Update app.py

app.py

@@ -1,64 +1,122 @@
 import gradio as gr
+import random
+import os
 from huggingface_hub import InferenceClient
+from sentence_transformers import SentenceTransformer
+import torch
 
-"""
-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")
+with open("knowledge.txt", "r", encoding="utf-8") as file:
+  recent = file.read()
+  # opens the text, saves as "file"
+  # reads the text and saves as water_cycle_text variable
 
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    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]})
+cleaned_text = recent.strip()
+# cleaning up the text
+chunks = cleaned_text.split("\n")
+# seperating the text into one sentence pieces
+cleaned_chunks = []
+# creating an empty list to put the cleaned chunks in
 
+for chunk in chunks:
+  stripped_chunk = chunk.strip()
+  if stripped_chunk:
+    cleaned_chunks.append(stripped_chunk)
+    # loop through chunks and add not empty chunks to cleaned_chunks list
+
+
+model = SentenceTransformer('all-MiniLM-L6-v2')
+
+chunk_embeddings = model.encode(cleaned_chunks, convert_to_tensor=True)
+# encode the model, pass through my cleaned chunks and convert to vector embeddings (not arrays)
+
+
+def get_top_chunks(query):
+  # create my function taking query as parameter
+  query_embedding = model.encode(query, convert_to_tensor=True)
+  # encode query to vector embedding for comparison
+  query_embedding_normalized = query_embedding / query_embedding.norm()
+  # normalize my query to 1; allows for comparison of meaning
+  chunk_embeddings_normalized = chunk_embeddings / chunk_embeddings.norm(dim=1, keepdim=True)
+  # normailizing chunks for comparison of meaning
+
+  similarities = torch.matmul(chunk_embeddings_normalized, query_embedding_normalized)
+
+  # using matmul (matrix multiplication method) to compare query to chunks
+  top_indices = torch.topk(similarities, k=3).indices
+
+  # get the indices of the chunks that are most similar to query
+
+  top_chunks = []
+
+  for i in top_indices:
+    chunk = chunks[i]
+    # for each index number in top_indices, get back the text
+    top_chunks.append(chunk)
+    # values of each index number is added to top_chunks
+  return top_chunks
+    
+client = InferenceClient(
+    model='Qwen/Qwen2.5-72B-Instruct',
+    #token = 'HF_TOKEN'
+) 
+
+#client is where you can change the LLM model!
+def respond(message,history):
+    if not message.strip():
+        return "Hello!"
+    gift_ideas = get_top_chunks(message)
+    messages = [{'role': 'system', 'content': f'You give really good gift ideas and are super helpful! You also tell me the price of each item. Give me 5 gift ideas if I ask. Use the following database for gift ideas: {gift_ideas}'}]
+    
+    if history:
+        messages.extend(history)
+    
     messages.append({"role": "user", "content": message})
 
     response = ""
-
     for message in client.chat_completion(
+        #max_tokens controls how many words your responses is 
         messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
+        max_tokens = 500,
+        stream = True,
+        
+        #temperature = 0.8, #code a decimal between 0-2
+        #top_p = .65 #code a decimal between 0-1
     ):
         token = message.choices[0].delta.content
-
         response += token
         yield response
+    #print(response["choices"][0]["message"]["content"].strip())
+    
+    #yield response["choices"][0]["message"]["content"].strip()
+    
+#with gr.Blocks(theme='hmb/amethyst') as demo:
+ #   with gr.Row(equal_height=True):
+  #      with gr.Column(scale=10):
+   #            """
+    #            # 🎁 Introducing WrapIT!
+     #           **WrapIT** helps users find personalized gift ideas and craft thoughtful card messages  
+      #          by inputting details like the recipient's interests, celebration type, and budget.
+       #         
+        #        ✨ *All you have to do is wrap it.*
+         #       """
+          #  )
+           # gr.ChatInterface(respond, type='messages')
 
+  
+    
+    #chatbot = gr.Chatbot()
+    #msg = gr.Textbox(placeholder="Say hi to WrapIT here!", label="Message")
+    #send = gr.Button("Send")
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    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"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
-
-
-if __name__ == "__main__":
-    demo.launch()
+    #msg.submit(respond, [msg, chatbot], [msg, chatbot])
+    #send.click(respond, [msg, chatbot], [msg, chatbot])
+suggestions = ["Hello", "How are you?"]
+chat_interface = gr.ChatInterface(
+    fn=respond, 
+    title="🎁 Introducing WrapIT!", 
+    theme='hmb/amethyst', 
+    description=" **WrapIT** helps users find personalized gift ideas and craft thoughtful card messages by inputting details like the recipient's interests, celebration type, and budget ✨ *All you have to do is wrap it.*")
+    examples=suggestions
+chat_interface.launch()