Sentiment API

A Rust-based API for Vercel Serverless Functions that provides sentiment analysis based on weighted post data.

Endpoints

1. GET /api/topics

Returns all available topics for a given category.

Query Parameters:

• category (required): The category to filter topics by

Example:

GET /api/topics?category=politics

Response:

[
  {
    "id": 1,
    "name": "2024 Election",
    "category": "politics"
  },
  {
    "id": 2,
    "name": "Climate Policy",
    "category": "politics"
  }
]

2. GET /api/posts

Returns top posts for a topic, sorted by weight * relevancy.

Query Parameters:

• topic_id (required): The ID of the topic
• limit (optional): Maximum number of posts to return (default: 10)

Example:

GET /api/posts?topic_id=1&limit=5

Response:

[
  {
    "id": 1,
    "topic_id": 1,
    "content": "Post content here...",
    "weight": 0.9,
    "relevancy": 0.85,
    "score": 0.765
  }
]

3. GET /api/sentiment

Returns normalized sentiment probabilities for all outcomes of a topic.

The sentiment is calculated by:

1. For each post and outcome: weight * relevancy * probability
2. Sum all scores for each outcome
3. Normalize probabilities so they sum to 1.0

Query Parameters:

• topic_id (required): The ID of the topic

Example:

GET /api/sentiment?topic_id=1

Response:

{
  "topic_id": 1,
  "topic_name": "2024 Election",
  "sentiments": [
    {
      "outcome_id": 1,
      "outcome_name": "Candidate A Wins",
      "probability": 0.55
    },
    {
      "outcome_id": 2,
      "outcome_name": "Candidate B Wins",
      "probability": 0.45
    }
  ]
}

Environment Variables

Create a .env file with the following variables:

DATABASE_URL=postgres://user:password@host:port/database

Database Schema

The API expects the following PostgreSQL tables:

CREATE TABLE topics (
    id BIGSERIAL PRIMARY KEY,
    name VARCHAR(255) NOT NULL,
    category VARCHAR(255) NOT NULL
);

CREATE TABLE posts (
    id BIGSERIAL PRIMARY KEY,
    topic_id BIGINT NOT NULL REFERENCES topics(id),
    content TEXT NOT NULL,
    weight DOUBLE PRECISION NOT NULL,
    relevancy DOUBLE PRECISION NOT NULL
);

CREATE TABLE outcomes (
    id BIGSERIAL PRIMARY KEY,
    topic_id BIGINT NOT NULL REFERENCES topics(id),
    name VARCHAR(255) NOT NULL
);

CREATE TABLE post_outcome_probabilities (
    id BIGSERIAL PRIMARY KEY,
    post_id BIGINT NOT NULL REFERENCES posts(id),
    outcome_id BIGINT NOT NULL REFERENCES outcomes(id),
    probability DOUBLE PRECISION NOT NULL
);

CREATE INDEX idx_topics_category ON topics(category);
CREATE INDEX idx_posts_topic_id ON posts(topic_id);
CREATE INDEX idx_outcomes_topic_id ON outcomes(topic_id);
CREATE INDEX idx_post_outcome_probabilities_post_id ON post_outcome_probabilities(post_id);

Development

Prerequisites

• Rust (latest stable)
• PostgreSQL database

Local Development

1. Install dependencies:

   cargo build

2. Set up your .env file with database credentials

3. Run locally with Vercel CLI:

   vercel dev

Deployment

Deploy to Vercel:

vercel

Make sure to set the DATABASE_URL environment variable in your Vercel project settings.

Project Structure

sentiment-api/
├── api/
│   ├── topics.rs      # Topics endpoint handler
│   ├── posts.rs       # Posts endpoint handler
│   └── sentiment.rs   # Sentiment endpoint handler
├── src/
│   └── lib.rs         # Shared database models and queries
├── Cargo.toml         # Rust dependencies
├── vercel.json        # Vercel configuration
└── README.md