main__3 notebook should be fully transfered over

app.py

@@ -9,7 +9,8 @@ from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStream
 from llama_index.core.prompts.prompts import SimpleInputPrompt
 from llama_index.llms.huggingface import HuggingFaceLLM
 from llama_index.legacy.embeddings.langchain import LangchainEmbedding
-from langchain.embeddings.huggingface import HuggingFaceEmbeddings # This import should now work
+#from langchain.embeddings.huggingface import HuggingFaceEmbeddings # This import should now work
+from langchain_huggingface import HuggingFaceEmbeddings
 from sentence_transformers import SentenceTransformer
 
 from llama_index.core import set_global_service_context, ServiceContext
@@ -38,6 +39,59 @@ As a derivate work of [Llama-2-7b-chat](https://huggingface.co/meta-llama/Llama-
 this demo is governed by the original [license](https://huggingface.co/spaces/huggingface-projects/llama-2-7b-chat/blob/main/LICENSE.txt) and [acceptable use policy](https://huggingface.co/spaces/huggingface-projects/llama-2-7b-chat/blob/main/USE_POLICY.md).
 """
 
+def read_pdf_to_documents(file_path):
+    doc = fitz.open(file_path)
+    documents = []
+    for page_num in range(len(doc)):
+        page = doc.load_page(page_num)
+        text = page.get_text()
+        documents.append(Document(text=text)) # Now Document is defined
+    return documents
+
+# Function to update the global system prompt
+def update_system_prompt(new_prompt):
+    global system_prompt
+    system_prompt = new_prompt
+    query_wrapper_prompt = SimpleInputPrompt("{query_str} [/INST]")
+    return "System prompt updated."
+
+def query_model(question):
+    llm = HuggingFaceLLM(
+        context_window=4096,
+        max_new_tokens=256,
+        system_prompt=system_prompt,
+        query_wrapper_prompt=query_wrapper_prompt,
+        model=model,
+        tokenizer=tokenizer
+    )
+    #embeddings = LangchainEmbedding(HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2"))
+    service_context = ServiceContext.from_defaults(chunk_size=1024, llm=llm, embed_model=embeddings)
+    set_global_service_context(service_context)
+
+    response = query_engine.query(question)
+    # formatted_response = format_paragraph(response.response)
+    return response.response
+
+def format_paragraph(text, line_length=80):
+    words = text.split()
+    lines = []
+    current_line = []
+    current_length = 0
+
+    for word in words:
+        if current_length + len(word) + 1 > line_length:
+            lines.append(' '.join(current_line))
+            current_line = [word]
+            current_length = len(word) + 1
+        else:
+            current_line.append(word)
+            current_length += len(word) + 1
+
+    if current_line:
+        lines.append(' '.join(current_line))
+
+    return '\n'.join(lines)
+
 if not torch.cuda.is_available():
     DESCRIPTION += "We won't be able to run this space! We need GPU processing"
 
@@ -48,7 +102,49 @@ if torch.cuda.is_available():
     tokenizer = AutoTokenizer.from_pretrained(model_id)
     tokenizer.use_default_system_prompt = False
 
+    system_prompt = """<s>[INST] <<SYS>>
+
+    <</SYS>>"""
+    # Throw together the query wrapper
+    query_wrapper_prompt = SimpleInputPrompt("{query_str} [/INST]")
+    llm = HuggingFaceLLM(context_window=4096,
+                     max_new_tokens=256,
+                     system_prompt=system_prompt,
+                     query_wrapper_prompt=query_wrapper_prompt,
+                     model=model, tokenizer=tokenizer)
+    embeddings = LangchainEmbedding(HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2"))
+    service_context = ServiceContext.from_defaults(chunk_size=1024, llm=llm, embed_model=embeddings)
+    set_global_service_context(service_context)
+    file_path = Path('/content/Full Pamplet.pdf')#make sure to change this to the document path
+    documents = read_pdf_to_documents(file_path)
+    index = VectorStoreIndex.from_documents(documents)
+    query_engine = index.as_query_engine()
+
+
+update_prompt_interface = gr.Interface(
+    fn=update_system_prompt,
+    inputs=gr.Textbox(lines=5, placeholder="Enter the system prompt here...", label="System Prompt", value=system_prompt),
+    outputs=gr.Textbox(label="Status"),
+    title="System Prompt Updater",
+    description="Update the system prompt used for context."
+)
+
+# Create Gradio interface for querying the model
+query_interface = gr.Interface(
+    fn=query_model,
+    inputs=gr.Textbox(lines=2, placeholder="Enter your question here...", label="User Question"),
+    outputs=gr.Textbox(label="Response"),
+    title="Document Query Assistant",
+    description="Ask questions based on the content of the loaded pamphlet."
+)
 
+# Combine the interfaces
+combined_interface = gr.TabbedInterface([update_prompt_interface, query_interface], ["Update System Prompt", "Query Assistant"])
+
+# Launch the combined interface
+combined_interface.launch()
+
+"""
 @spaces.GPU(duration=240)
 def generate(
     message: str,
@@ -143,10 +239,12 @@ chat_interface = gr.ChatInterface(
         ["Write a 100-word article on 'Benefits of Open-Source in AI research'"],
     ],
 )
+"""
 
 with gr.Blocks(css="style.css") as demo:
     gr.Markdown(DESCRIPTION)
-    chat_interface.render()
+    #chat_interface.render()
+    combined_interface.render()
     gr.Markdown(LICENSE)
 
 if __name__ == "__main__":

requirements.txt

@@ -11,4 +11,6 @@ sentence-transformers
 PyPDF2
 PyMuPDF
 langchain
-langchain-community
+langchain_huggingface
+llama-index
+llama-index-llms-huggingface