Update app.py

app.py

@@ -8,6 +8,9 @@ from oauth2client.service_account import ServiceAccountCredentials
 # Set up OpenAI client
 client = OpenAI(api_key=st.secrets["OPENAI_API_KEY"])
 
+# Constants
+WORD_LIMIT = 8000
+
 # Google Sheets setup remains the same
 scope = ["https://spreadsheets.google.com/feeds", "https://www.googleapis.com/auth/drive"]
 creds = ServiceAccountCredentials.from_json_keyfile_name("genexam-2c8c645ecc0d.json", scope)
@@ -34,6 +37,21 @@ def update_api_usage(username):
     except Exception as e:
         st.error(f"Error updating API usage: {str(e)}")
 
+def extract_text_from_pdf(pdf_file):
+    """Simple PDF text extraction with word limit check"""
+    try:
+        pdf_reader = PyPDF2.PdfReader(io.BytesIO(pdf_file.read()))
+        text_content = ""
+        for page in pdf_reader.pages:
+            text_content += page.extract_text() + "\n"
+        
+        word_count = len(text_content.split())
+        if word_count > WORD_LIMIT:
+            return None, f"PDF content exceeds {WORD_LIMIT:,} words (contains {word_count:,} words). Please use a shorter document."
+        return text_content, None
+    except Exception as e:
+        return None, f"Error processing PDF: {str(e)}"
+
 def generate_questions_with_retry(knowledge_material, question_type, cognitive_level, extra_instructions, case_based, num_choices=None, max_retries=3):
     # Adjust number of questions based on type
     if question_type == "Multiple Choice":
@@ -143,6 +161,7 @@ Please format the output with:
             time.sleep(2)
 
 # ระบบ login
+# Main Streamlit interface
 if 'username' not in st.session_state:
     st.title("Login")
     username_input = st.text_input("Enter your username:")
@@ -157,106 +176,74 @@ if 'username' not in st.session_state:
         else:
             st.warning("Please enter a valid username.")
 else:
-    # Main App after login
     st.title(f"Welcome, {st.session_state['username']}! Generate your exam questions")
-
-    # Input field for knowledge material (text) with 3,000-word limit
-    knowledge_material = st.text_area("Enter knowledge material to generate exam questions:")
     
-    # Word count check
-    if len(knowledge_material.split()) > 3000:
-        st.warning("Please limit the knowledge material to 3,000 words or fewer.")
-
-    # File uploader for PDFs (limited to 5 MB)
-    uploaded_file = st.file_uploader("Upload a file (PDF)", type="pdf")
+    # Create tabs for input methods
+    tab1, tab2 = st.tabs(["Text Input", "PDF Upload"])
     
-    if uploaded_file is not None:
-        if uploaded_file.size > 5 * 1024 * 1024:  # 5 MB limit
-            st.warning("File size exceeds 5 MB. Please upload a smaller file.")
-        else:
-            st.success("File uploaded successfully! (Text extraction not implemented yet.)")
-
-    # Select question type
-    question_type = st.selectbox("Select question type:", 
-                                 ["Multiple Choice", "Fill in the Blank", "Open-ended", "True/False"])
-
-    # For multiple choice, let users select the number of choices
-    num_choices = None
-    if question_type == "Multiple Choice":
-        num_choices = st.selectbox("Select the number of choices for each question:", [3, 4, 5])
-
-    # Select cognitive level
-    cognitive_level = st.selectbox("Select cognitive level:", 
-                                   ["Recall", "Understanding", "Application", "Analysis", "Synthesis", "Evaluation"])
-
-    # Checkbox for Case-Based Medical Situations
-    case_based = st.checkbox("Generate case-based medical exam questions")
+    with tab1:
+        knowledge_material = st.text_area("Enter knowledge material to generate exam questions:")
+        word_count = len(knowledge_material.split())
+        if word_count > WORD_LIMIT:
+            st.error(f"Text exceeds {WORD_LIMIT:,} words. Please shorten your content.")
+    
+    with tab2:
+        st.info(f"Maximum content length: {WORD_LIMIT:,} words")
+        uploaded_file = st.file_uploader("Upload a PDF file", type="pdf")
+        
+        if uploaded_file is not None:
+            pdf_content, error = extract_text_from_pdf(uploaded_file)
+            if error:
+                st.error(error)
+            else:
+                st.success("PDF processed successfully!")
+                knowledge_material = pdf_content
 
