Feature/multi track ingestion

app/alembic/versions/60f5233a9ec0_add_audio_tracks_table_for_multi_track_.py

@@ -0,0 +1,37 @@
+"""Add audio_tracks table for multi-track support
+
+Revision ID: 60f5233a9ec0
+Revises: 434bd387074a
+Create Date: 2025-07-12 19:03:42.738087
+
+"""
+
+from collections.abc import Sequence
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "60f5233a9ec0"
+down_revision: str | Sequence[str] | None = "434bd387074a"
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column("projects", "original_audio_s3_key")
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column(
+        "projects",
+        sa.Column(
+            "original_audio_s3_key", sa.VARCHAR(), autoincrement=False, nullable=True
+        ),
+    )
+    # ### end Alembic commands ###

app/api/v1/endpoints/projects.py

@@ -8,15 +8,10 @@
 from app.api.v1 import dependencies
 from app.core.config import settings
 from app.db import crud
-
-# Import the SQLAlchemy model with an alias to prevent name conflicts
 from app.db.models.project import Project as ProjectModel
 from app.db.models.user import User
-from app.schemas.project import (
-    Project,
-    ProjectCreate,
-    ProjectUpdate,
-)
+from app.schemas.audio_track import AudioTrackCreate
+from app.schemas.project import Project, ProjectCreate, ProjectUpdate
 from app.schemas.s3 import (
     PresignedPostRequest,
     PresignedPostResponse,
@@ -36,19 +31,11 @@ def create_project(
     db: Session = Depends(dependencies.get_db),
     current_user: User = Depends(dependencies.get_current_user),
 ) -> ProjectModel:
-    """
-    Create a new project record for the current user.
-    This does NOT trigger processing; that happens after file upload is complete.
-    """
-    logger.info(
-        "Request received to create project",
-        project_name=project_in.project_name,
-        user_id=current_user.id,
-    )
+    """Create a new project record. Processing begins after file upload completion."""
+    logger.info("Creating project", user_id=current_user.id)
     project = crud.project.create_with_owner(
         db=db, obj_in=project_in, user_id=current_user.id
     )
-    logger.info("Project record created successfully", project_id=str(project.id))
     return project
 
 
@@ -60,28 +47,26 @@ def create_upload_url(
     db: Session = Depends(dependencies.get_db),
     current_user: User = Depends(dependencies.get_current_user),
 ) -> dict[str, Any]:
-    """
-    Generate a pre-signed URL for a client to upload a file directly to S3.
-    """
+    """Generate pre-signed S3 URL for audio track upload."""
     log = logger.bind(project_id=str(project_id), user_id=current_user.id)
-    log.info("Request received to generate S3 upload URL", filename=request_in.filename)
-
     project = crud.project.get(db=db, id=project_id)
+
     if not project or project.user_id != current_user.id:
         raise HTTPException(status_code=404, detail="Project not found")
 
+    log.info("Generating S3 upload URL", filename=request_in.filename)
     object_key = f"uploads/{project.id}/{uuid.uuid4()}-{request_in.filename}"
+
     presigned_data = s3_service.generate_presigned_post(
         bucket_name=settings.AWS_S3_BUCKET_NAME, object_name=object_key
     )
-
     if not presigned_data:
-        log.error("Failed to generate S3 presigned POST URL")
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
-            detail="Could not generate upload URL.",
-        )
-    return cast(dict[str, Any], presigned_data)
+        raise HTTPException(status_code=500, detail="Could not generate upload URL.")
+
+    # Cast to dict[str, Any] to satisfy MyPy
+    result = cast(dict[str, Any], presigned_data)
+    result["s3_key"] = object_key
+    return result
 
 
 @router.post("/{project_id}/complete-upload", response_model=Project)
@@ -92,43 +77,35 @@ def complete_upload(
     db: Session = Depends(dependencies.get_db),
     current_user: User = Depends(dependencies.get_current_user),
 ) -> ProjectModel:
-    """
-    Finalize a multipart upload and trigger the processing pipeline.
-    """
+    """Finalize upload, create track records, and trigger processing."""
     log = logger.bind(project_id=str(project_id), user_id=current_user.id)
-    log.info(
-        "Request received to complete S3 upload", upload_id=completion_in.upload_id
-    )
+    log.info("Completing upload for tracks", track_count=len(completion_in.tracks))
 
     project = crud.project.get(db=db, id=project_id)
     if not project or project.user_id != current_user.id:
         raise HTTPException(status_code=404, detail="Project not found")
 
