Files
dobotai-platform/internal/outboundimage/rewriter_test.go

378 lines
14 KiB
Go

package outboundimage
import (
"context"
"errors"
"strings"
"testing"
)
type fakeFetcher struct {
calls map[string]int
fail map[string]bool
}
func (f *fakeFetcher) Fetch(_ context.Context, ref string) ([]byte, string, error) {
if f.calls == nil {
f.calls = map[string]int{}
}
f.calls[ref]++
if f.fail[ref] {
return nil, "", errors.New("fetch failed")
}
return []byte("image:" + ref), "image/png", nil
}
type fakePublisher struct {
calls int
}
func (p *fakePublisher) Publish(_ context.Context, data []byte, contentType string) (string, error) {
p.calls++
if contentType != "image/png" || len(data) == 0 {
return "", errors.New("bad image")
}
return "https://img.example/" + string(rune('0'+p.calls)) + ".png?sig=short", nil
}
func TestRewriteMarkdownResourceImagesDeduplicatesAndPreservesOtherURLs(t *testing.T) {
fetcher := &fakeFetcher{}
publisher := &fakePublisher{}
r := NewRewriter(fetcher, publisher, 5)
input := "步骤一\n![图1](resource://AbCdEfGhIjKlMnOpQrStUv)\n重复\n![图1重复](resource://AbCdEfGhIjKlMnOpQrStUv)\n外部\n![官网](https://example.com/a.png)"
got := r.Rewrite(context.Background(), input)
if strings.Contains(got, "resource://") {
t.Fatalf("resource reference remained: %s", got)
}
if strings.Count(got, "https://img.example/1.png?sig=short") != 2 {
t.Fatalf("deduplicated signed URL count mismatch: %s", got)
}
if !strings.Contains(got, "![官网](https://example.com/a.png)") {
t.Fatalf("ordinary HTTPS image changed: %s", got)
}
if fetcher.calls["resource://AbCdEfGhIjKlMnOpQrStUv"] != 1 || publisher.calls != 1 {
t.Fatalf("calls fetch=%v publish=%d", fetcher.calls, publisher.calls)
}
}
func TestRewriteMarkdownResourceImagesDropsFailuresAndImagesBeyondLimit(t *testing.T) {
failed := "resource://FailedRefAbCdEfGhIjKlMn"
fetcher := &fakeFetcher{fail: map[string]bool{failed: true}}
publisher := &fakePublisher{}
r := NewRewriter(fetcher, publisher, 3)
input := "保留正文\n![坏图](" + failed + ")\n![一](resource://FirstRefAbCdEfGhIjKlMnO)\n![二](resource://SecondRefAbCdEfGhIjKlMn)\n![三](resource://ThirdRefAbCdEfGhIjKlMnO)"
got := r.Rewrite(context.Background(), input)
if !strings.Contains(got, "保留正文") {
t.Fatalf("text was lost: %s", got)
}
if strings.Contains(got, "resource://") || strings.Contains(got, "坏图") || strings.Contains(got, "![三]") {
t.Fatalf("broken or over-limit image remained: %s", got)
}
if publisher.calls != 2 {
t.Fatalf("publisher calls=%d, want 2", publisher.calls)
}
}
func TestRewriteHTMLResourceImagesAsDingTalkMarkdown(t *testing.T) {
resource := "resource://AbCdEfGhIjKlMnOpQrStUv"
fetcher := &fakeFetcher{}
publisher := &fakePublisher{}
r := NewRewriter(fetcher, publisher, 5)
input := "前文\n<img alt=\"步骤图\" width=\"100\" src=\"" + resource + "\"/>\n" +
"重复\n<img src='" + resource + "' class='diagram'>\n后文"
got := r.Rewrite(context.Background(), input)
if strings.Contains(got, "resource://") || strings.Contains(strings.ToLower(got), "<img") {
t.Fatalf("internal HTML image remained: %s", got)
}
if !strings.Contains(got, "![步骤图](https://img.example/1.png?sig=short)") {
t.Fatalf("alt image not normalized: %s", got)
}
if !strings.Contains(got, "![图片](https://img.example/1.png?sig=short)") {
t.Fatalf("default-alt image not normalized: %s", got)
}
if fetcher.calls[resource] != 1 || publisher.calls != 1 {
t.Fatalf("calls fetch=%v publish=%d", fetcher.calls, publisher.calls)
}
}
func TestSanitizePartialRemovesInternalImageNodesAndBareReferences(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := "正文\n![图](resource://AbCdEfGhIjKlMnOpQrStUv)\n" +
"<img alt='图' src='resource://SecondRefAbCdEfGhIjKlMn'>\n" +
"内部引用 resource://BareRefAbCdEfGhIjKlMn"
got := r.SanitizePartial(input)
if strings.Contains(got, "resource://") || strings.Contains(strings.ToLower(got), "<img") {
t.Fatalf("partial content leaked an internal image: %s", got)
}
if !strings.Contains(got, "正文") || !strings.Contains(got, "内部引用") {
t.Fatalf("partial text was lost: %s", got)
}
}
func TestRewriteFailsClosedOnUnquotedHTMLAndBareResourceReferences(t *testing.T) {
fetcher := &fakeFetcher{}
r := NewRewriter(fetcher, &fakePublisher{}, 5)
input := "正文\n<img src=resource://UnsafeRefAbCdEfGhIjKlMn>\n内部引用 resource://BareRefAbCdEfGhIjKlMn"
got := r.Rewrite(context.Background(), input)
if strings.Contains(got, "resource://") || strings.Contains(strings.ToLower(got), "<img") {
t.Fatalf("final content leaked an internal reference: %s", got)
}
if !strings.Contains(got, "正文") || !strings.Contains(got, "内部引用") {
t.Fatalf("final text was lost: %s", got)
}
if len(fetcher.calls) != 0 {
t.Fatalf("unsafe references must not be fetched: %v", fetcher.calls)
}
}
func TestRewriteRemovesUnsafeImageLocatorsAndSignedURLs(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := strings.Join([]string{
"正文",
"![local](local://bucket/a.png)",
"![oss](oss://private/a.png)",
"![private](https://127.0.0.1/a.png)",
"![signed](https://bucket.oss-cn-shenzhen.aliyuncs.com/a.png?x-oss-signature=secret)",
