Update app.py

app.py

@@ -6,8 +6,10 @@ from streamlit_drawable_canvas import st_canvas
 import torch
 from diffusers import AutoPipelineForInpainting
 import numpy as np
+from sentence_transformers import SentenceTransformer,util
 from streamlit_image_select import image_select
 import os
+import PyPDF2
 import requests 
 from streamlit_navigation_bar import st_navbar
 from langchain_community.llms import Ollama
@@ -23,7 +25,7 @@ from streamlit_modal import Modal
 import streamlit.components.v1 as components
 from datetime import datetime
 from streamlit_js_eval import streamlit_js_eval
-
+from streamlit_pdf_viewer import pdf_viewer
 def consume_llm_api(prompt):
     """
     Sends a prompt to the LLM API and processes the streamed response.
@@ -45,7 +47,9 @@ def consume_llm_api(prompt):
         print(f"Error consuming API: {e}")
     except Exception as e:
         print(f"Unexpected error: {e}")
-
+def send_prompt():
+    return "please respond according to the prompt asked below from the above context"
+    
 def image_to_base64(image_path):
     with open(image_path, "rb") as img_file:
         return base64.b64encode(img_file.read()).decode()
@@ -188,15 +192,17 @@ with column2:
                 for index,prompts_ in enumerate(dictionary['every_prompt_with_val'][::-1]):
                     if prompts_[-1]=="@working":
                         if index==0:
-                            st.write(prompts_[0].upper())
+
+                            st.write(prompts_[0].split(send_prompt())[-1].upper() if send_prompt() in prompts_[0] else prompts_[0].upper())
                             data_need=st.write_stream(consume_llm_api(prompts_[0]))
                             dictionary['every_prompt_with_val'][-1]=(prompts_[0],str(data_need))
                             
                     elif isinstance(prompts_[-1],str):
+                        show_case_text=prompts_[0].split(send_prompt())[-1].upper() if send_prompt() in prompts_[0] else prompts_[0].upper()
                         if index==0:
-                            st.text_area(label=prompts_[0].upper(),value=prompts_[-1],height=500,key=str(index))
+                            st.text_area(label=show_case_text,value=prompts_[-1],height=500,key=str(index))
                         else:
-                            st.text_area(label=prompts_[0].upper(),value=prompts_[-1],key=str(index))
+                            st.text_area(label=show_case_text,value=prompts_[-1],key=str(index))
 
                     else:
                         st.write(prompts_[0].upper())
@@ -397,20 +403,47 @@ with column1:
     changes,implementation,current=st.columns([0.01,0.9,0.01])
 
     with implementation:
-                st.write("<br>"*5,unsafe_allow_html=True)
-                canvas_result = st_canvas(
-                    fill_color="rgba(255, 165, 0, 0.3)",  # Fixed fill color with some opacity
-                    stroke_width=stroke_width,
-                    stroke_color=stroke_color,
-                    background_color=bg_color,
-                    background_image=gen_image if gen_image else Image.open("/home/user/app/ALL_image_formation/image_gen.png"),
-                    update_streamlit=True,
-                    height=screen_height//2.16 if screen_height!=1180 else screen_height/2,
-                    width=screen_width//2.3  if screen_width!=820 else screen_width/2,
-                    drawing_mode=drawing_mode,
-                    point_display_radius=point_display_radius if drawing_mode == 'point' else 0,
-                    key="canvas",
-                )
+                if bg_doc:
+                    
+                    canvas_result=None
+                    binary_data = bg_doc.getvalue()
+                    pdf_viewer(input=binary_data,
+                            width=int(screen_width//2.3)  if screen_width!=820 else screen_width//2,
+                            height=int(screen_height//2.16) if screen_height!=1180 else screen_height//2)
+                    with open("temp.pdf", "wb") as f:
+                        f.write(bg_doc.getbuffer())
+                    
+                    # Process the uploaded PDF file
+                    data = process_pdf("temp.pdf")
+                    text_splitter = RecursiveCharacterTextSplitter(chunk_size=7500, chunk_overlap=100)
+                    chunks = text_splitter.split_documents(data)
+                    # chunk_texts = [str(chunk.page_content) for chunk in chunks] 
+                    # print("testing",chunk_texts)
+                    model_name = "all-MiniLM-L6-v2"
+                    model = SentenceTransformer(model_name)
+                    embeddings = [model.encode(str(chunk.page_content)) for chunk in chunks]
+
+                    vector_store = []
+                    for chunk, embedding in zip(chunks, embeddings):
+                        vector_store.append((embedding, chunk.page_content) )
+                        
+                else:
+
+                    
+                    st.write("<br>"*5,unsafe_allow_html=True)
+                    canvas_result = st_canvas(
+                        fill_color="rgba(255, 165, 0, 0.3)",  # Fixed fill color with some opacity
+                        stroke_width=stroke_width,
+                        stroke_color=stroke_color,
+                        background_color=bg_color,
+                        background_image=gen_image if gen_image else Image.open("/home/user/app/ALL_image_formation/image_gen.png"),
+                        update_streamlit=True,
+                        height=screen_height//2.16 if screen_height!=1180 else screen_height/2,
+                        width=screen_width//2.3  if screen_width!=820 else screen_width/2,
+                        drawing_mode=drawing_mode,
+                        point_display_radius=point_display_radius if drawing_mode == 'point' else 0,
+                        key="canvas",
+                    )
 
 
 
@@ -424,9 +457,15 @@ with column1:
 
     # run=st.button("run_experiment")
 
+if bg_doc and prompt:
+    query_embedding = model.encode([prompt])
+    retrieved_chunks = max([(util.cos_sim(match[0],query_embedding),match[-1])for  match in vector_store])[-1]
+    print(retrieved_chunks)
+    prompt = "Context: "+ retrieved_chunks +"\n"+send_prompt()+ "\n"+prompt
 
-
-if canvas_result.image_data is not None:
+    modifiedValue="@working"
+    dictionary['every_prompt_with_val'].append((prompt,modifiedValue))
+elif not bg_doc and canvas_result.image_data is not None:
     if prompt:
 
         text_or_image=multimodel_output(prompt)