fix: Add URL extraction to AsyncDBWriter to populate journal_urls table

The journal_urls table was remaining empty despite successful sync operations
because the AsyncDBWriter._batch_write_journals() method was not extracting
and storing URLs from journal data.

This fix adds comprehensive URL extraction logic that:
- Extracts URLs from top-level 'urls' field in journal records
- Extracts URLs from metadata['urls'] (Algerian Ministry format)
- Extracts URLs from metadata['website_url'] (Kscien format)
- Extracts URLs from metadata['source_url'] (general source URLs)
- Deduplicates URLs per journal using set() for efficiency
- Uses INSERT OR REPLACE SQL pattern consistent with existing code
- Follows database schema with proper journal_id foreign key relationships

Resolves issue where journal URL data was being collected by data sources
but lost during the batch write process, leaving journal_urls table empty.

Author: florath

src/aletheia_probe/cache_sync.py

@@ -137,6 +137,7 @@ def _batch_write_journals(
                 name_inserts = []
                 assessment_inserts = []
                 metadata_inserts = []
+                url_inserts = []
 
                 # First, collect all normalized names that will be processed
                 normalized_names = [
@@ -206,6 +207,39 @@ def _batch_write_journals(
                     # Source assessments
                     assessment_inserts.append((journal_id, source_id, list_type, 1.0))
 
+                    # URLs - Extract from multiple sources and deduplicate
+                    urls_to_insert = set()  # Use set for automatic deduplication
+
+                    # Check for top-level urls field
+                    if journal.get("urls"):
+                        for url in journal["urls"]:
+                            if url and isinstance(url, str) and url.strip():
+                                urls_to_insert.add(url.strip())
+
+                    # Extract URLs from metadata
+                    if metadata:
+                        # Handle Algerian Ministry format: metadata["urls"] as list
+                        if "urls" in metadata and isinstance(metadata["urls"], list):
+                            for url in metadata["urls"]:
+                                if url and isinstance(url, str) and url.strip():
+                                    urls_to_insert.add(url.strip())
+
+                        # Handle Kscien format: metadata["website_url"] as string
+                        if "website_url" in metadata and metadata["website_url"]:
+                            url = metadata["website_url"]
+                            if isinstance(url, str) and url.strip():
+                                urls_to_insert.add(url.strip())
+
+                        # Handle other potential URL fields in metadata
+                        if "source_url" in metadata and metadata["source_url"]:
+                            url = metadata["source_url"]
+                            if isinstance(url, str) and url.strip():
+                                urls_to_insert.add(url.strip())
+
+                    # Add deduplicated URLs to batch inserts
+                    for url in urls_to_insert:
+                        url_inserts.append((journal_id, url))
+
                     # Metadata
                     if metadata:
                         for key, value in metadata.items():
@@ -253,6 +287,15 @@ def _batch_write_journals(
                         metadata_inserts,
                     )
 
+                # Batch insert/update journal URLs
+                if url_inserts:
+                    cursor.executemany(
+                        """INSERT OR REPLACE INTO journal_urls
+                           (journal_id, url, last_seen_at)
+                           VALUES (?, ?, CURRENT_TIMESTAMP)""",
+                        url_inserts,
+                    )
+
                 # Commit transaction
                 conn.execute("COMMIT")