====== Product Sentiment Analysis App ======

This is a web application that searches X/Twitter for tweets about a product, analyzes their sentiment using Azure Cognitive Services and Azure OpenAI, and displays a summary.

===== Infrastructure =====

==== Required Services ====
  * **Azure OpenAI Service** – Used to summarize tweet sentiments via the Chat Completions API.
  * **Azure Text Analytics** – For classifying sentiment (positive, negative, neutral) of individual tweets via the Sentiment Analysis API.
  * **Twitter API v2** – To fetch tweets related to the searched keyword using the recent search endpoint.
  * **Azure App Service** – For hosting the backend using the FastAPI framework.

==== Environment Variables ====
  * `AZURE_OPENAI_KEY`
  * `AZURE_OPENAI_ENDPOINT`
  * `AZURE_OPENAI_DEPLOYMENT_NAME`
  * `AZURE_TEXT_ANALYTICS_KEY`
  * `AZURE_TEXT_ANALYTICS_ENDPOINT`
  * `X_BEARER_TOKEN`

Store them in a `.env` file or as App Settings in Azure App Service.

===== Backend Code (app.py) =====

==== Libraries Used ====
  * `fastapi` – Defines HTTP endpoints.
  * `requests` – Used to interact with the Twitter API.
  * `openai` – Azure OpenAI SDK for calling GPT models.
  * `azure.ai.textanalytics` – Azure SDK for Text Analytics API.
  * `jinja2` – Template rendering.
  * `uvicorn` – Development server for FastAPI.

==== Endpoints ====
  * `GET /` – Renders the form (HTML template).
  * `POST /analyze` – Handles form submission, performs:
    - Tweet search.
    - Sentiment classification.
    - Summary generation.

==== Key Functions ====

=== search_tweets(product_name) ===
  - Calls Twitter API v2:
    - Endpoint: `https://api.twitter.com/2/tweets/search/recent`
    - Auth: Bearer Token (`X_BEARER_TOKEN`)
    - Returns up to 10 recent English tweets about the keyword.

=== analyze_sentiment(texts) ===
  - Calls Azure Text Analytics API:
    - Endpoint: `<AZURE_TEXT_ANALYTICS_ENDPOINT>/text/analytics/v3.1/sentiment`
    - Auth: API Key (`AZURE_TEXT_ANALYTICS_KEY`)
    - Returns document sentiment (positive, negative, neutral) and confidence scores.

=== summarize_with_gpt(tweets) ===
  - Calls Azure OpenAI Chat Completions API:
    - Endpoint: `<AZURE_OPENAI_ENDPOINT>/openai/deployments/<DEPLOYMENT_NAME>/chat/completions`
    - Model: gpt-35-turbo or compatible
    - Prompts GPT to summarize public sentiment based on labeled tweet data.

===== Frontend Code (index.html) =====

Uses Bootstrap for styling. Features:
  * Form for entering a product.
  * Display area showing:
    - Sentiment-labeled tweets.
    - GPT-generated summary.

===== User Steps =====

==== 1. Configure Environment ====
Create `.env` with:
```
AZURE_OPENAI_KEY=your_key
AZURE_OPENAI_ENDPOINT=https://your-endpoint.openai.azure.com/
AZURE_OPENAI_DEPLOYMENT_NAME=your_deployment
AZURE_TEXT_ANALYTICS_KEY=your_text_analytics_key
AZURE_TEXT_ANALYTICS_ENDPOINT=https://your-text-analytics.cognitiveservices.azure.com/
TWITTER_BEARER_TOKEN=your_twitter_bearer_token
```

==== 2. Install Dependencies ====
Run:
  pip install -r requirements.txt

==== 3. Launch the App ====
Run:
  uvicorn app:app --reload

Then open your browser to:
  http://localhost:8000

==== 4. Use the App ====
  * Enter a product keyword (e.g., "AirPods").
  * Submit.
  * View tweet list with sentiment tags.
  * View summary generated by GPT.

===== File Structure =====

^ File ^ Purpose ^
| `app.py` | Main backend logic |
| `templates/index.html` | Frontend |
| `requirements.txt` | Dependencies |

===== How It Works =====

  - User inputs a keyword.
  - Tweets are fetched using Twitter API v2.
  - Sentiment analysis is performed using Azure Text Analytics.
  - Tweets are summarized using Azure OpenAI GPT.
  - Result is rendered via HTML template.