Make pending enrichment query aware of parser version

CHANGELOG.md

@@ -6,6 +6,8 @@ The format is based on [Keep a Changelog 1.0.0].
 
 ## Unreleased
 
+- **Breaking:** `Client.get_pending_enrichment_for_version` now requires both a target enrichment version and a target parser version, and will not include documents which have not been parsed with the target version.
+
 ## [Release 20.0.0]
 
 - **Feature:** New `Client.get_pending_parse_for_version` and `Client.get_highest_parser_version` methods to help find documents in need of re-parsing.

src/caselawclient/Client.py

@@ -968,12 +968,16 @@ def get_highest_enrichment_version(self) -> tuple[int, int]:
         return (int(table[1][1]), int(table[1][2]))
 
     def get_pending_enrichment_for_version(
-        self, target_version: tuple[int, int]
+        self,
+        target_enrichment_version: tuple[int, int],
+        target_parser_version: tuple[int, int],
     ) -> list[list[Any]]:
         """Retrieve documents which are not yet enriched with a given version."""
         vars: query_dicts.GetPendingEnrichmentForVersionDict = {
-            "target_major_version": target_version[0],
-            "target_minor_version": target_version[1],
+            "target_enrichment_major_version": target_enrichment_version[0],
+            "target_enrichment_minor_version": target_enrichment_version[1],
+            "target_parser_major_version": target_parser_version[0],
+            "target_parser_minor_version": target_parser_version[1],
         }
         results: list[list[Any]] = json.loads(
             get_single_string_from_marklogic_response(

src/caselawclient/xquery/get_pending_enrichment_for_version.xqy

@@ -1,28 +1,35 @@
 xquery version "1.0-ml";
 
-declare variable $target_major_version as xs:int external;
-declare variable $target_minor_version as xs:int external;
+declare variable $target_enrichment_major_version as xs:int external;
+declare variable $target_enrichment_minor_version as xs:int external;
+declare variable $target_parser_major_version as xs:int external;
+declare variable $target_parser_minor_version as xs:int external;
 
 xdmp:to-json(xdmp:sql(
   "SELECT process_data.uri, enrich_version_string, minutes_since_enrichment_request
   FROM (
     SELECT
       process_data.uri,
       enrich_version_string, enrich_major_version, enrich_minor_version,
+      parser_major_version, parser_minor_version,
       DATEDIFF('minute', last_sent_to_enrichment, CURRENT_TIMESTAMP) AS minutes_since_enrichment_request
     FROM documents.process_data
     JOIN documents.process_property_data ON process_data.uri = process_property_data.uri
   )
   WHERE (
     (enrich_version_string IS NULL) OR
-    (enrich_major_version <= @target_major_version AND enrich_minor_version < @target_minor_version)
+    (enrich_major_version <= @target_enrichment_major_version AND enrich_minor_version < @target_enrichment_minor_version)
+  ) AND (
+    (parser_major_version = @target_parser_major_version AND parser_minor_version = @target_parser_minor_version)
   )
   AND (minutes_since_enrichment_request > 43200 OR minutes_since_enrichment_request IS NULL)
   ORDER BY enrich_major_version ASC NULLS FIRST, enrich_minor_version ASC",
   "array",
   map:new((
-    map:entry("target_major_version", $target_major_version),
-    map:entry("target_minor_version", $target_minor_version)
+    map:entry("target_enrichment_major_version", $target_enrichment_major_version),
+    map:entry("target_enrichment_minor_version", $target_enrichment_minor_version),
+    map:entry("target_parser_major_version", $target_parser_major_version),
+    map:entry("target_parser_minor_version", $target_parser_minor_version)
   ))
 ))
 

src/caselawclient/xquery_type_dicts.py

@@ -80,8 +80,10 @@ class GetLastModifiedDict(MarkLogicAPIDict):
 
 # get_pending_enrichment_for_version.xqy
 class GetPendingEnrichmentForVersionDict(MarkLogicAPIDict):
-    target_major_version: int
-    target_minor_version: int
+    target_enrichment_major_version: int
+    target_enrichment_minor_version: int
+    target_parser_major_version: int
+    target_parser_minor_version: int
 
 
 # get_pending_parse_for_version.xqy