Upload 3 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Code[[:space:]]Civil[[:space:]]vectorised.json filter=lfs diff=lfs merge=lfs -text

Code Civil vectorised.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b1bd358abd17993d9c49fd622ce5e353f35eb1d815d87ead88c914e1db18041
+size 47281293

app.py

@@ -1,12 +1,43 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
+import json
+import numpy as np
+from sentence_transformers import SentenceTransformer
+from sklearn.metrics.pairwise import cosine_similarity
 
 """
 For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
 """
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+# Load embeddings from a JSON file
+def load_embeddings(file_path):
+    with open(file_path, 'r', encoding='utf-8') as file:
+        return json.load(file)
 
+# Function to get relevant articles based on user query
+def get_relevant_documents(query, embeddings_data, model, top_k=3):
+    query_embedding = model.encode(query)
+    similarities = []
+
+    for entry in embeddings_data:
+        embedding = np.array(entry['embedding'])
+        similarity = cosine_similarity([query_embedding], [embedding])[0][0]
+        similarities.append((entry, similarity))
+
+    # Sort by similarity and return top_k relevant entries
+    similarities.sort(key=lambda x: x[1], reverse=True)
+    top_entries = [entry for entry, _ in similarities[:top_k]]
+    
+    return top_entries
+
+# Function to format relevant documents into a string
+def format_documents(documents):
+    formatted = ""
+    for doc in documents:
+        formatted += f"Relevant article: {doc['name']}\n{doc['content']}\n\n"
+    return formatted
+
+# Main chatbot function that integrates RAG
 def respond(
     message,
     history: list[tuple[str, str]],
@@ -14,8 +45,17 @@ def respond(
     max_tokens,
     temperature,
     top_p,
+    embeddings_data,
+    model
 ):
-    messages = [{"role": "system", "content": system_message}]
+    # Search for relevant documents based on user input
+    relevant_docs = get_relevant_documents(message, embeddings_data, model)
+    retrieved_context = format_documents(relevant_docs)
+    
+    # Add the retrieved context as part of the system message
+    system_message_with_context = system_message + "\n\n" + "Relevant documents:\n" + retrieved_context
+
+    messages = [{"role": "system", "content": system_message_with_context}]
 
     for val in history:
         if val[0]:
@@ -26,7 +66,8 @@ def respond(
     messages.append({"role": "user", "content": message})
 
     response = ""
-
+    client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+    
     for message in client.chat_completion(
         messages,
         max_tokens=max_tokens,
@@ -35,15 +76,19 @@ def respond(
         top_p=top_p,
     ):
         token = message.choices[0].delta.content
-
         response += token
         yield response
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+# Load embeddings and model once at startup
+embeddings_file = 'Code Civil vectorised.json'
+embeddings_data = load_embeddings(embeddings_file)
+embedding_model = SentenceTransformer('Lajavaness/bilingual-embedding-small', trust_remote_code=True)
+
+# Gradio interface
 demo = gr.ChatInterface(
-    respond,
+    lambda message, history, system_message, max_tokens, temperature, top_p: respond(
+        message, history, system_message, max_tokens, temperature, top_p, embeddings_data, embedding_model
+    ),
     additional_inputs=[
         gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
@@ -58,6 +103,5 @@ demo = gr.ChatInterface(
     ],
 )
 
-
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()

requirements.txt

@@ -1 +1,6 @@
-huggingface_hub==0.22.2
+huggingface_hub==0.22.2
+gradio==3.25.0
+huggingface_hub==0.22.2
+sentence-transformers==2.2.2
+scikit-learn==1.3.0
+numpy==1.24.2