detoX API

This API serves as the backend for the detoX browser extension designed to address hate speech detection with a particular focus on election-related and politics-related content in the Philippines. It employs a fine-tuned RoBERTA Tagalog Base model, a variant of the BERT (Bidirectional Encoder Representations from Transformers) model known for its exceptional performance in NLP tasks.

The model was trained using a combined dataset from two sources. The first dataset was a pre-labeled Filipino hate speech dataset. The second dataset consisted of newly crawled 2022 Philippine Presidential Elections-related Tweets.

🌐 Relevant Links

🗂️ Datasets

• Hate Speech Filipino: hate_speech_filipino
• Combined Dataset: mapsoriano/2016_2022_hate_speech_filipino

🤖 Models

• RoBERTA Tagalog Base Model: jcblaise/roberta-tagalog-base
• Fine-tuned Model: mapsoriano/roberta-tagalog-base-philippine-elections-2016-2022-hate-speech

🧰 Browser Extension

• Development repo: alliyah95/detoX
• Live: detoX

💻 Local Setup and Installation

1. Download or clone the repository to your local machine using the following command:

   git clone https://github.com/alliyah95/detoX-api.git
   

2. Navigate to the project directory:

   cd detoX-api
   

3. Install the project dependencies:

   pip install -r requirements.txt
   

4. Run the app by executing the following command:

   uvicorn main:app --reload
   

5. To test the API endpoints with a GUI, navigate to the following URL in your browser. The default port is 8000.

   http://127.0.0.1:<PORT>/docs#/
   

🔍 Endpoints

/ (Root endpoint)

• Used for checking whether the server is running.

  Response:

  {
      "1": "Server is up and running"
  }

/api/v1/detect

• Returns the result of the hate speech detection.

  Parameter:

  ?content - the text content used for hate speech classification.

  Example:

  /api/v1/detect?content=hahaha%20basta%20pulangaw%20bobo
  

  Response:

  {
      "result": 1
  }

🛠️ Built With

💙 Acknowledgements

Name	Contribution
🌟 Dr. Mary Jane Rabena	Our Thesis Adviser
🌟 Dr. Arlan Dela Cruz	Our Thesis Co-Adviser
🌟 Ms. Abijah Louise Dela Cruz	Dataset Validator
🌟 Dr. Alma Theresa Manaloto	Expert Evaluator
🌟 Dr. Ocirne Jun-Jun Liwanag	Expert Evaluator
🌟 Mr. Arvin del Rosario	Expert Evaluator
🌟 Mr. John Montes	Expert Evaluator
🌟 Mr. Nelson Dizon	Expert Evaluator
🌟 Mr. Alberto Castro Jr.	Expert Evaluator
🌟 Mr. Blaise Cruz	Publisher of the RoBERTa Tagalog Base model and the Hate Speech Filipino Dataset

🧠 Authors

• 👧 Danica L. Castro
• 👧 Lenina Jemima V. Dizon
• 👧 Alliyah Joyce M. Sarip
• 👦 Mark Aaron P. Soriano