Patent Search Engine

---

Description

This is a patent search engine that uses machine learning models and Elasticsearch to retrieve relevant patent documents based on user queries. The search engine retrieves documents using two methods: semantic search using a BERT-based model and keyword search using Elasticsearch.

Requirements

• Python 3.6 or higher
• Elasticsearch
• Flask
• SentenceTransformers
• Faiss
• Pandas
• Flask-Paginate

Installation

1. Install Python 3.6 or higher on your machine.[optional]
2. Install Elasticsearch by following the instructions in the official documentation: https://www.elastic.co/guide/en/elasticsearch/reference/current/install-elasticsearch.html.[optional]
3. Run Elasticsearch by navigating to the Elasticsearch directory and running .\bin\elasticsearch.bat on Windows or ./bin/elasticsearch on Unix/Linux.[optional]
4. Clone this repository using $git clone https://github.com/AnishmMore/patent_search.git.
5. Navigate to the cloned repository by running $cd patent_search .
6. Install the required Python packages by running $pip install --ignore-installed -r requirements.txt OR simply run the run.sh script [windows]: $bash .\run.sh OR [unix/linux]: $.\run.sh.

Usage

1. Run the Python Flask application by running python3 app.py.
2. Open a web browser and go to http://127.0.0.1:8080 to access the search engine.
3. Enter a query in the search box and click "Search" to retrieve relevant patent documents.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Credits

• The machine learning models used in this project were developed by the SentenceTransformers and Faiss libraries.
• The Flask web framework was used to develop the web application.
• The Elasticsearch library was used to index and search patent documents.
• The Flask-Paginate library was used to implement pagination in the search results.

Team Members and Contributors

• Anish More
• Akash Bilgi
• Amrutha Alewoor
• Mrunali Lachake
• Kaushik Daiv