📋 TidBit (Backend)

Backend API for TidBit

TidBit is a document summarization application, currently with a Chrome extension front-end. When you hit the TidBit button on the Chrome extension, it will try to summarize and get the gist out of the webpage.

Technologies

• 🐍 Python - Python 3 came out with asynchronous IO features, which will allow us to take greater advantage of concurrency and serve more requests for a given server.
• 🐋 Docker - Containerization and easy development/production environment bootstrapping
• 🐇 RabbitMQ - Message Queue that allows us to send messages between our backend and microservices, all with complex routing
• 📡 Postgres - Relational database of choice
• 📈 Neo4J - Graph-based database
• 📁 HDFS - Distributed Filesystem

Getting Started

Make sure to have Docker and Docker-Compose installed. Then, follow through the following steps:

1. cd to the base of the project directory
2. Rename backend_sample.env and relational_db_sample.env to backend.env and relational_db.env
3. Fill in the appropriate information in both backend.env and relational_db.env, making sure to specify all values accordingly
4. Run docker-compose build
5. Run docker-compose up and now, you should be able to access the API routes at the localhost or whatever hostname your copy of Docker is working on!
6. To shut containers down, run docker-compose down

When you update the source code, you will have to run docker-compose down followed by docker-compose up again. These commands must always be executed at the base of the project directory!

API Routes

The API routes are accessible under the hostname/api/<current-version> URL namespace. Output is in JSON format.

An access token can be obtained by loggin in and can be used on a request by specifying the Authorization header with the value: Bearer <token>.

Responses have the structure {status, message, errors}, where the errors prop is provided whenever the status denotes an HTTP error status.

• /user
  • POST / - Create a new user
    • Request: {firstName, lastName, username, password}
  • GET /{userId} - Get information about the user
    • With appropriate token:
      • Response: {firstName, lastName, username, joinedAt, lastLoginAt, history: [{url, accessedAt, summarizer_type}]}
    • Without token:
      • Response: {firstName, lastName, username, joinedAt, lastLoginAt}
  • POST /{userId} - Update information about the user
    • Requires appropriate token
    • Request: {firstName, lastName, username, password}
  • DELETE /{userId} - Delete the user
    • Requires appropriate token
  • POST /{user_id}/summary - Get a summary for the provided webpage contents
    • Request: {url, domContent, summarizerType}
    • Response: {summary, summarizerType}
    • Requires appropriate token
  • GET /{user_id}/history - Get a history of the documents this user has summarized
    • Request: None
    • Response: {history: [{url, accessedAt, summarizer_type}]}
    • Requires appropriate token
• /authenticate
  • POST / - Login with provided credentials and obtain a token
    • Request: {username, password}
    • Response: {token}
• /summary
  • POST / - Get a summary of the page content
    • Request: {url, domContent:optional, summarizerType}
  • GET /types - Returns a list of the available summarization algorithms.
    • Request: None
    • Response: {summarizerTypes}

Relational Schema

The following entities exist in the database and represented in a normalized and relational fashion into a relational database.

• user
  • id - uuid
  • first_name - text
  • last_name - text
  • username - varchar(30)
  • password - varchar(100)
  • joined_at - timestamptz
  • last_login_at - timestamptz
• history
  • id - uuid
  • user_id - foreign(user, id)
  • document_id - foreign(document, id)
  • summarizer_type - text
  • accessed_at - timestamptz
• document
  • id - int
  • user_id - foreign(user, id)
  • url - text
  • contents - text
  • summarized_at - timestamptz