-    object_key = f"uploads/{project.id}/{completion_in.filename}"
-    parts_data = [part.model_dump() for part in completion_in.parts]
-
-    success = s3_service.complete_multipart_upload(
-        bucket_name=settings.AWS_S3_BUCKET_NAME,
-        object_name=object_key,
-        upload_id=completion_in.upload_id,
-        parts=parts_data,
-    )
-
-    if not success:
-        log.error("Failed to complete multipart upload in S3")
-        raise HTTPException(status_code=500, detail="Could not finalize file upload.")
+    # Create AudioTrack records for each uploaded file
+    for track_data in completion_in.tracks:
+        # Convert CompletedTrack to AudioTrackCreate with only valid fields
+        track_create = AudioTrackCreate(
+            s3_key=track_data.s3_key,
+            speaker_label=track_data.speaker_label,
+        )
+        crud.audio_track.create_with_project(
+            db=db, obj_in=track_create, project_id=uuid.UUID(str(project.id))
+        )
 
-    updated_project = crud.project.update(
-        db=db,
-        db_obj=project,
-        obj_in={"original_audio_s3_key": object_key, "status": "PROCESSING"},
+    # Update project status to trigger processing
+    project = crud.project.update(
+        db=db, db_obj=project, obj_in={"status": "PROCESSING"}
     )
-    log.info("Project status updated to PROCESSING")
+    db.commit()
 
     post_upload_processing.delay(str(project.id))
     log.info("Dispatched background processing task")
 
-    return updated_project
+    return project
 
 
 @router.get("/", response_model=list[Project])
@@ -153,15 +130,10 @@ def read_project(
     db: Session = Depends(dependencies.get_db),
     current_user: User = Depends(dependencies.get_current_user),
 ) -> ProjectModel:
-    """Get a specific project by its ID, ensuring it belongs to the current user."""
-    log = logger.bind(project_id=str(project_id), user_id=current_user.id)
-    log.info("Request received to get project by ID")
-
+    """Retrieve a specific project by ID."""
     project = crud.project.get(db=db, id=project_id)
     if not project or project.user_id != current_user.id:
-        log.warning("User attempted to access a project they do not own")
         raise HTTPException(status_code=404, detail="Project not found")
-
     return project
 
 
@@ -173,17 +145,13 @@ def update_project(
     db: Session = Depends(dependencies.get_db),
     current_user: User = Depends(dependencies.get_current_user),
 ) -> ProjectModel:
-    """Update a project, ensuring it belongs to the current user."""
-    log = logger.bind(project_id=str(project_id), user_id=current_user.id)
-    log.info("Request received to update project")
-
+    """Update project properties."""
     project = crud.project.get(db=db, id=project_id)
     if not project or project.user_id != current_user.id:
-        log.warning("User attempted to update a project they do not own")
         raise HTTPException(status_code=404, detail="Project not found")
 
     updated_project = crud.project.update(db=db, db_obj=project, obj_in=project_in)
-    log.info("Project updated successfully")
+    logger.info("Project updated successfully", project_id=project_id)
     return updated_project
 
 
@@ -194,15 +162,14 @@ def delete_project(
     db: Session = Depends(dependencies.get_db),
     current_user: User = Depends(dependencies.get_current_user),
 ) -> ProjectModel:
-    """Delete a project, ensuring it belongs to the current user."""
-    log = logger.bind(project_id=str(project_id), user_id=current_user.id)
-    log.info("Request received to delete project")
-
+    """Delete a project and associated resources."""
     project = crud.project.get(db=db, id=project_id)
     if not project or project.user_id != current_user.id:
-        log.warning("User attempted to delete a project they do not own")
         raise HTTPException(status_code=404, detail="Project not found")
 
     deleted_project = crud.project.remove(db=db, id=project_id)
-    log.info("Project deleted successfully")
-    return cast(ProjectModel, deleted_project)
+    if not deleted_project:
+        raise HTTPException(status_code=500, detail="Failed to delete project")
+
+    logger.info("Project deleted successfully", project_id=project_id)
+    return deleted_project

app/db/crud/__init__.py

@@ -1,4 +1,6 @@
+from .audio_track import audio_track
 from .project import project
+from .transcript_segment import transcript_segment
 from .user import user
 
