139 lines
5.3 KiB
Markdown
139 lines
5.3 KiB
Markdown
# Migration Plan: Alwrity (AI-Writer) to Enterprise-Ready Architecture
|
||
|
||
## 1. Background & Motivation
|
||
Alwrity (AI-Writer) is currently an open-source, Streamlit-based project for AI-powered content creation, SEO, analytics, and more. To serve enterprise customers, we need to move to a scalable, secure, and maintainable architecture, reusing as much of the existing Python codebase as possible while replacing the UI and improving backend robustness.
|
||
|
||
---
|
||
|
||
## 2. Current State
|
||
- **UI:** Streamlit (great for prototyping, not for enterprise)
|
||
- **Backend:** Python modules for AI writing, SEO, analytics, chatbot, etc.
|
||
- **Database:** SQLite, ChromaDB, some service layers for Twitter and content
|
||
- **AI/ML:** Integrates with OpenAI, Gemini, and other providers
|
||
|
||
---
|
||
|
||
## 3. Design Directions & Tech Stack Recommendations
|
||
|
||
### A. Frontend
|
||
- **React** (TypeScript) for scalable, maintainable UI
|
||
- **UI Library:** Material-UI (MUI) or Ant Design
|
||
- **State/Data:** React Query, Context API or Redux Toolkit
|
||
|
||
### B. Backend
|
||
- **FastAPI** (Python): async, high-performance, easy to wrap existing modules
|
||
- **Task Queue:** Celery + Redis for background jobs (if needed)
|
||
|
||
### C. Database & Storage
|
||
- **PostgreSQL** for structured data
|
||
- **Redis** for caching and task queue
|
||
- **Vector DB:** Pinecone, Weaviate, or Qdrant for semantic search (if needed)
|
||
- **Blob Storage:** AWS S3 or Azure Blob for files
|
||
|
||
### D. AI/ML Integration
|
||
- Reuse existing Python modules
|
||
- Serve custom models via FastAPI endpoints
|
||
|
||
### E. Authentication
|
||
- **Auth0** or **Keycloak** for OAuth2/SSO, or FastAPI JWT for MVP
|
||
|
||
### F. DevOps
|
||
- **Docker** for containerization
|
||
- **GitHub Actions** for CI/CD
|
||
- **(Optional) Kubernetes** for orchestration
|
||
|
||
### G. Security & Compliance
|
||
- SSO, RBAC, audit logs, encryption, GDPR/SOC2 readiness
|
||
|
||
---
|
||
|
||
## 4. Migration Plan: Step-by-Step
|
||
|
||
### Phase 1: Preparation
|
||
- Audit codebase for reusable business logic
|
||
- Separate UI code from backend logic
|
||
- Set up monorepo or separate repos for backend (Python/FastAPI) and frontend (React)
|
||
|
||
### Phase 2: Backend API Layer
|
||
- Scaffold FastAPI app
|
||
- Wrap existing Python modules as API endpoints (content generation, SEO, analytics, etc.)
|
||
- Add authentication (JWT for MVP, SSO for production)
|
||
- Write unit/integration tests
|
||
|
||
### Phase 3: Frontend Migration
|
||
- Scaffold React app (TypeScript)
|
||
- Set up routing, authentication, dashboard layout
|
||
- For each Streamlit feature, create a React page/component
|
||
- Use MUI/Ant Design for UI
|
||
- Fetch data from FastAPI using React Query
|
||
|
||
### Phase 4: Feature Parity & Enhancements
|
||
- Migrate all features, one by one, to new stack
|
||
- Use Celery + Redis for long-running jobs
|
||
- Add UI/UX improvements (loading, error handling, feedback)
|
||
|
||
### Phase 5: Productionization
|
||
- Dockerize frontend and backend
|
||
- Set up CI/CD with GitHub Actions
|
||
- Add logging, monitoring (Sentry, Prometheus, Grafana)
|
||
- Harden security (HTTPS, CORS, secure cookies, etc.)
|
||
|
||
### Phase 6: Launch & Iterate
|
||
- Deploy to cloud
|
||
- Gather user feedback and iterate
|
||
|
||
---
|
||
|
||
## 5. Prioritized Modules for Migration
|
||
|
||
### Best-fit modules to start with (already decoupled from UI):
|
||
1. **AI Writers (lib/ai_writers/):** Blog, news, social, email, story, YouTube script writers
|
||
2. **SEO Tools (lib/ai_seo_tools/):** Keyword analyzer, meta generator, content gap, enterprise SEO, content calendar
|
||
3. **Website Analyzer (lib/utils/website_analyzer/):** Performance, SEO, content quality analysis
|
||
4. **Analytics/Performance (lib/content_performance_predictor/):** Content analytics and prediction
|
||
5. **Chatbot Core (lib/chatbot_custom/core/):** Workflow engine, tool router, intent analyzer, context manager
|
||
6. **Database Services (lib/database/):** Twitter and content management service layers
|
||
7. **AI Marketing Tools (lib/ai_marketing_tools/ai_backlinker/):** Backlinking and marketing automation
|
||
|
||
### Modules to avoid for now:
|
||
- Streamlit UI scripts and thin wrappers
|
||
|
||
---
|
||
|
||
## 6. Summary Table
|
||
|
||
| Layer | Stack/Tooling | Why? |
|
||
|---------------|-----------------------------|--------------------------------------------|
|
||
| Frontend | React + TypeScript + MUI | Modern, scalable, huge ecosystem |
|
||
| Backend | FastAPI (Python) | Async, high-perf, easy to wrap old code |
|
||
| Auth | FastAPI JWT/Auth0/Keycloak | Secure, enterprise-ready |
|
||
| DB | PostgreSQL, Redis | Reliable, scalable, Python-friendly |
|
||
| AI/ML | Existing Python modules | Maximum code reuse |
|
||
| Task Queue | Celery + Redis | For background/async jobs |
|
||
| DevOps | Docker, GitHub Actions | Easy deployment, automation |
|
||
|
||
---
|
||
|
||
## 7. Next Steps
|
||
- Start with AI Writers and SEO Tools: wrap as FastAPI endpoints
|
||
- Gradually add Website Analyzer, Analytics, and Chatbot features
|
||
- Leave UI and Streamlit code aside; focus on modules that don’t depend on Streamlit
|
||
- Build React frontend to consume new API endpoints
|
||
|
||
---
|
||
|
||
## 8. Optional: Sample FastAPI Endpoint (for reference)
|
||
```python
|
||
from fastapi import FastAPI
|
||
from lib.ai_writers.blog_writer import generate_blog_post
|
||
|
||
app = FastAPI()
|
||
|
||
@app.post("/generate-blog/")
|
||
def generate_blog(data: BlogRequest):
|
||
return generate_blog_post(data.topic, data.keywords)
|
||
```
|
||
|
||
---
|
||
|
||
**This document should be updated as the migration progresses and new architectural decisions are made.** |