Files
upstream-mirror 52ae0c5802
Some checks failed
cache-seeder / prune (push) Has been skipped
cache-seeder / frontend (push) Failing after 3s
cache-seeder / python (push) Failing after 3s
cache-seeder / gobuild (push) Failing after 4s
cache-seeder / lint (bindata, lint-backend) (push) Failing after 3s
giteabot backport / giteabot (push) Has been skipped
giteabot / giteabot (push) Has been skipped
cache-prune / prune (push) Has been skipped
release-nightly-snapcraft / build-and-publish (ubuntu-24.04) (push) Has been cancelled
release-nightly-snapcraft / build-and-publish (ubuntu-24.04-arm) (push) Has been cancelled
release-nightly / nightly-binary (push) Has been cancelled
release-nightly / nightly-container (push) Has been cancelled
cron-renovate / cron-renovate (push) Has been skipped
snapshot of https://github.com/go-gitea/gitea.git (main, 2026-09-15; blobs > 131072B omitted)
2026-09-15 21:41:59 +05:00

77 lines
1.6 KiB
Go

// Copyright 2024 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package path
import (
"fmt"
"testing"
"github.com/blevesearch/bleve/v2/analysis"
"github.com/blevesearch/bleve/v2/analysis/tokenizer/unicode"
"github.com/stretchr/testify/assert"
)
type Scenario struct {
Input string
Tokens []string
}
func TestTokenFilter(t *testing.T) {
scenarios := []struct {
Input string
Terms []string
}{
{
Input: "Dockerfile",
Terms: []string{"Dockerfile"},
},
{
Input: "Dockerfile.rootless",
Terms: []string{"Dockerfile.rootless"},
},
{
Input: "a/b/c/Dockerfile.rootless",
Terms: []string{"a", "a/b", "a/b/c", "a/b/c/Dockerfile.rootless", "Dockerfile.rootless", "Dockerfile.rootless/c", "Dockerfile.rootless/c/b", "Dockerfile.rootless/c/b/a"},
},
{
Input: "",
Terms: []string{},
},
}
for _, scenario := range scenarios {
t.Run(fmt.Sprintf("ensure terms of '%s'", scenario.Input), func(t *testing.T) {
terms := extractTerms(scenario.Input)
assert.Len(t, terms, len(scenario.Terms))
for _, term := range terms {
assert.Contains(t, scenario.Terms, term)
}
})
}
}
func extractTerms(input string) []string {
tokens := tokenize(input)
filteredTokens := filter(tokens)
terms := make([]string, 0, len(filteredTokens))
for _, token := range filteredTokens {
terms = append(terms, string(token.Term))
}
return terms
}
func filter(input analysis.TokenStream) analysis.TokenStream {
filter := NewTokenFilter()
return filter.Filter(input)
}
func tokenize(input string) analysis.TokenStream {
tokenizer := unicode.NewUnicodeTokenizer()
return tokenizer.Tokenize([]byte(input))
}