Pocket Sommelier is a wine recommendation web app built with FastAPI.
It uses TF-IDF vectorization of wine descriptions to suggest similar wines based on text similarity.
All model artifacts are precomputed offline and loaded at runtime for fast responses.

• 🔍 Content-based recommendations: Uses TF-IDF + cosine similarity between wine descriptions
• ⚡ FastAPI backend: Simple REST API for querying wine similarities
• 🧠 Precomputed artifacts: TF-IDF vectorizer, sparse matrix, and metadata CSV built via Jupyter Notebook
• 🐳 Dockerized environment: One-command setup for local development
• 🧩 Reproducible pipeline: Notebook for preprocessing and asset generation

1. Data Preparation:

   • Combine multiple open-source wine datasets
   • Clean and normalize descriptions
   • Build TF-IDF model and compute cosine similarity matrix

2. Model Artifacts:

   • tfidf_vectorizer.joblib – serialized scikit-learn vectorizer
   • tfidf_matrix.npz – sparse matrix representation
   • wines_data.csv – metadata (name, country, variety, etc.)

3. API Endpoints:

   Endpoint	Method	Description
   /by_description?description=...	GET	Returns top 5 similar wines to input text

These instructions assume you have Docker and docker-compose installed.

1. Start the API container:

   docker-compose up -d

2. Go to http://0.0.0.0:8000/

The datasets used in this project are publicly available and were sourced from the following repositories:

• Wine Ratings Dataset
• 21st Century Bordeaux Wine Dataset
• Wine Reviews Dataset

Parts of the user interface make use of ready-made CSS templates from uiverse.io.
These templates are licensed under the MIT License.

• CSS by 0xNihilism (0xNihilist) — © 2025 0xNihilism
• CSS by Vamsi Devendra Kumar — © 2025 Vamsi Devendra Kumar

Used with permission under the MIT License.