Update app.py

app.py

@@ -92,21 +92,31 @@ def load_css():
     """, unsafe_allow_html=True)
 
 # --- Helper Functions ---
+def normalise_hyphens(text):
+    # Replace hyphen variants with U+002D for internal consistency
+    return text.replace('\u2011', '-').replace('\u2013', '-').replace('\u2014', '-')
+
 def encode_text_fragment(text):
     # Encode text for W3C Text Fragments, preserving only hyphens
     # En dashes (–) and em dashes (—) are encoded as %E2%80%93 and %E2%80%94
     return urllib.parse.quote(text, safe='-')
 
 def generate_citation_hash(author, year, url, fragment_text, cited_text, username, task_name, current_date, current_time):
-    data = f"{author}, {year} | {url} | {fragment_text} | {cited_text} | {username} | {task_name} | {current_date} | {current_time}"
+    # Normalise hyphens for consistent hash generation
+    normalised_fragment_text = normalise_hyphens(fragment_text)
+    normalised_cited_text = normalise_hyphens(cited_text)
+    normalised_task_name = normalise_hyphens(task_name)
+    data = f"{author}, {year} | {url} | {normalised_fragment_text} | {normalised_cited_text} | {username} | {normalised_task_name} | {current_date} | {current_time}"
     return hashlib.sha256(data.encode('utf-8')).hexdigest()
 
 def format_citation_html(url, fragment_text, author, year, scc_hash):
+    # Use original fragment_text for text fragment URL to match external source
     encoded_fragment = encode_text_fragment(fragment_text)
     full_url = f"{url}#:~:text={encoded_fragment}"
     return f'<a href="{full_url}" data-hash="{scc_hash}">{author} ({year})</a>'
 
 def format_metadata_html(url, author, year, scc_hash, username, task_name, current_date, current_time):
+    # Use original task_name with em dashes for text fragment URL
     metadata = f"{username}—{task_name}—{current_date}—{current_time}"
     encoded_metadata = encode_text_fragment(metadata)
     full_url = f"{url}#:~:text={encoded_metadata}"
@@ -345,7 +355,10 @@ with tabs[0]:
             </div>
             """, unsafe_allow_html=True)
         else:
-            scc_hash = generate_citation_hash(author_name, publication_year, source_url, annotated_text, annotated_text, username, task_name, current_date, current_time)
+            # Normalise hyphens in user inputs for hash generation
+            normalised_annotated_text = normalise_hyphens(annotated_text)
+            normalised_task_name = normalise_hyphens(task_name)
+            scc_hash = generate_citation_hash(author_name, publication_year, source_url, normalised_annotated_text, normalised_annotated_text, username, normalised_task_name, current_date, current_time)
             citation_link_start = format_citation_html(source_url, annotated_text, author_name, publication_year, scc_hash)
             citation_link_end = f'<a href="{source_url}#:~:text={encode_text_fragment(annotated_text)}" data-hash="{scc_hash}">({author_name}, {publication_year})</a>'
             metadata_link = format_metadata_html(source_url, author_name, publication_year, scc_hash, username, task_name, current_date, current_time)
@@ -404,9 +417,12 @@ with tabs[1]:
             elif citation_base_url != hash_base_url:
                 st.error("The citation URL and SCC index URL must point to the same base URL.")
             else:
+                # Normalise hyphens for hash recomputation
+                normalised_citation_fragment = normalise_hyphens(citation_fragment)
+                normalised_task_name = normalise_hyphens(task_name)
                 # Recompute hash
                 recomputed_hash = generate_citation_hash(
-                    author, year, citation_base_url, citation_fragment, citation_fragment, username, task_name, date, time
+                    author, year, citation_base_url, normalised_citation_fragment, normalised_citation_fragment, username, normalised_task_name, date, time
                 )
 
                 if recomputed_hash == scc_hash: