Merge pull request #29425 from vespa-engine/toregge/check-for-valid-successor

Check for valid successor.

2 files changed, 43 insertions, 28 deletions

diff --git a/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.cpp b/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.cpp
index 040bd9ccc98..9d6e2e9815d 100644
--- a/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.cpp
+++ b/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.cpp
@@ -1,6 +1,7 @@
 // Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
 
 #include "dfa_fuzzy_matcher.h"
+#include "i_enum_store_dictionary.h"
 #include <vespa/vespalib/text/utf8.h>
 #include <vespa/vespalib/text/lowercase.h>
 
@@ -94,4 +95,43 @@ DfaFuzzyMatcher::is_match(const char* word) const
     return match.matches();
 }
 
+template <typename DictionaryConstIteratorType>
+bool
+DfaFuzzyMatcher::is_match(const char* word, DictionaryConstIteratorType& itr, const DfaStringComparator::DataStoreType& data_store) {
+    if (_prefix_size > 0) {
+        word = skip_prefix(word);
+        if (_prefix.size() < _prefix_size) {
+            if (*word == '\0') {
+                return true;
+            }
+            _successor.resize(_prefix.size());
+            _successor.emplace_back(beyond_unicode);
+        } else {
+            _successor.resize(_prefix.size());
+            auto match = _dfa.match(word, _successor);
+            if (match.matches()) {
+                return true;
+            }
+        }
+    } else {
+        _successor.clear();
+        auto match = _dfa.match(word, _successor);
+        if (match.matches()) {
+            return true;
+        }
+    }
+    DfaStringComparator cmp(data_store, _successor, _cased);
+    assert(cmp.less(itr.getKey().load_acquire(), vespalib::datastore::EntryRef()));
+    itr.seek(vespalib::datastore::AtomicEntryRef(), cmp);
+    return false;
+}
+
+template
+bool
+DfaFuzzyMatcher::is_match(const char* word, EnumPostingTree::ConstIterator& itr, const DfaStringComparator::DataStoreType& data_store);
+
+template
+bool
+DfaFuzzyMatcher::is_match(const char* word, EnumTree::ConstIterator& itr, const DfaStringComparator::DataStoreType& data_store);
+
 }
diff --git a/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.h b/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.h
index 9de77085fae..51457129637 100644
--- a/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.h
+++ b/searchlib/src/vespa/searchlib/attribute/dfa_fuzzy_matcher.h
@@ -5,7 +5,6 @@
 #include "dfa_string_comparator.h"
 #include <vespa/vespalib/datastore/atomic_entry_ref.h>
 #include <vespa/vespalib/fuzzy/levenshtein_dfa.h>
-#include <iostream>
 
 namespace search::attribute {
 
@@ -23,6 +22,8 @@ private:
     uint32_t                        _prefix_size;
     bool                            _cased;
 
+    static constexpr uint32_t beyond_unicode = 0x110000;
+
     const char* skip_prefix(const char* word) const;
 public:
     DfaFuzzyMatcher(std::string_view target, uint8_t max_edits, uint32_t prefix_size, bool cased, vespalib::fuzzy::LevenshteinDfa::DfaType dfa_type);
@@ -39,33 +40,7 @@ public:
      * functionality in the dictionary.
      */
     template <typename DictionaryConstIteratorType>
-    bool is_match(const char* word, DictionaryConstIteratorType& itr, const DfaStringComparator::DataStoreType& data_store) {
-        if (_prefix_size > 0) {
-            word = skip_prefix(word);
-            if (_prefix.size() < _prefix_size) {
-                if (*word == '\0') {
-                    return true;
-                }
-                _successor.resize(_prefix.size());
-                _successor.emplace_back(1);
-            } else {
-                _successor.resize(_prefix.size());
-                auto match = _dfa.match(word, _successor);
-                if (match.matches()) {
-                    return true;
-                }
-            }
-        } else {
-            _successor.clear();
-            auto match = _dfa.match(word, _successor);
-            if (match.matches()) {
-                return true;
-            }
-        }
-        DfaStringComparator cmp(data_store, _successor, _cased);
-        itr.seek(vespalib::datastore::AtomicEntryRef(), cmp);
-        return false;
-    }
+    bool is_match(const char* word, DictionaryConstIteratorType& itr, const DfaStringComparator::DataStoreType& data_store);
 };
 
 }