ),
         )

backend/transcribee_backend/models/task.py

@@ -3,7 +3,7 @@
 import uuid
 from typing import Any, Dict, List, Literal, Optional
 
-from sqlmodel import JSON, Column, Field, ForeignKey, Relationship, SQLModel
+from sqlmodel import JSON, Column, Field, ForeignKey, Relationship, SQLModel, col
 from sqlmodel.sql.sqltypes import GUID
 from transcribee_backend.config import settings
 from transcribee_backend.helpers.time import now_tz_aware
@@ -74,6 +74,7 @@ class Task(TaskBase, table=True):
         sa_column=Column(
             GUID, ForeignKey("taskattempt.id", ondelete="SET NULL", use_alter=True)
         ),
+        default=None,
     )
     current_attempt: Optional["TaskAttempt"] = Relationship(
         sa_relationship_kwargs={
@@ -112,7 +113,9 @@ class TaskAttempt(SQLModel, table=True):
     )
 
     task_id: uuid.UUID = Field(
-        sa_column=Column(GUID, ForeignKey(Task.id, ondelete="CASCADE"), nullable=False),
+        sa_column=Column(
+            GUID, ForeignKey(col(Task.id), ondelete="CASCADE"), nullable=False
+        ),
         unique=False,
     )
     task: Task = Relationship(