Update app.py

app.py

@@ -18,20 +18,36 @@ st.set_page_config(
 # ========== CACHED DATA LOADING ==========
 @st.cache_data(show_spinner=False)
 def load_questions():
-    with open('data/questions.txt') as f:
-        return f.readlines()
+    try:
+        with open('data/questions.txt') as f:
+            return [q.strip() for q in f.readlines() if q.strip()]
+    except Exception as e:
+        st.error(f"Error loading questions: {str(e)}")
+        return []
 
 @st.cache_data(show_spinner=False)
 def load_questions_short():
-    with open('data/questions_short.txt') as f:
-        return f.readlines()
+    try:
+        with open('data/questions_short.txt') as f:
+            return [q.strip() for q in f.readlines() if q.strip()]
+    except Exception as e:
+        st.error(f"Error loading short questions: {str(e)}")
+        return []
 
 # ========== UTILITY FUNCTIONS ==========
 def extract_text_from_pdf(uploaded_file):
-    with pdfplumber.open(uploaded_file) as pdf:
-        return "\n".join(page.extract_text() or "" for page in pdf.pages)
+    try:
+        with pdfplumber.open(uploaded_file) as pdf:
+            text = "\n".join(page.extract_text() or "" for page in pdf.pages)
+            return text if text.strip() else ""
+    except Exception as e:
+        st.error(f"PDF extraction error: {str(e)}")
+        return ""
 
 def highlight_differences(text1, text2):
+    if not text1 or not text2:
+        return ""
+    
     differ = difflib.Differ()
     diff = list(differ.compare(text1.split(), text2.split()))
     
@@ -48,32 +64,50 @@ def highlight_differences(text1, text2):
     return highlighted_text
 
 def calculate_similarity(text1, text2):
-    vectorizer = TfidfVectorizer()
-    tfidf_matrix = vectorizer.fit_transform([text1, text2])
-    similarity = cosine_similarity(tfidf_matrix[0:1], tfidf_matrix[1:2])
-    return similarity[0][0] * 100
+    if not text1.strip() or not text2.strip():
+        return 0.0
+    
+    try:
+        vectorizer = TfidfVectorizer(token_pattern=r'(?u)\b\w+\b')
+        tfidf_matrix = vectorizer.fit_transform([text1, text2])
+        similarity = cosine_similarity(tfidf_matrix[0:1], tfidf_matrix[1:2])
+        return similarity[0][0] * 100
+    except ValueError:
+        return difflib.SequenceMatcher(None, text1, text2).ratio() * 100
 
 def load_contract(file):
+    if file is None:
+        return ""
+    
     ext = file.name.split('.')[-1].lower()
-    if ext == 'txt':
-        return StringIO(file.getvalue().decode("utf-8")).read()
-    elif ext == 'pdf':
-        try:
-            pdfReader = PyPDF4.PdfFileReader(file)
-            return '\n'.join([pdfReader.getPage(i).extractText() for i in range(pdfReader.numPages)])
-        except:
-            st.warning('Unable to read PDF, please try another file')
-    elif ext == 'docx':
-        return docx2txt.process(file)
-    else:
-        st.warning('Unsupported file type')
-    return ""
+    try:
+        if ext == 'txt':
+            content = StringIO(file.getvalue().decode("utf-8")).read()
+        elif ext == 'pdf':
+            content = extract_text_from_pdf(file)
+            if not content:
+                # Fallback to PyPDF4
+                pdfReader = PyPDF4.PdfFileReader(file)
+                content = '\n'.join([pdfReader.getPage(i).extractText() for i in range(pdfReader.numPages)])
+        elif ext == 'docx':
+            content = docx2txt.process(file)
+        else:
+            st.warning('Unsupported file type')
+            return ""
+        return content.strip() if content else ""
+    except Exception as e:
+        st.error(f"Error loading {ext.upper()} file: {str(e)}")
+        return ""
 
 # ========== MAIN APP ==========
 def main():
     questions = load_questions()
     questions_short = load_questions_short()
 
