Merge pull request #3516 from opencloud-eu/fix/search-kql-path-normalization

fix(search): normalize KQL path once

services/search/pkg/opensearch/internal/convert/kql_transpile.go

@@ -119,9 +119,6 @@ func (t kqlOpensearchTranspiler) toBuilder(node ast.Node) (osu.Builder, error) {
 		}
 
 		field, value := node.Key, node.Value
-		if query.FieldIsPath(node.Key) {
-			value = strings.TrimSuffix(value, "/")
-		}
 		if node.CaseInsensitive {
 			field += mapping.LowercaseSuffix
 			value = strings.ToLower(value)

services/search/pkg/parity/README.md

@@ -250,6 +250,15 @@ Fixtures:
 | PATH-06 | `path:"./DOCUMENTS"` | docs-upper | docs-upper | docs-upper | ✅ |
 | PATH-07 | `path:"./Documents"` | docs-mixed | docs-mixed | docs-mixed | ✅ |
 | PATH-08 | `path:"./parent/"` | child.jpg, parent | child.jpg, parent | child.jpg, parent | ✅ |
+| PATH-09 | `path:"/"` | child.jpg, docs-lower, docs-mixed, docs-upper, parent | child.jpg, docs-lower, docs-mixed, docs-upper, parent | child.jpg, docs-lower, docs-mixed, docs-upper, parent | ✅ |
+| PATH-10 | `path:""` | child.jpg, docs-lower, docs-mixed, docs-upper, parent | child.jpg, docs-lower, docs-mixed, docs-upper, parent | child.jpg, docs-lower, docs-mixed, docs-upper, parent | ✅ |
+| PATH-11 | `path:"/parent"` | child.jpg, parent | child.jpg, parent | child.jpg, parent | ✅ |
+| PATH-12 | `path:"parent"` | child.jpg, parent | child.jpg, parent | child.jpg, parent | ✅ |
+| PATH-13 | `path:"parent/*"` | child.jpg | child.jpg | child.jpg | ✅ |
+| PATH-14 | `path:"*/parent"` | child.jpg, parent | child.jpg, parent | child.jpg, parent | ✅ |
+| PATH-15 | `path:"?/parent"` | child.jpg, parent | child.jpg, parent | child.jpg, parent | ✅ |
+| PATH-16 | `path:"*/Documents"` | docs-mixed | docs-mixed | docs-mixed | ✅ |
+| PATH-17 | `path:"*/parent/"` | child.jpg, parent | child.jpg, parent | child.jpg, parent | ✅ |
 
 ### fields
 

services/search/pkg/parity/query_path_test.go

@@ -23,6 +23,15 @@ func pathGroup() queryGroup {
 			{id: 6, query: `path:"./DOCUMENTS"`, want: []string{"docs-upper"}},
 			{id: 7, query: `path:"./Documents"`, want: []string{"docs-mixed"}},
 			{id: 8, query: `path:"./parent/"`, want: []string{"parent", "child.jpg"}},
+			{id: 9, query: `path:"/"`, want: []string{"parent", "child.jpg", "docs-lower", "docs-upper", "docs-mixed"}},
+			{id: 10, query: `path:""`, want: []string{"parent", "child.jpg", "docs-lower", "docs-upper", "docs-mixed"}},
+			{id: 11, query: `path:"/parent"`, want: []string{"parent", "child.jpg"}},
+			{id: 12, query: `path:"parent"`, want: []string{"parent", "child.jpg"}},
+			{id: 13, query: `path:"parent/*"`, want: []string{"child.jpg"}},
+			{id: 14, query: `path:"*/parent"`, want: []string{"parent", "child.jpg"}},
+			{id: 15, query: `path:"?/parent"`, want: []string{"parent", "child.jpg"}},
+			{id: 16, query: `path:"*/Documents"`, want: []string{"docs-mixed"}},
+			{id: 17, query: `path:"*/parent/"`, want: []string{"parent", "child.jpg"}},
 		},
 	}
 }

services/search/pkg/query/bleve/compiler.go

@@ -95,9 +95,6 @@ func walk(offset int, nodes []ast.Node) (bleveQuery.Query, int, error) {
 			// bleve treats `/` and `+` as literals mid-term, so a literal MIME like
 			// image/svg+xml still matches exactly.
 			val := n.Value
-			if searchQuery.FieldIsPath(n.Key) {
-				val = strings.TrimSuffix(val, "/")
-			}
 			k := n.Key
 			v := val
 			if k != "ID" && k != "Size" && k != "MimeType" {

services/search/pkg/query/bleve/compiler_test.go

@@ -59,7 +59,7 @@ func Test_compile(t *testing.T) {
 				},
 			},
 			want: func() query.Query {
-				tq := query.NewTermQuery("/Foo")
+				tq := query.NewTermQuery("./Foo")
 				tq.SetField("Path")
 				return query.NewConjunctionQuery([]query.Query{tq})
 			}(),

services/search/pkg/query/normalize.go

@@ -5,6 +5,7 @@ import (
 	"strings"
 
 	provider "github.com/cs3org/go-cs3apis/cs3/storage/provider/v1beta1"
+	"github.com/opencloud-eu/reva/v2/pkg/utils"
 	"reflect"
 
 	"github.com/opencloud-eu/opencloud/pkg/ast"
@@ -36,6 +37,14 @@ func normalizeNodes(nodes []ast.Node, resolve func(string) string, defaultKey st
 		switch node := n.(type) {
 		case *ast.StringNode:
 			node.Key = resolveKey(node.Key)
+			if FieldIsPath(node.Key) {
+				if strings.IndexAny(node.Value, "*?") == 0 {
+					// a leading wildcard stands in for the "./" prefix, anchoring would break it
+					node.Value = strings.TrimSuffix(node.Value, "/")
+				} else {
+					node.Value = utils.MakeRelativePath(node.Value)
+				}
+			}
 			if FieldValueIsNormalized(node.Key) {
 				node.Value = strings.ToLower(node.Value)
 			}

services/search/pkg/query/normalize_test.go

@@ -127,6 +127,27 @@ var _ = Describe("Normalize", func() {
 		}))
 	})
 
+	DescribeTable("normalizes path values to the stored relative form",
+		func(in, want string) {
+			got := norm(&ast.StringNode{Key: "path", Value: in})
+			Expect(got).To(Equal([]ast.Node{&ast.StringNode{Key: "Path", Value: want}}))
+		},
+		Entry("relative", "./parent", "./parent"),
+		Entry("trailing slash", "./parent/", "./parent"),
+		Entry("absolute", "/parent", "./parent"),
+		Entry("bare", "parent", "./parent"),
+		Entry("nested bare", "parent/child.jpg", "./parent/child.jpg"),
+		Entry("root", "/", "."),
+		Entry("empty", "", "."),
+		Entry("dot", ".", "."),
+		Entry("wildcard", "./par*/", "./par*"),
+		Entry("bare wildcard", "parent/*", "./parent/*"),
+		Entry("leading wildcard", "*child*", "*child*"),
+		Entry("leading wildcard covers the prefix", "*/parent", "*/parent"),
+		Entry("leading single-char wildcard covers the prefix", "?/parent", "?/parent"),
+		Entry("leading wildcard trailing slash", "*/parent/", "*/parent"),
+	)
+
 	It("converts value nodes to pointers", func() {
 		got := norm(
 			ast.StringNode{Key: "name", Value: "x"},