Update app.py

app.py

@@ -1,15 +1,14 @@
 import os
 import gradio as gr
-import google.generativeai as genai
-from typing import List, Tuple
+from google.generativeai import GenerativeModel, configure, types
 import fitz  # PyMuPDF
 from sentence_transformers import SentenceTransformer
 import numpy as np
 import faiss
 
-# Initialize Google API Key
+# Set up the Google API for the Gemini model
 GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
-genai.configure(api_key=GOOGLE_API_KEY)
+configure(api_key=GOOGLE_API_KEY)
 
 # Placeholder for the app's state
 class MyApp:
@@ -31,55 +30,105 @@ class MyApp:
         print("PDF processed successfully!")
 
     def build_vector_db(self) -> None:
-        """Builds a vector database using FAISS and SentenceTransformer embeddings."""
-        model = SentenceTransformer("all-MiniLM-L6-v2")
-        embeddings = model.encode([doc["content"] for doc in self.documents])
-        self.embeddings = np.array(embeddings, dtype="float32")
+        """Builds a vector database using the content of the PDF."""
+        model = SentenceTransformer('all-MiniLM-L6-v2')
+        self.embeddings = model.encode([doc["content"] for doc in self.documents], show_progress_bar=True)
         self.index = faiss.IndexFlatL2(self.embeddings.shape[1])
-        self.index.add(self.embeddings)
+        self.index.add(np.array(self.embeddings))
         print("Vector database built successfully!")
 
-    def search(self, query: str, top_k: int = 5) -> List[Tuple[int, str]]:
-        """Searches for the most similar documents based on the query."""
-        query_embedding = SentenceTransformer("all-MiniLM-L6-v2").encode([query])
-        distances, indices = self.index.search(np.array(query_embedding, dtype="float32"), top_k)
-        return [(self.documents[idx]["page"], self.documents[idx]["content"]) for idx in indices[0]]
+    def search_documents(self, query: str, k: int = 3) -> List[str]:
+        """Searches for relevant documents using vector similarity."""
+        model = SentenceTransformer('all-MiniLM-L6-v2')
+        query_embedding = model.encode([query], show_progress_bar=False)
+        D, I = self.index.search(np.array(query_embedding), k)
+        results = [self.documents[i]["content"] for i in I[0]]
+        return results if results else ["No relevant documents found."]
 
-    def generate_response(self, query: str) -> str:
-        """Generates a response using the Gemini model based on the query."""
-        if not GOOGLE_API_KEY:
-            raise ValueError("GOOGLE_API_KEY is not set. Please set it up.")
+app = MyApp()
 
-        generation_config = genai.types.GenerationConfig(
-            temperature=0.7,
-            max_output_tokens=512
+def respond(message: str, history: List[Tuple[str, str]]):
+    system_message = "You are a supportive and empathetic Dialectical Behaviour Therapist assistant. You politely guide users through DBT exercises based on the given DBT book. You must say one thing at a time and ask follow-up questions to continue the chat."
+    messages = [{"role": "system", "content": system_message}]
+
+    for val in history:
+        if val[0]:
+            messages.append({"role": "user", "content": val[0]})
+        if val[1]:
+            messages.append({"role": "assistant", "content": val[1]})
+
+    messages.append({"role": "user", "content": message})
+
+    # RAG - Retrieve relevant documents if the query suggests exercises or specific information
+    if any(keyword in message.lower() for keyword in ["exercise", "technique", "information", "guide", "help", "how to"]):
+        retrieved_docs = app.search_documents(message)
+        context = "\n".join(retrieved_docs)
+        if context.strip():
+            messages.append({"role": "system", "content": "Relevant documents: " + context})
+
+    model = GenerativeModel("gemini-1.5-pro-latest")
+    generation_config = types.GenerationConfig(
+        temperature=0.7,
+        max_output_tokens=1024
+    )
+    response = model.generate_content([message], generation_config=generation_config)
+
+    response_content = response[0].text if response else "No response generated."
+    history.append((message, response_content))
+    return history, ""
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🧘‍♀️ **Dialectical Behaviour Therapy**")
+    gr.Markdown(
+        "‼️Disclaimer: This chatbot is based on a DBT exercise book that is publicly available. "
+        "We are not medical practitioners, and the use of this chatbot is at your own responsibility."
+    )
+
+    chatbot = gr.Chatbot()
+
+    with gr.Row():
+        txt_input = gr.Textbox(
+            show_label=False,
+            placeholder="Type your message here...",
+            lines=1
         )
+        submit_btn = gr.Button("Submit", scale=1)
+        refresh_btn = gr.Button("Refresh Chat", scale=1, variant="secondary")
+
+    example_questions = [
+        ["What are some techniques to handle distressing situations?"],
+        ["How does DBT help with emotional regulation?"],
+        ["Can you give me an example of an interpersonal effectiveness skill?"],
+        ["I want to practice mindfulness. Can you help me?"],
+        ["I want to practice distraction techniques. What can I do?"],
+        ["How do I plan self-accommodation?"],
+        ["What are some distress tolerance skills?"],
+        ["Can you help me with emotional regulation techniques?"],
+        ["How can I improve my interpersonal effectiveness?"],
+        ["What are some ways to cope with stress using DBT?"],
+        ["Can you guide me through a grounding exercise?"],
+        ["How do I use DBT skills to handle intense emotions?"],
+        ["What are some self-soothing techniques I can practice?"],
+        ["How can I create a sensory-friendly safe space?"],
+        ["Can you help me create a personal crisis plan?"],
+        ["What are some affirmations for neurodivergent individuals?"],
+        ["How can I manage rejection sensitive dysphoria?"],
+        ["Can you guide me through observing with my senses?"],
+        ["What are some accessible mindfulness exercises?"],
+        ["How do I engage my wise mind?"],
+        ["What are some values that I can identify with?"],
+        ["How can I practice mindful appreciation?"],
+        ["What is the STOP skill in distress tolerance?"],
+        ["How can I use the TIPP skill to manage distress?"],
+        ["What are some tips for managing meltdowns?"],
+        ["Can you provide a list of stims that I can use?"],
+        ["How do I improve my environment to reduce distress?"]
+    ]
+
+    gr.Examples(examples=example_questions, inputs=[txt_input])
 
-        model_name = "gemini-1.5-pro-latest"
-        model = genai.GenerativeModel(model_name)
-        response = model.generate_content([query], generation_config=generation_config)
-        
-        return response[0].text if response else "No response generated."
-
-# Gradio UI setup for interaction
-def main():
-    app = MyApp()
-
-    def handle_query(query):
-        search_results = app.search(query)
-        response = app.generate_response(query)
-        return {"Search Results": search_results, "Response": response}
-
-    gr.Interface(
-        fn=handle_query,
-        inputs=gr.Textbox(placeholder="Enter your query here"),
-        outputs=[
-            gr.JSON(label="Search Results"),
-            gr.Textbox(label="Generated Response")
-        ],
-        title="Dialectical Behavioral Exercise with Gemini",
-        description="This app uses Google Gemini to generate responses based on document content."
-    ).launch()
+    submit_btn.click(fn=respond, inputs=[txt_input, chatbot], outputs=[chatbot, txt_input])
+    refresh_btn.click(lambda: [], None, chatbot)
 
 if __name__ == "__main__":
-    main()
+    demo.launch()