app.py

app.py

@@ -1,17 +1,17 @@
+import os
+import gradio as gr
+import subprocess
 import streamlit as st
 from openai import OpenAI
 import time
-import os
-import re
-import pandas as pd
 import PyPDF2
+from docx import Document
 from datetime import datetime
 from pydub import AudioSegment
-from docx import Document
 from io import BytesIO
 
+# Streamlit Page Config
 st.set_page_config(page_title="Schlager ContractAi")
-
 st.title("Schlager ContractAi")
 st.caption("Chat with your contract or manage meeting minutes")
 
@@ -19,13 +19,62 @@ st.caption("Chat with your contract or manage meeting minutes")
 with st.sidebar:
     OPENAI_API_KEY = st.text_input("Enter your C2 Group of Technologies Access Key", type="password")
 
-# Tabs for Contract, Technical, and Minutes
-tab1, tab2, tab3 = st.tabs(["Contract", "Technical", "Minutes"])
+# Tabs
+tab1, tab2, tab3, tab4 = st.tabs(["Contract", "Technical", "Minutes", "Document Preparation"])
 
 SUPPORTED_AUDIO_FORMATS = (".mp3", ".wav", ".m4a")
 SUPPORTED_TEXT_FORMATS = (".txt", ".docx", ".csv", ".xlsx", ".pdf")
 
+def install_dependencies():
+    try:
+        subprocess.run(['apt-get', 'update'], check=True)
+        subprocess.run(['apt-get', 'install', '-y', 'poppler-utils'], check=True)
+        subprocess.run(['apt-get', 'install', '-y', 'tesseract-ocr'], check=True)
+        subprocess.run(['apt-get', 'install', '-y', 'tesseract-ocr-eng'], check=True)
+        print("Packages installed successfully!")
+    except subprocess.CalledProcessError as e:
+        print(f"An error occurred: {e}")
+
+install_dependencies()
+
+def process_pdf(file):
+    input_pdf = file.name
+    os.system(f'pdftoppm -png "{input_pdf}" img')
+    
+    for image in os.listdir():
+        if image.startswith('img') and image.endswith('.png'):
+            output_txt = f"ocr_{image}.txt"
+            os.system(f'tesseract "{image}" "{output_txt[:-4]}"')
+    
+    output_txt_file = f"{input_pdf[:-4]}.txt"
+    with open(output_txt_file, 'w') as output_file:
+        for text_file in os.listdir():
+            if text_file.startswith('ocr_img') and text_file.endswith('.txt'):
+                with open(text_file, 'r') as f:
+                    output_file.write(f.read())
+                    output_file.write("\n")
+    
+    for file in os.listdir():
+        if file.startswith('img') or file.startswith('ocr_img'):
+            os.remove(file)
+    
+    return output_txt_file
+
+with tab4:
+    st.subheader("Document Preparation")
+    uploaded_file = st.file_uploader("Upload a PDF (Max: 200MB)", type=["pdf"], accept_multiple_files=False)
+    
+    if uploaded_file:
+        st.write("Processing the uploaded document...")
+        result_file = process_pdf(uploaded_file)
+        st.download_button(
+            label="Download Extracted Text",
+            data=open(result_file, "rb").read(),
+            file_name=result_file,
+            mime="text/plain"
+        )
 
+# Contract Chat Section
 def contract_chat_section(tab, assistant_id, session_key, input_key):
     with tab:
         st.subheader("Chat")
@@ -87,81 +136,3 @@ contract_chat_section(tab1, ASSISTANT_CONTRACT_ID, "contract_messages", "contrac
 # Technical Chat Section
 contract_chat_section(tab2, ASSISTANT_TECHNICAL_ID, "technical_messages", "technical_input")
 
-with tab3:
-    st.subheader("Minutes")
-    
-    if "generated_minutes" not in st.session_state:
-        st.session_state["generated_minutes"] = ""
-
-    uploaded_files = st.file_uploader("Upload meeting minutes (PDF/DOCX/Audio)", 
-                                      type=["pdf", "docx", "mp3", "wav", "m4a"], 
-                                      accept_multiple_files=True)
-    
-    if uploaded_files:
-        st.write("### Uploaded Files:")
-        for uploaded_file in uploaded_files:
-            st.write(f"- {uploaded_file.name}")
-        
-        combined_text = ""
-        for uploaded_file in uploaded_files:
-            if uploaded_file.name.lower().endswith(SUPPORTED_AUDIO_FORMATS):
-                audio = AudioSegment.from_file(uploaded_file)
-                temp_audio_path = "temp_audio.mp3"
-                audio.export(temp_audio_path, format="mp3")
-                with open(temp_audio_path, "rb") as audio_file:
-                    transcription = client.audio.transcriptions.create(
-                        model="whisper-1",
-                        file=audio_file
-                    )
-                combined_text += transcription.text + "\n"
-                os.remove(temp_audio_path)
-            else:
-                if uploaded_file.name.endswith(".docx"):
-                    doc = Document(uploaded_file)
-                    combined_text += "\n".join([para.text for para in doc.paragraphs])
-                elif uploaded_file.name.endswith(".pdf"):
-                    pdf_reader = PyPDF2.PdfReader(uploaded_file)
-                    combined_text += "\n".join([page.extract_text() for page in pdf_reader.pages if page.extract_text()])
-        
-        if combined_text:
-            st.write("### Transcribed and Extracted Text:")
-            st.text_area("Meeting Transcript", combined_text, height=300)
-            
-            if st.button("Generate Meeting Minutes", key="generate_minutes"):
-                prompt = f"""
-                Based on the following meeting transcript, generate structured meeting minutes in the format below:
-                
-                ---
-                **Meeting Name:** [Enter meeting name]  
-                **Location:** [Enter location]  
-                **Date:** [Enter date]  
-                **Time:** [Enter time]  
-                **Attendees:** [List attendees]  
-                ### Agenda Items
-                - [Agenda Item 1]
-                - [Agenda Item 2]
-                - [Agenda Item 3]
-                ### Action Items
-                | Action Item | Owner(s) | Deadline | Status |
-                |------------|---------|----------|--------|
-                | [Action Item 1] | [Owner(s) 1] | [Deadline 1] | [Status 1] |
-                | [Action Item 2] | [Owner(s) 2] | [Deadline 2] | [Status 2] |
-                | [Action Item 3] | [Owner(s) 3] | [Deadline 3] | [Status 3] |
-                ---
-                **Meeting Summary:**  
-                [Brief summary of key points, discussions, and decisions]
-                **Transcript:**  
-                {combined_text}
-                """
-
-                response = client.chat.completions.create(
-                    model="gpt-4-turbo",
-                    messages=[
-                        {"role": "system", "content": "You are an AI assistant that generates professional meeting minutes in a structured template."},
-                        {"role": "user", "content": prompt}
-                    ]
-                )
-                st.session_state["generated_minutes"] = response.choices[0].message.content
-                
-                st.write("### Generated Meeting Minutes:")
-                st.text_area("Meeting Minutes", st.session_state["generated_minutes"], height=400)