fix(indexer): correct bleve indexer token filters (#38853)

* fix #36228
* fix #37221

---------

Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>

modules/indexer/issues/bleve/camelcase_keep_whole.go

@@ -0,0 +1,64 @@
+// Copyright 2026 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package bleve
+
+import (
+	"gitea.dev/modules/util"
+
+	"github.com/blevesearch/bleve/v2/analysis"
+	"github.com/blevesearch/bleve/v2/analysis/token/camelcase"
+	"github.com/blevesearch/bleve/v2/registry"
+)
+
+const camelCaseKeepWholeName = "camelCaseKeepWhole"
+
+// camelCaseKeepWholeFilter behaves like bleve's built-in "camelCase" token filter,
+// it also uses the whole word for a token. For example: when indexing "someThing",
+// CamelCaseFilter only emits "some" and "thing", this filter also emits "something".
+// It is questionable why the "issue indexer" used the CamelCaseFilter, it just can't search "someThing".
+// To avoid breaking existing user experiences, this "whole token filter" is introduced to make the full word can be searched.
+type camelCaseKeepWholeFilter struct {
+	inner *camelcase.CamelCaseFilter
+}
+
+func (f *camelCaseKeepWholeFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
+	// First, do exactly what the stock camelCase filter does: split by "camelCase" tokens
+	split := f.inner.Filter(input)
+
+	// Index the resulting position of the *first* sub-token produced for
+	// each original token (matched by start offset), so the duplicated
+	// whole-word token we add below lines up at the same position as the
+	// sub-word it stands in for, instead of drifting out of sync for
+	// fields with more than one original token.
+	posByStart := make(map[int]int, len(split))
+	for _, tok := range split {
+		if _, ok := posByStart[tok.Start]; !ok {
+			posByStart[tok.Start] = tok.Position
+		}
+	}
+
+	rv := make(analysis.TokenStream, 0, len(split)+len(input))
+	rv = append(rv, split...)
+
+	// Then append one extra, un-split copy of every original token, so the
+	// whole word survives as a standalone, independently searchable term.
+	for _, token := range input {
+		dup := *token
+		dup.Term = append([]byte(nil), token.Term...)
+		if pos, ok := posByStart[token.Start]; ok {
+			dup.Position = pos
+		}
+		rv = append(rv, &dup)
+	}
+
+	return rv
+}
+
+func camelCaseKeepWholeFilterConstructor(_ map[string]any, _ *registry.Cache) (analysis.TokenFilter, error) {
+	return &camelCaseKeepWholeFilter{inner: camelcase.NewCamelCaseFilter()}, nil
+}
+
+func init() {
+	util.MustNoError(registry.RegisterTokenFilter(camelCaseKeepWholeName, camelCaseKeepWholeFilterConstructor))
+}