Initial clean commit with hosted model

Dockerfile

@@ -1,21 +1,17 @@
-# Base image
 FROM python:3.10-slim
 
 # Set working directory
 WORKDIR /app
 
-# Install system dependencies
-RUN apt-get update && apt-get install -y git
-
 # Copy files
 COPY . .
 
-# Install Python dependencies
-RUN pip install --upgrade pip
-RUN pip install -r requirements.txt
+# Install dependencies
+RUN pip install --upgrade pip && \
+    pip install -r requirements.txt
 
-# Expose Streamlit default port
-EXPOSE 8501
+# Expose Streamlit port
+EXPOSE 7860
 
 # Run the app
-CMD ["streamlit", "run", "app.py", "--server.port=8501", "--server.enableCORS=false"]
+CMD ["streamlit", "run", "app.py", "--server.port=7860", "--server.address=0.0.0.0"]

README.md

@@ -1,122 +1,93 @@
----
-title: Counselor Assistant
-emoji: 👀
-colorFrom: red
-colorTo: blue
-sdk: streamlit
-sdk_version: 1.44.1
-app_file: app.py
-pinned: false
-license: mit
----
 # 🧠 Mental Health Counselor Assistant
 
-This project is a **counselor-facing AI assistant** designed to support mental health professionals by offering helpful suggestions in response to patient messages. It detects crisis and violent messages, generates supportive or informative suggestions, and provides logging and export features.
+This app helps mental health counselors generate helpful, informed, and compassionate suggestions based on patient messages. It classifies the message intent and offers guidance for how a counselor might respond, with built-in safety checks for crisis or violent content.
 
 ---
 
-## 📘 How to use this app
+## 💡 Features
 
-1. Enter a message from a patient in the text box.
-2. The app will generate helpful suggestions for how the counselor might reply.
-3. If a safety risk is detected (e.g., crisis or violence), a safe response will be shown instead.
-4. You can save the conversation history as a CSV file.
+- ✅ Response type classification (advice, validation, information, question)
+- 💬 LLM-generated counselor suggestions using `google/flan-t5-xl`
+- 🛡️ Built-in safety detection for crisis and toxic content
+- 💾 Conversation history logging and export
+- 🌐 Web-based Streamlit interface
 
 ---
 
-## 🧩 Key Features
-
-- ✅ Fine-tuned classifier (DistilBERT) for predicting intent (advice, question, validation, information)
-- 🔁 Streaming text generation using Flan-T5-XL
-- 🔍 Safety detection for crisis, violence, and toxicity
-- 💡 Suggestion prompts tailored to each response type
-- 💾 Option to save conversation logs
-- 📸 Screenshots for walkthrough
-
----
+## 📦 Hosted Model
 
-## 🧠 Model Architecture
+This app uses a fine-tuned classification model hosted at:
 
-- **Response Type Classifier**: `DistilBERT` fine-tuned on a labeled response dataset.
-- **Text Generator**: `google/flan-t5-xl` used to generate helpful, human-like counselor suggestions.
-- **Safety Filter**: Includes custom regex patterns and `Detoxify` toxicity detector.
+➡️ [`scdong/distilbert-response-type`](https://huggingface.co/scdong/distilbert-response-type)
 
 ---
 
-## 🧪 Datasets Used
+## 🚀 How to Use
 
-- `counselchat-data.csv` – Labeled examples of counselor replies
-- `pair_data.csv` – Counselor-patient message pairs
-- `Kaggle_Mental_Health_Conversations_train.csv`
-- `distilbert_labeled_responses.csv` – Augmented training set
-- `cleaned_combined_mental_health_dataset.csv`
+1. Enter a patient's message in the chat box.
+2. The app detects safety concerns (e.g. suicidal, violent, or toxic content).
+3. If safe, it classifies the message and generates counselor-style suggestions.
+4. You can export the session log as a CSV file.
 
 ---
 
-## 📁 Project Structure
+## 🖥️ Running Locally
 
 ```bash
-counselor-assistant/
-├── app.py                     # Main Streamlit app
-├── Dockerfile                 # Containerization for deployment
-├── requirements.txt           # Python dependencies
-├── pedal.yaml                 # Optional: Pedal config
-├── datasets/                  # Training + testing datasets
-├── models/                    # Fine-tuned model + tokenizer
-├── notebooks/                 # Data prep + training notebooks
-├── logs/                      # Saved chat logs
-├── screenshots/               # App walkthrough examples
-├── cache/                     # Prompt-response cache
-└── alert/                     # Alert logs (for flagged inputs)
+# Install dependencies
+pip install -r requirements.txt
+
+# Run the app
+streamlit run app.py
 ```
 
 ---
 
-## 🐳 Docker Deployment
-
-To build and run the app locally:
+## 🐳 Docker (optional)
 
 ```bash
+# Build Docker image
 docker build -t counselor-assistant .
+
+# Run it
 docker run -p 7860:7860 counselor-assistant
 ```
 
-Then navigate to `http://localhost:7860` in your browser.
+Then go to `http://localhost:7860` in your browser.
 
 ---
 
-## ☁️ Hugging Face Deployment
-
-You can deploy to [Hugging Face Spaces](https://huggingface.co/spaces/) by:
+## 🧠 Example Use Case
 
-1. Adding this repo to a Space with SDK = `streamlit`
-2. Uploading the model folder to the Space (or hosting it via Hugging Face Hub)
-3. Pushing your app via `git push`
+> **Patient:** I feel like I’m not good enough, no matter how hard I try.  
+> **Suggestions:**
+> - You're doing more than enough — even if it doesn't feel like it right now.  
+> - Try to be as kind to yourself as you would to a close friend.  
+> - Even small steps forward count. You're showing up, and that matters.
 
 ---
 
-## 📸 Screenshots
-
-### 💬 Classifier Prediction + Suggestions
-![example1](screenshots/example1.png)
-
-### 🚨 Safe Response for Crisis Input
-![example2](screenshots/example2.png)
+## 🛑 Crisis Handling
 
-### 🧠 Multiple Suggestions
-![example3](screenshots/example3.png)
+If the input includes high-risk language (e.g., "I want to kill myself"), the app immediately responds with:
 
-### 💾 Logging the Conversation
-![example4](screenshots/example4.png)
+> _"I'm really sorry you're feeling this way. You're not alone. Please consider calling 988 or speaking with a mental health professional right away."_
 
 ---
 
-## 📦 Requirements
-
-Install dependencies:
+## 📁 File Structure
 
 ```bash
-pip install -r requirements.txt
+.
+├── app.py              # Main Streamlit app
+├── requirements.txt    # Python dependencies
+├── Dockerfile          # (Optional) Docker setup
+├── cache.json          # Prompt/response cache
+├── README.md           # This file
 ```
 
+---
+
+## 🤝 License
 
+MIT License.