fork of https://github.com/sourcegraph/zoekt
Add tests for file-based streaming strategy (#537)
Ensures the new "flush collect sender" is exercised in tests.
author
Julie Tibshirani
committer
GitHub
date
(Feb 17, 2023, 9:22 AM -0800)
commit
58cf4748
58cf4748830ac0eded1517cc8c2454694c531fbd
parent
e3a85ad2
e3a85ad21637d86c2e2a3f5701776349a1950a11
+70
-54
+70
-54
shards/shards_test.go
+70
-54
shards/shards_test.go
···
929
929
func TestFileBasedSearch(t *testing.T) {
930
930
cases := []struct {
931
931
name string
932
-
testShardedSearch func(t *testing.T, q query.Q, ib *zoekt.IndexBuilder) []zoekt.FileMatch
932
+
testShardedSearch func(t *testing.T, q query.Q, ib *zoekt.IndexBuilder, useDocumentRanks bool) []zoekt.FileMatch
933
933
}{
934
934
{"Search", testShardedSearch},
935
935
{"StreamSearch", testShardedStreamSearch},
···
945
945
)
946
946
947
947
for _, tt := range cases {
948
-
t.Run(tt.name, func(t *testing.T) {
949
-
matches := tt.testShardedSearch(t, &query.Substring{
950
-
CaseSensitive: false,
951
-
Pattern: "ananas",
952
-
}, b)
948
+
for _, useDocumentRanks := range []bool{false, true} {
949
+
t.Run(tt.name, func(t *testing.T) {
950
+
matches := tt.testShardedSearch(t, &query.Substring{
951
+
CaseSensitive: false,
952
+
Pattern: "ananas",
953
+
}, b, useDocumentRanks)
953
954
954
-
if len(matches) != 2 {
955
-
t.Fatalf("got %v, want 2 matches", matches)
956
-
}
957
-
if matches[0].FileName != "f2" || matches[1].FileName != "f1" {
958
-
t.Fatalf("got %v, want matches {f1,f2}", matches)
959
-
}
960
-
if matches[0].LineMatches[0].LineFragments[0].Offset != 10 || matches[1].LineMatches[0].LineFragments[0].Offset != 8 {
961
-
t.Fatalf("got %#v, want offsets 10,8", matches)
962
-
}
963
-
})
955
+
if len(matches) != 2 {
956
+
t.Fatalf("got %v, want 2 matches", matches)
957
+
}
958
+
if matches[0].FileName != "f2" || matches[1].FileName != "f1" {
959
+
t.Fatalf("got %v, want matches {f1,f2}", matches)
960
+
}
961
+
if matches[0].LineMatches[0].LineFragments[0].Offset != 10 || matches[1].LineMatches[0].LineFragments[0].Offset != 8 {
962
+
t.Fatalf("got %#v, want offsets 10,8", matches)
963
+
}
964
+
})
965
+
}
964
966
}
965
967
}
966
968
967
969
func TestWordBoundaryRanking(t *testing.T) {
968
970
cases := []struct {
969
971
name string
970
-
testShardedSearch func(t *testing.T, q query.Q, ib *zoekt.IndexBuilder) []zoekt.FileMatch
972
+
testShardedSearch func(t *testing.T, q query.Q, ib *zoekt.IndexBuilder, useDocumentRanks bool) []zoekt.FileMatch
971
973
}{
972
974
{"Search", testShardedSearch},
973
975
{"StreamSearch", testShardedStreamSearch},
···
980
982
zoekt.Document{Name: "f3", Content: []byte("xbytex ybytex")})
981
983
982
984
for _, tt := range cases {
983
-
t.Run(tt.name, func(t *testing.T) {
984
-
files := tt.testShardedSearch(t, &query.Substring{Pattern: "byte"}, b)
985
+
for _, useDocumentRanks := range []bool{false, true} {
986
+
t.Run(tt.name, func(t *testing.T) {
987
+
files := tt.testShardedSearch(t, &query.Substring{Pattern: "byte"}, b, useDocumentRanks)
985
988
986
-
if len(files) != 3 {
987
-
t.Fatalf("got %#v, want 3 files", files)
988
-
}
989
+
if len(files) != 3 {
990
+
t.Fatalf("got %#v, want 3 files", files)
991
+
}
989
992
990
-
file0 := files[0]
991
-
if file0.FileName != "f2" || len(file0.LineMatches) != 3 {
992
-
t.Fatalf("got file %s, num matches %d (%#v), want 3 matches in file f2", file0.FileName, len(file0.LineMatches), file0)
993
-
}
993
+
file0 := files[0]
994
+
if file0.FileName != "f2" || len(file0.LineMatches) != 3 {
995
+
t.Fatalf("got file %s, num matches %d (%#v), want 3 matches in file f2", file0.FileName, len(file0.LineMatches), file0)
996
+
}
994
997
995
-
if file0.LineMatches[0].LineFragments[0].Offset != 13 {
996
-
t.Fatalf("got first match %#v, want full word match", files[0].LineMatches[0])
997
-
}
998
-
if file0.LineMatches[1].LineFragments[0].Offset != 7 {