-    # Extra input field for additional instructions (placed below cognitive level)
-    extra_instructions = st.text_area("Enter additional instructions (e.g., how you want the questions to be phrased):")
+    # Question generation options
+    col1, col2 = st.columns(2)
+    
+    with col1:
+        question_type = st.selectbox(
+            "Select question type:",
+            ["Multiple Choice", "Fill in the Blank", "Open-ended", "True/False"]
+        )
+        
+        if question_type == "Multiple Choice":
+            num_choices = st.selectbox("Select number of choices:", [3, 4, 5])
+        
+        cognitive_level = st.selectbox(
+            "Select cognitive level:",
+            ["Recall", "Understanding", "Application", "Analysis", "Synthesis", "Evaluation"]
+        )
+    
+    with col2:
+        case_based = st.checkbox("Generate case-based medical exam questions")
+        extra_instructions = st.text_area("Additional instructions (optional):")
 
     # Generate questions button
-    if 'previous_questions' not in st.session_state:
-        st.session_state['previous_questions'] = []
-
     if st.button("Generate Questions"):
-        if len(knowledge_material.split()) <= 3000:
-            # Generate questions with retry logic
-            questions = generate_questions_with_retry(
-                knowledge_material, 
-                question_type, 
-                cognitive_level, 
-                extra_instructions, 
-                case_based, 
-                num_choices
-            )
-            
-            if questions:
-                st.write("Generated Exam Questions:")
-                st.write(questions)
-
-                # Avoid showing repeated content in future requests
-                st.session_state['previous_questions'].append(questions)
-
-                # Option to download the questions as a text file
-                st.download_button(
-                    label="Download Questions",
-                    data=questions,
-                    file_name='generated_questions.txt',
-                    mime='text/plain'
-                )
-        else:
-            st.warning("Please reduce the word count to 3,000 or fewer.")
-
-    # Button to generate more questions based on the same material
-    if st.button("Generate More Questions"):
-        if len(knowledge_material.split()) <= 3000:
-            # Regenerate new questions, trying to avoid repeated content
-            questions = generate_questions_with_retry(
-                knowledge_material, 
-                question_type, 
-                cognitive_level, 
-                extra_instructions, 
-                case_based, 
-                num_choices
-            )
-
-            # Check if the new set of questions is not the same as the previous set
-            if questions and questions not in st.session_state['previous_questions']:
-                st.write("Generated More Exam Questions:")
-                st.write(questions)
-
-                # Append the new questions to the session state
-                st.session_state['previous_questions'].append(questions)
-
-                # Option to download the new set of questions
-                st.download_button(
-                    label="Download More Questions",
-                    data=questions,
-                    file_name='more_generated_questions.txt',
-                    mime='text/plain'
+        if 'knowledge_material' in locals() and knowledge_material.strip():
+            with st.spinner("Generating questions..."):
+                # Your existing generate_questions_with_retry function call here
+                questions = generate_questions_with_retry(
+                    knowledge_material,
+                    question_type,
+                    cognitive_level,
+                    extra_instructions,
+                    case_based,
+                    num_choices if question_type == "Multiple Choice" else None
                 )
-            else:
-                st.warning("New questions seem to overlap with the previous ones. Try adjusting the instructions.")
+                
+                if questions:
+                    st.write("### Generated Exam Questions:")
+                    st.write(questions)
+                    
+                    # Download button
+                    st.download_button(
+                        label="Download Questions",
+                        data=questions,
+                        file_name='generated_questions.txt',
+                        mime='text/plain'
+                    )
         else:
-            st.warning("Please reduce the word count to 3,000 or fewer.")
+            st.warning("Please enter knowledge material or upload a PDF file first.")