fix 18

app.py

@@ -6,6 +6,7 @@ from difflib import unified_diff
 import tempfile
 from docx.shared import RGBColor
 import re
+import subprocess
 
 def pdf_to_text_with_layout(pdf_file):
     doc = fitz.open(stream=pdf_file.read(), filetype="pdf")
@@ -38,32 +39,8 @@ def apply_pipeline(file, model_name, balance_type, apsn_transactions, max_fees_p
     )
 
 def redline_changes(original_path, revised_path, output_path):
-    original_doc = Document(original_path)
-    revised_doc = Document(revised_path)
-
-    original_text = "\n".join([para.text for para in original_doc.paragraphs])
-    revised_text = "\n".join([para.text for para in revised_doc.paragraphs])
-
-    diff = unified_diff(original_text.splitlines(), revised_text.splitlines(), lineterm='')
-
-    diff_doc = Document()
-    for line in diff:
-        if line.startswith('-'):
-            p = diff_doc.add_paragraph(style='Normal')
-            run = p.add_run(line)
-            run.font.color.rgb = RGBColor(255, 0, 0)  # Red
-        elif line.startswith('+'):
-            p = diff_doc.add_paragraph(style='Normal')
-            run = p.add_run(line)
-            run.font.color.rgb = RGBColor(0, 128, 0)  # Green
-        elif line.startswith('@@'):
-            p = diff_doc.add_paragraph(style='Normal')
-            run = p.add_run(line)
-            run.font.color.rgb = RGBColor(0, 0, 255)  # Blue
-        else:
-            diff_doc.add_paragraph(line, style='Normal')
-
-    diff_doc.save(output_path)
+    # Using docxcompose to create a redlined document
+    subprocess.run(['docxcompose', 'compose', original_path, revised_path, output_path])
 
 # Streamlit App
 st.title("Canarie AI Prototype")
@@ -108,6 +85,22 @@ if uploaded_file is not None:
             redlined_output_path = tempfile.NamedTemporaryFile(delete=False, suffix=".docx").name
             redline_changes(original_word_path, revised_word_path, redlined_output_path)
 
+            with open(original_word_path, "rb") as f:
+                st.download_button(
+                    label="Download Original Document",
+                    data=f,
+                    file_name="original_document.docx",
+                    mime="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+                )
+            
+            with open(revised_word_path, "rb") as f:
+                st.download_button(
+                    label="Download Revised Document",
+                    data=f,
+                    file_name="revised_document.docx",
+                    mime="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+                )
+
             with open(redlined_output_path, "rb") as f:
                 st.download_button(
                     label="Download Redlined Document",
@@ -115,7 +108,8 @@ if uploaded_file is not None:
                     file_name="redlined_document.docx",
                     mime="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
                 )
-            st.success("Redlined document created successfully!")
+
+            st.success("Documents created successfully!")
 
         except Exception as e:
             st.exception(e)

langchain_pipeline.py

@@ -54,7 +54,7 @@ def pipeline(file, model_name, balance_type, apsn_transactions, max_fees_per_day
         What is the minimum transaction amount to trigger an overdraft?: {min_transaction_overdraft}
 
         Please output in the following format: 
-        {{entire updated disclosure text}}
+        {{full text of the updated disclosure with formatting retained}}
         ------
         {{reasons_for_changes with cases or law cited}}
         """

requirements.txt

@@ -19,4 +19,5 @@ langchain_astradb
 langchain_core
 langchain_openai
 langchain_anthropic
-langchain_google_genai
+langchain_google_genai
+docxcompose