999
-
t.Fatalf("got second match %#v, want partial word match", files[0].LineMatches[1])
1000
-
}
1001
-
})
998
+
if file0.LineMatches[0].LineFragments[0].Offset != 13 {
999
+
t.Fatalf("got first match %#v, want full word match", files[0].LineMatches[0])
1000
+
}
1001
+
if file0.LineMatches[1].LineFragments[0].Offset != 7 {
1002
+
t.Fatalf("got second match %#v, want partial word match", files[0].LineMatches[1])
1003
+
}
1004
+
})
1005
+
}
1002
1006
}
1003
1007
}
1004
1008
1005
1009
func TestAtomCountScore(t *testing.T) {
1006
1010
cases := []struct {
1007
1011
name string
1008
-
testShardedSearch func(t *testing.T, q query.Q, ib *zoekt.IndexBuilder) []zoekt.FileMatch
1012
+
testShardedSearch func(t *testing.T, q query.Q, ib *zoekt.IndexBuilder, useDocumentRanks bool) []zoekt.FileMatch
1009
1013
}{
1010
1014
{"Search", testShardedSearch},
1011
1015
{"StreamSearch", testShardedStreamSearch},
···
1023
1027
zoekt.Document{Name: "needle-file", Content: []byte("needle content"), Branches: []string{"branches"}})
1024
1028
1025
1029
for _, tt := range cases {
1026
-
t.Run(tt.name, func(t *testing.T) {
1027
-
files := tt.testShardedSearch(t,
1028
-
query.NewOr(
1029
-
&query.Substring{Pattern: "needle"},
1030
-
&query.Substring{Pattern: "needle", FileName: true},
1031
-
&query.Branch{Pattern: "needle"},
1032
-
), b)
1033
-
var got []string
1034
-
for _, f := range files {
1035
-
got = append(got, f.FileName)
1036
-
}
1037
-
want := []string{"needle-file-branch", "needle-file", "f1"}
1038
-
if !reflect.DeepEqual(got, want) {
1039
-
t.Errorf("got %v, want %v", got, want)
1040
-
}
1041
-
})
1030
+
for _, useDocumentRanks := range []bool{false, true} {
1031
+
t.Run(tt.name, func(t *testing.T) {
1032
+
files := tt.testShardedSearch(t,
1033
+
query.NewOr(
1034
+
&query.Substring{Pattern: "needle"},
1035
+
&query.Substring{Pattern: "needle", FileName: true},
1036
+
&query.Branch{Pattern: "needle"},
1037
+
), b, useDocumentRanks)
1038
+
var got []string
1039
+
for _, f := range files {
1040
+
got = append(got, f.FileName)
1041
+
}
1042
+
want := []string{"needle-file-branch", "needle-file", "f1"}
1043
+
if !reflect.DeepEqual(got, want) {
1044
+
t.Errorf("got %v, want %v", got, want)
1045
+
}
1046
+
})
1047
+
}
1042
1048
}
1043
1049
}
1044
1050
1045
-
func testShardedStreamSearch(t *testing.T, q query.Q, ib *zoekt.IndexBuilder) []zoekt.FileMatch {
1051
+
func testShardedStreamSearch(t *testing.T, q query.Q, ib *zoekt.IndexBuilder, useDocumentRanks bool) []zoekt.FileMatch {
1046
1052
ss := newShardedSearcher(1)
1047
1053
searcher := searcherForTest(t, ib)
1048
1054
ss.replace(map[string]zoekt.Searcher{"r1": searcher})
···
1052
1058
files = append(files, result.Files...)
1053
1059
})
1054
1060
1055
-
if err := ss.StreamSearch(context.Background(), q, &zoekt.SearchOptions{}, sender); err != nil {
1061
+
opts := zoekt.SearchOptions{}
1062
+
if useDocumentRanks {
1063
+
opts.UseDocumentRanks = true
1064
+
opts.FlushWallTime = 10 * time.Millisecond
1065
+
}
1066
+
if err := ss.StreamSearch(context.Background(), q, &opts, sender); err != nil {
1056
1067
t.Fatal(err)
1057
1068
}
1058
1069
return files
1059
1070
}
1060
1071
1061
-
func testShardedSearch(t *testing.T, q query.Q, ib *zoekt.IndexBuilder) []zoekt.FileMatch {
1072
+
func testShardedSearch(t *testing.T, q query.Q, ib *zoekt.IndexBuilder, useDocumentRanks bool) []zoekt.FileMatch {
1062
1073
ss := newShardedSearcher(1)
1063
1074
searcher := searcherForTest(t, ib)
1064
1075
ss.replace(map[string]zoekt.Searcher{"r1": searcher})
1065
1076
1066
-
sres, _ := ss.Search(context.Background(), q, &zoekt.SearchOptions{})
1077
+
opts := zoekt.SearchOptions{}
1078
+
if useDocumentRanks {
1079
+
opts.UseDocumentRanks = true
1080
+
opts.FlushWallTime = 50 * time.Millisecond
1081
+
}
1082
+
sres, _ := ss.Search(context.Background(), q, &opts)
1067
1083
return sres.Files
1068
1084
}
1069
1085