Backend

AI-powered fashion + fragrance recommendation engine built using:

• Flask (REST API)
• Gemini multimodal embeddings (placeholder integration in code)
• FAISS vector similarity search
• Firestore (item metadata + embeddings)
• Google Cloud Storage (image storage)
• Firebase Auth (optional user authentication)

This backend powers the Weafore personalization engine.
Frontend, infra, and datasets are not included here by design.

---

🚀 Features

✔ Recommendation API

• Accepts image + mood + fragrance notes
• Generates multimodal embeddings (placeholder — plug in Gemini)
• Searches FAISS vector store for similar items
• Fetches metadata from Firestore
• Reranks using scoring logic (mood + fragrance compatibility)
• Generates short explanations using Gemini text generation

✔ GCS Integration

• Uploads user images to Google Cloud Storage
• Returns public URL

✔ Firestore Integration

• Stores outfit/fragrance metadata
• Reads embeddings for similarity search

✔ Firebase Authentication (Optional)

• Validate Bearer tokens
• Secure your recommendation pipeline

---