Merge pull request #1995 from UlrichB22/fix_1017

Rename and rework get_indexer function

Author: RogerHaase

src/moin/storage/middleware/indexing.py

@@ -1,7 +1,7 @@
 # Copyright: 2011 MoinMoin:RonnyPfannschmidt
 # Copyright: 2011 MoinMoin:ThomasWaldmann
 # Copyright: 2011 MoinMoin:MichaelMayorov
-# Copyright: 2024 MoinMoin:UlrichB
+# Copyright: 2024-2025 MoinMoin:UlrichB
 # License: GNU GPL v2 (or any later version), see LICENSE.txt for details.
 
 """
@@ -105,29 +105,6 @@
 INDEXER_TIMEOUT = 20.0
 
 
-def get_indexer(fn, **kw):
-    """
-    Return a valid indexer or raise a KeyError.
-
-    Under heavy loads, the Whoosh AsyncWriter writer may be delayed in writing
-    indexes to storage. Try several times before failing.
-
-    FIXME: runs into timeout for a non-existing revid
-
-    :param fn: the indexer function
-    :param **kw: "revid" is required, index name optional
-    """
-    until = time.time() + INDEXER_TIMEOUT
-    while True:
-        indexer = fn(**kw)
-        if indexer is not None:
-            break
-        time.sleep(2)
-        if time.time() > until:
-            raise KeyError(kw.get("revid", "") + " - server overload or corrupt index")
-    return indexer
-
-
 def parent_names(names):
     """
     Compute list of parent names (same order as in names, but no dupes)
@@ -1045,6 +1022,31 @@ def existing_item(self, **query):
         """
         return Item.existing(self, **query)
 
+    def get_document(self, revid: str, retry: bool = False, **kwargs):
+        """
+        Return a valid indexer document or raise a KeyError.
+
+        Under heavy loads, the Whoosh AsyncWriter writer may be delayed in writing
+        indexes to storage. Try several times before failing.
+
+        :param revid: revision to search
+        :param retry: retry backend search if document not found, required when server load is high
+        :param kwargs: idx_name, name of index used for searching (optional)
+        """
+        until = time.monotonic() + INDEXER_TIMEOUT
+        while True:
+            doc = self._document(revid=revid, **kwargs)
+            if doc is not None:
+                break
+            if not retry:
+                msg = f"revid: {revid} not found. Please check meta data and indexes"
+                raise KeyError(msg)
+            if time.monotonic() > until:
+                msg = f"revid: {revid} - Server overload may have corrupted the index; rebuild it."
+                raise KeyError(msg)
+            time.sleep(2)
+        return doc
+
 
 class PropertiesMixin:
     """
@@ -1357,9 +1359,9 @@ def store_revision(
         self.indexer.index_revision(meta, content, backend_name, force_latest=not overwrite)
         gc.collect()  # triggers close of index files from is_latest search
         if not overwrite:
-            self._current = get_indexer(self.indexer._document, revid=revid)
+            self._current = self.indexer.get_document(revid=revid, retry=True)
         if return_rev:
-            return Revision(self, revid)
+            return Revision(self, revid, retry=True)
 
     def store_all_revisions(self, meta, data):
         """
@@ -1404,13 +1406,13 @@ class Revision(PropertiesMixin):
     An existing revision (exists in the backend).
     """
 
-    def __init__(self, item: Item, revid: str, doc=None, name=None):
+    def __init__(self, item: Item, revid: str, doc=None, name=None, retry=False):
         is_current = revid == CURRENT
         if doc is None:
             if is_current:
                 doc = item._current
             else:
-                doc = get_indexer(item.indexer._document, idx_name=ALL_REVS, revid=revid)
+                doc = item.indexer.get_document(idx_name=ALL_REVS, revid=revid, retry=retry)
 
         if is_current:
             revid = doc.get(REVID)