+    if not questions or not questions_short or len(questions) != len(questions_short):
+        st.error("Failed to load questions or questions mismatch. Please check data files.")
+        return
+
     st.title("📑 Contract Analysis Suite")
     st.markdown("""
     Compare documents and analyze legal clauses using AI-powered question answering.
@@ -90,8 +124,8 @@ def main():
             key="file1"
         )
         contract_text1 = load_contract(uploaded_file1) if uploaded_file1 else ""
-        st.text_area("Document 1 Content", value=contract_text1, height=200, key="area1")
-
+        doc1_display = st.empty()
+        
     with col2:
         uploaded_file2 = st.file_uploader(
             "Upload Second Document", 
@@ -99,7 +133,19 @@ def main():
             key="file2"
         )
         contract_text2 = load_contract(uploaded_file2) if uploaded_file2 else ""
-        st.text_area("Document 2 Content", value=contract_text2, height=200, key="area2")
+        doc2_display = st.empty()
+
+    # Update document displays
+    if uploaded_file1:
+        doc1_display.text_area("Document 1 Content", 
+                             value=contract_text1, 
+                             height=200, 
+                             key="area1")
+    if uploaded_file2:
+        doc2_display.text_area("Document 2 Content", 
+                             value=contract_text2, 
+                             height=200, 
+                             key="area2")
 
     if not (uploaded_file1 and uploaded_file2):
         st.warning("Please upload both documents to proceed")
@@ -110,6 +156,10 @@ def main():
     with st.expander("Show Document Differences", expanded=True):
         if st.button("Compare Documents"):
             with st.spinner("Analyzing documents..."):
+                if not contract_text1.strip() or not contract_text2.strip():
+                    st.error("One or both documents appear to be empty or couldn't be read properly")
+                    return
+                
                 similarity_score = calculate_similarity(contract_text1, contract_text2)
                 st.metric("Document Similarity Score", f"{similarity_score:.2f}%")
                 
@@ -125,18 +175,23 @@ def main():
 
     # ===== QUESTION ANALYSIS SECTION =====
     st.header("3. Clause Analysis")
-    question_selected = st.selectbox(
-        'Select a legal question to analyze:',
-        questions_short,
-        index=0,
-        key="question_select"
-    )
-    question_idx = questions_short.index(question_selected)
-    selected_question = questions[question_idx]
+    
+    try:
+        question_selected = st.selectbox(
+            'Select a legal question to analyze:',
+            questions_short,
+            index=0,
+            key="question_select"
+        )
+        question_idx = questions_short.index(question_selected)
+        selected_question = questions[question_idx]
+    except Exception as e:
+        st.error(f"Error selecting question: {str(e)}")
+        return
 
     if st.button("Analyze Both Documents"):
-        if not (contract_text1 and contract_text2):
-            st.error("Please ensure both documents have content")
+        if not (contract_text1.strip() and contract_text2.strip()):
+            st.error("Please ensure both documents have readable content")
             return
             
         col1, col2 = st.columns(2)
@@ -144,16 +199,22 @@ def main():
         with col1:
             st.subheader("First Document Analysis")
             with st.spinner('Processing first document...'):
-                predictions1 = run_prediction([selected_question], contract_text1, 'marshmellow77/roberta-base-cuad', n_best_size=5)
-                answer1 = predictions1.get('0', 'No answer found')
-                st.success(answer1)
+                try:
+                    predictions1 = run_prediction([selected_question], contract_text1, 'marshmellow77/roberta-base-cuad', n_best_size=5)
+                    answer1 = predictions1.get('0', 'No answer found')
+                    st.success(answer1 if answer1 else "No relevant clause found")
+                except Exception as e:
+                    st.error(f"Analysis failed for Document 1: {str(e)}")
         
         with col2:
             st.subheader("Second Document Analysis")
             with st.spinner('Processing second document...'):
-                predictions2 = run_prediction([selected_question], contract_text2, 'marshmellow77/roberta-base-cuad', n_best_size=5)
-                answer2 = predictions2.get('0', 'No answer found')
-                st.success(answer2)
+                try:
+                    predictions2 = run_prediction([selected_question], contract_text2, 'marshmellow77/roberta-base-cuad', n_best_size=5)
+                    answer2 = predictions2.get('0', 'No answer found')
+                    st.success(answer2 if answer2 else "No relevant clause found")
+                except Exception as e:
+                    st.error(f"Analysis failed for Document 2: {str(e)}")
 
 if __name__ == "__main__":
     main()