-__all__ = ["project", "user"]
+__all__ = ["project", "user", "audio_track", "transcript_segment"]

app/db/crud/audio_track.py

@@ -0,0 +1,24 @@
+import uuid
+
+from sqlalchemy.orm import Session
+
+from app.db.models.audio_track import AudioTrack
+from app.schemas.audio_track import AudioTrackCreate, AudioTrackUpdate
+
+from .base import CRUDBase
+
+
+class CRUDAudioTrack(CRUDBase[AudioTrack, AudioTrackCreate, AudioTrackUpdate]):
+    def create_with_project(
+        self, db: Session, *, obj_in: AudioTrackCreate, project_id: uuid.UUID
+    ) -> AudioTrack:
+        """
+        Create a new audio track and assign it to a project.
+        """
+        db_obj = self.model(**obj_in.model_dump(), project_id=project_id)
+        db.add(db_obj)
+        db.flush()
+        return db_obj
+
+
+audio_track = CRUDAudioTrack(AudioTrack)

app/db/crud/transcript_segment.py

@@ -0,0 +1,31 @@
+import uuid
+
+from sqlalchemy.orm import Session
+
+from app.db.models.transcript_segment import TranscriptSegment
+from app.schemas.transcript_segment import (
+    TranscriptSegmentCreate,
+    TranscriptSegmentUpdate,
+)
+
+from .base import CRUDBase
+
+
+class CRUDTranscriptSegment(
+    CRUDBase[TranscriptSegment, TranscriptSegmentCreate, TranscriptSegmentUpdate]
+):
+    def create_with_project(
+        self, db: Session, *, obj_in: TranscriptSegmentCreate, project_id: uuid.UUID
+    ) -> TranscriptSegment:
+        """
+        Create a new transcript segment and assign it to a project.
+        """
+        # Create the SQLAlchemy model instance, including the project_id
+        db_obj = self.model(**obj_in.model_dump(), project_id=project_id)
+        db.add(db_obj)
+        db.flush()  # Send the change to the DB without committing the transaction
+        db.refresh(db_obj)
+        return db_obj
+
+
+transcript_segment = CRUDTranscriptSegment(TranscriptSegment)

app/db/models/__init__.py

@@ -1,5 +1,7 @@
+from .audio_track import AudioTrack
 from .base import Base
 from .project import Project
+from .transcript_segment import TranscriptSegment
 from .user import User
 
-__all__ = ["Base", "Project", "User"]
+__all__ = ["Base", "Project", "User", "AudioTrack", "TranscriptSegment"]

app/db/models/audio_track.py

@@ -0,0 +1,19 @@
+import uuid
+
+from sqlalchemy import ForeignKey, String
+from sqlalchemy.orm import Mapped, mapped_column, relationship
+
+from .base import Base
+
+
+class AudioTrack(Base):
+    __tablename__ = "audio_tracks"
+
+    id: Mapped[uuid.UUID] = mapped_column(primary_key=True, default=uuid.uuid4)
+    project_id: Mapped[uuid.UUID] = mapped_column(ForeignKey("projects.id"))
+    s3_key: Mapped[str] = mapped_column(String, nullable=False)
+    speaker_label: Mapped[str] = mapped_column(
+        String, nullable=False
+    )  # e.g., "Host", "Guest 1"
+
+    project = relationship("Project", back_populates="audio_tracks")

app/db/models/project.py

@@ -23,7 +23,9 @@ class Project(Base):
     project_name = Column(String(255), nullable=False)
     status = Column(String(50), nullable=False, default="UPLOADING")
     language_code = Column(String(10))
-    original_audio_s3_key: Mapped[str | None] = mapped_column(String, nullable=True)
+    audio_tracks = relationship(
+        "AudioTrack", back_populates="project", cascade="all, delete-orphan"
+    )
     final_audio_s3_key: Mapped[str | None] = mapped_column(String, nullable=True)
     error_message = Column(Text)
 

app/db/models/transcript_segment.py

@@ -0,0 +1,23 @@
+import uuid
+
+from sqlalchemy import Float, ForeignKey, String, Text
+from sqlalchemy.orm import Mapped, mapped_column, relationship
+
+from .base import Base
+
+
+class TranscriptSegment(Base):
+    __tablename__ = "transcript_segments"
+
+    id: Mapped[int] = mapped_column(primary_key=True, index=True)
+    project_id: Mapped[uuid.UUID] = mapped_column(ForeignKey("projects.id"))
+
+    start_time: Mapped[float] = mapped_column(Float, nullable=False)
+    end_time: Mapped[float] = mapped_column(Float, nullable=False)
+    original_text: Mapped[str] = mapped_column(Text, nullable=False)
+    speaker_label: Mapped[str] = mapped_column(String(100), nullable=False)
+
+    # This field will be used later for the editing feature
+    edited_text: Mapped[str | None] = mapped_column(Text, nullable=True)
+
+    project = relationship("Project")