Add redirect_target column to wikiq output.

Exposes page.redirect as a nullable string column so downstream pipelines
can build redirect maps for link target resolution.

src/wikiq/__init__.py

@@ -133,6 +133,7 @@ def build_table(
         tables.RevisionArticleId(),
         tables.RevisionPageTitle(),
         tables.RevisionNamespace(),
+        tables.RevisionRedirectTarget(),
         tables.RevisionDeleted(),
         tables.RevisionEditorId(),
         tables.RevisionEditSummary(),
@@ -1004,7 +1005,7 @@ class WikiqParser:
                         if rev.id == page_resume_revid:
                             found_resume_point = True
                             is_resume_page = False
-                            print(f"Resuming output after revid {rev.id}", file=sys.stderr)
+                            print(f"Resuming output after revid {rev.id}", file=sys.stderr, flush=True)
                         continue
 
                 rev_count += 1

src/wikiq/tables.py

@@ -1,4 +1,3 @@
-import sys
 from abc import abstractmethod, ABC
 from datetime import datetime, timezone
 from hashlib import sha1
@@ -144,6 +143,13 @@ class RevisionNamespace(RevisionField[int]):
         return page.namespace
 
 
+class RevisionRedirectTarget(RevisionField[Union[str, None]]):
+    field = pa.field("redirect_target", pa.string(), nullable=True)
+
+    def extract(self, page: mwtypes.Page, revisions: list[mwxml.Revision]) -> Union[str, None]:
+        return page.redirect
+
+
 class RevisionSha1(RevisionField[str]):
     field = pa.field("sha1", pa.string())