MuziSitsha/mzansi-news-summarizer

A deep learning capstone project built with PyTorch, Hugging Face Transformers, and Gradio. This application automatically summarizes South African news articles and performs sentiment analysis to reveal public mood and emotional tone. # Mzansi News Summarizer South Africa-focused news summarizer built with Streamlit, Hugging Face models, and RSS aggregation. Created by **Muziwakhe Sitsha**. ## Features - Summarize pasted article text or an article URL - Sentiment analysis with confidence score - SA-focused tags and category signals - "Breaking Mzansi" RSS feed browser - "What's Hot in SA" trends dashboard - Summary language selection across SA language options ## Requirements - Python 3.10+ - `pip` ## Local Setup ```bash python -m venv venv venv\Scripts\activate pip install -r requirements.txt ``` ## Run Locally ```bash streamlit run app.py ``` If `streamlit` is not on your PATH: ```bash venv\Scripts\python.exe -m streamlit run app.py ``` ## Streamlit Cloud Deployment 1. Push this repo to GitHub. 2. In Streamlit Community Cloud, click **New app** and select this repo. 3. Set: - **Main file path**: `app.py` - **Python version**: 3.10 or 3.11 4. Add secrets in **App Settings -> Secrets**. 5. Deploy. ### Recommended Secrets (Streamlit Cloud) ```toml HF_API_TOKEN = "your_hf_token" HUGGINGFACEHUB_API_TOKEN = "your_hf_token" HF_TOKEN = "your_hf_token" SUMMARY_PROVIDER = "hf_api" SENTIMENT_PROVIDER = "hf_api" ``` Optional provider keys: ```toml OPENAI_API_KEY = "your_openai_key" COHERE_API_KEY = "your_cohere_key" ``` ## App Configuration - Streamlit config is in `.streamlit/config.toml`. - Dependencies are pinned in `requirements.txt`. ## Troubleshooting - If deployment fails, open Streamlit Cloud logs and check missing package/token errors. - If RSS feels slow on first load, wait for cache warm-up; subsequent loads are faster. - If translation is unavailable, verify your app has network access and valid provider token secrets.