Migrate Flask Setup to FastAPI

[naman9271]

Summary

This issue proposes migrating the current Flask implementation to FastAPI to enhance performance, scalability, and maintainability for the TensorMap application.

Why is this change necessary?

The TensorMap project is focused on Machine Learning (ML) and Deep Learning (DL) applications, which demand high-performance APIs, better concurrency handling, and improved scalability. Flask, while widely used, has some limitations that FastAPI addresses effectively:

---

Migration plan

Phase 1: Setup FastAPI & Initial Configuration

Tasks

• Create a new FastAPI project (tensormap-server-fastapi).
• Install dependencies: fastapi uvicorn pydantic SQLAlchemy alembic python-dotenv
• Set up main.py as the new entry point (equivalent to app.py in Flask).
• Configure CORS middleware for frontend communication.
• Migrate environment configurations from Flask to FastAPI (setup/settings.py).

Files to Migrate

• app.py → main.py
• setup/settings.py
• setup/test_settings.py

Can Be Copied Directly:

• .gitignore

---

Phase 2: Migrate Core API Endpoints

Tasks

• Convert Flask @app.route to FastAPI @app.get, @app.post, etc.
• Replace Flask’s request.json handling with Pydantic models for validation.
• Refactor API responses to return FastAPI’s JSONResponse.
• Update urls.py for each module to use FastAPI’s APIRouter.

Files to Migrate

Data Processing Endpoints

• endpoints/DataProcess/views.py
• endpoints/DataProcess/models.py
• endpoints/DataProcess/services.py
• endpoints/DataProcess/validators.py
• endpoints/DataProcess/urls.py

Data Upload Endpoints

• endpoints/DataUpload/views.py
• endpoints/DataUpload/models.py
• endpoints/DataUpload/services.py
• endpoints/DataUpload/validators.py
• endpoints/DataUpload/urls.py

Deep Learning Endpoints

• endpoints/DeepLearning/modelRun.py
• endpoints/DeepLearning/models.py
• endpoints/DeepLearning/services.py
• endpoints/DeepLearning/validators.py
• endpoints/DeepLearning/urls.py

Can Be Copied Directly:

• constants/errors.py

---

Phase 3: Migrate Database & Migrations

Tasks

• Replace Flask-SQLAlchemy with SQLAlchemy + FastAPI Dependency Injection.
• Update Alembic configurations for FastAPI.
• Ensure database models and session handling work correctly.
• Migrate Alembic migrations.

Files to Migrate

• migrations/env.py
• migrations/alembic.ini
• migrations/script.py.mako
• migrations/versions/*.py
• setup/settings.py (for database settings)

Can Be Copied Directly:

• migrations/README

---

Phase 4: Migrate Testing & Deployment

Tasks

• Refactor tests/unit/ to use FastAPI’s TestClient.
• Update Dockerfile for FastAPI and change Gunicorn to Uvicorn:
• Update .env.example for any required environment variables.
• Test Docker compatibility with FastAPI.
• Ensure all API endpoints work correctly and validate responses.

Files to Migrate

• tests/unit/test_DataProcess.py
• tests/unit/test_DataUpload.py
• tests/unit/conftest.py
• tests/unit/__init__.py
• .flake8
• .gitignore
• .env.example
• Dockerfile

Can Be Copied Directly:

• .flake8
• .gitignore
• .env.example

---

Phase 5: Final Integration & Documentation

Tasks

• Update README.md with FastAPI usage instructions.
• Document all API endpoints with FastAPI’s interactive Swagger UI.
• Final refactoring and performance optimizations.
• Complete final testing before merging PRs.

Files to Migrate

• README.md
• shared/request/response.py
• shared/services/code/generation.py
• shared/services/model/generation.py
• shared/services/model/config.py

Can Be Copied Directly:

• shared/request/
• shared/services/

---

Expected Outcome

• All API endpoints fully migrated to FastAPI.
• Improved performance and async support.
• Cleaner and modularized backend architecture.
• Comprehensive test coverage with pytest and httpx.

---

• Yes, I am willing to open a PR for this issue

[naman9271]

Hey @charithccmc @ivantha please assign this issue to me so that i can start work on this as it will also take time for migration

[naman9271]

Thanks @rajutkarsh07 for assigning this issue
Could you please tell the workflow should i follow for the migration i am thinking to make a

• separate folder for fastapi backend like tensormap-server-fastapi and migrate the whole backend
  What do you say ??

[naman9271]

hey @rajutkarsh07 @ivantha

• I have update the issue with the whole migration plan that how the workflow of this issue will go and i am thinking to open 5 PRs dedicated to each phase so it will get merge and got tested side by side in each phase

• Kindly please Review the Migration Plan so I start working upon it or please tell if i have to make some changes in this plan