Set scores where there are none

Author: thecoop

server/src/main/java/org/elasticsearch/action/search/RankFeaturePhase.java

@@ -20,6 +20,7 @@
 import org.elasticsearch.search.builder.SearchSourceBuilder;
 import org.elasticsearch.search.dfs.AggregatedDfs;
 import org.elasticsearch.search.internal.ShardSearchContextId;
+import org.elasticsearch.search.rank.RankDoc;
 import org.elasticsearch.search.rank.context.RankFeaturePhaseRankCoordinatorContext;
 import org.elasticsearch.search.rank.feature.RankFeatureDoc;
 import org.elasticsearch.search.rank.feature.RankFeatureResult;
@@ -187,7 +188,7 @@ private void onPhaseDone(
             new ActionListener<>() {
                 @Override
                 public void onResponse(RankFeatureDoc[] docsWithUpdatedScores) {
-                    RankFeatureDoc[] topResults = rankFeaturePhaseRankCoordinatorContext.rankAndPaginate(docsWithUpdatedScores, true);
+                    RankDoc[] topResults = rankFeaturePhaseRankCoordinatorContext.rankAndPaginate(docsWithUpdatedScores, true);
                     SearchPhaseController.ReducedQueryPhase reducedRankFeaturePhase = newReducedQueryPhaseResults(
                         reducedQueryPhase,
                         topResults
@@ -200,13 +201,18 @@ public void onFailure(Exception e) {
                     if (rankFeaturePhaseRankCoordinatorContext.failuresAllowed()) {
                         // TODO: handle the exception somewhere
                         // don't want to log the entire stack trace, it's not helpful here
-                        logger.warn("Exception computing updated ranks: {}. Continuing with existing ranks.", e.toString());
+                        logger.warn("Exception computing updated ranks, continuing with existing ranks: {}", e.toString());
                         // use the existing score docs as-is
-                        RankFeatureDoc[] existingScores = Arrays.stream(reducedQueryPhase.sortedTopDocs().scoreDocs())
-                            .map(sd -> new RankFeatureDoc(sd.doc, sd.score, sd.shardIndex))
-                            .toArray(RankFeatureDoc[]::new);
-
-                        RankFeatureDoc[] topResults = rankFeaturePhaseRankCoordinatorContext.rankAndPaginate(existingScores, false);
+                        // downstream things expect every doc to have a score, so we need to infer a score here
+                        // if the doc doesn't otherwise have a score. We can use the rank.
+                        ScoreDoc[] inputDocs = reducedQueryPhase.sortedTopDocs().scoreDocs();
+                        // use RankDoc to indicate there was a problem using the specified features
+                        RankFeatureDoc[] rankDocs = new RankFeatureDoc[inputDocs.length];
+                        for (int i = 0; i < inputDocs.length; i++) {
+                            ScoreDoc doc = inputDocs[i];
+                            rankDocs[i] = new RankFeatureDoc(doc.doc, Float.isNaN(doc.score) ? 1f / (i+1) : doc.score, doc.shardIndex);
+                        }
+                        RankDoc[] topResults = rankFeaturePhaseRankCoordinatorContext.rankAndPaginate(rankDocs, false);
                         SearchPhaseController.ReducedQueryPhase reducedRankFeaturePhase = newReducedQueryPhaseResults(
                             reducedQueryPhase,
                             topResults

x-pack/plugin/rank-rrf/src/yamlRestTest/resources/rest-api-spec/test/rrf/800_rrf_with_text_similarity_reranker_retriever.yml

@@ -334,3 +334,83 @@ setup:
   - match: {hits.hits.0._explanation.details.1.description: "/rrf.score:.\\[0.5\\].*/" }
   - match: {hits.hits.0._explanation.details.1.details.0.description: "/text_similarity_reranker.match.using.inference.endpoint:.\\[my-rerank-model\\].on.document.field:.\\[text\\].*/" }
   - match: {hits.hits.0._explanation.details.1.details.0.details.0.description: "/weight.*astronomy.*/" }
+
+---
+"rrf retriever with failed text similarity reranker":
+
+  - do:
+      search:
+        index: test-index
+        body:
+          track_total_hits: true
+          fields: [ "text", "topic" ]
+          retriever:
+            rrf: {
+              retrievers:
+                [
+                  {
+                    standard: {
+                      query: {
+                        bool: {
+                          should:
+                            [
+                              {
+                                constant_score: {
+                                  filter: {
+                                    term: {
+                                      integer: 1
+                                    }
+                                  },
+                                  boost: 10
+                                }
+                              },
+                              {
+                                constant_score:
+                                  {
+                                    filter:
+                                      {
+                                        term:
+                                          {
+                                            integer: 2
+                                          }
+                                      },
+                                    boost: 1
+                                  }
+                              }
+                            ]
+                        }
+                      }
+                    }
+                  },
+                  {
+                    text_similarity_reranker: {
+                      retriever:
+                        {
+                          standard: {
+                            query: {
+                              match_all: {}
+                            },
+                            sort: {
+                              integer: "asc"
+                            }
+                          }
+                        },
+                      rank_window_size: 10,
+                      inference_id: failure-rerank-model,
+                      inference_text: "How often does the moon hide the sun?",
+                      field: text,
+                      allow_rerank_failures: true
+                    }
+                  }
+                ],
+              rank_window_size: 10,
+              rank_constant: 1
+            }
+          size: 10
+
+  - match: { hits.total.value: 3 }
+  - length: { hits.hits: 3 }
+
+  - match: { hits.hits.0._id: "doc_1" }
+  - match: { hits.hits.1._id: "doc_2" }
+  - match: { hits.hits.2._id: "doc_3" }