"![public](https://example.com/a.png)",
}, "\n")
got := r.Rewrite(context.Background(), input)
for _, forbidden := range []string{"local://", "oss://", "127.0.0.1", "x-oss-signature"} {
if strings.Contains(strings.ToLower(got), forbidden) {
t.Fatalf("final content retained %q: %s", forbidden, got)
}
}
if !strings.Contains(got, "![public](https://example.com/a.png)") {
t.Fatalf("public HTTPS image was removed: %s", got)
}
}
func TestRewriteNormalizesPublicHTMLImageAsDingTalkMarkdown(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := `<img class="diagram" src="https://example.com/a.png" alt="公网图" width="100">`
got := r.Rewrite(context.Background(), input)
if got != "![公网图](https://example.com/a.png)" {
t.Fatalf("normalized public HTML image=%q", got)
}
}
func TestRewriteRemovesDataImageWithOptionalTitle(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := "正文\n![占位图](data:image/png;base64,iVBORw0KGgo= \"\")\n后文"
got := r.Rewrite(context.Background(), input)
if strings.Contains(got, "data:image") || strings.Contains(got, "占位图") {
t.Fatalf("final content retained a data image: %s", got)
}
if !strings.Contains(got, "正文") || !strings.Contains(got, "后文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestSanitizePartialRemovesDataImageWithOptionalTitle(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := "正文\n![占位图](data:image/png;base64,iVBORw0KGgo= \"\")\n后文"
got := r.SanitizePartial(input)
if strings.Contains(got, "data:image") || strings.Contains(got, "占位图") {
t.Fatalf("partial content retained a data image: %s", got)
}
if !strings.Contains(got, "正文") || !strings.Contains(got, "后文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestRewritePublishesResourceImageWithOptionalTitle(t *testing.T) {
resource := "resource://AbCdEfGhIjKlMnOpQrStUv"
fetcher := &fakeFetcher{}
publisher := &fakePublisher{}
r := NewRewriter(fetcher, publisher, 5)
input := "![步骤图](" + resource + " \"安装说明\")"
got := r.Rewrite(context.Background(), input)
if got != "![步骤图](https://img.example/1.png?sig=short \"安装说明\")" {
t.Fatalf("resource image with title=%q", got)
}
if fetcher.calls[resource] != 1 || publisher.calls != 1 {
t.Fatalf("calls fetch=%v publish=%d", fetcher.calls, publisher.calls)
}
}
func TestSanitizePartialRemovesDataHTMLImage(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := `正文<img alt="占位图" src="data:image/png;base64,iVBORw0KGgo=">后文`
got := r.SanitizePartial(input)
if strings.Contains(got, "data:image") || strings.Contains(strings.ToLower(got), "<img") {
t.Fatalf("partial content retained a data HTML image: %s", got)
}
if !strings.Contains(got, "正文") || !strings.Contains(got, "后文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestRewritePreservesPublicHTTPSImageWithOptionalTitle(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := `![官网](https://example.com/a.png "产品图")`
if got := r.Rewrite(context.Background(), input); got != input {
t.Fatalf("final public HTTPS image=%q", got)
}
}
func TestSanitizePartialPreservesPublicHTTPSImageWithOptionalTitle(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := `![官网](https://example.com/a.png "产品图")`
if got := r.SanitizePartial(input); got != input {
t.Fatalf("partial public HTTPS image=%q", got)
}
}
func TestSanitizePartialRemovesIncompleteDataMarkdownImage(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := "正文\n![占位图](data:image/png;base64,iVBORw0K"
got := r.SanitizePartial(input)
if strings.Contains(got, "data:image") || strings.Contains(got, "占位图") {
t.Fatalf("partial content retained an incomplete data image: %s", got)
}
if !strings.Contains(got, "正文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestSanitizePartialRemovesIncompleteDataHTMLImage(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := `正文<img alt="占位图" src="data:image/png;base64,iVBORw0K`
got := r.SanitizePartial(input)
if strings.Contains(got, "data:image") || strings.Contains(strings.ToLower(got), "<img") {
t.Fatalf("partial content retained an incomplete data HTML image: %s", got)
}
if !strings.Contains(got, "正文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestRewriteRemovesBareDataImageReference(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := "正文 data:image/png;base64,iVBORw0KGgo= 后文"
got := r.Rewrite(context.Background(), input)
if strings.Contains(got, "data:image") {
t.Fatalf("final content retained a bare data image: %s", got)
}
if !strings.Contains(got, "正文") || !strings.Contains(got, "后文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestRewriteRemovesIncompleteDataMarkdownImage(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := "正文\n![占位图](data:image/png;base64,iVBORw0K"
got := r.Rewrite(context.Background(), input)
if strings.Contains(got, "data:image") || strings.Contains(got, "占位图") || strings.Contains(got, "![") {
t.Fatalf("final content retained a broken data image marker: %s", got)
}
if !strings.Contains(got, "正文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestRewriteWithoutPublisherRemovesDataImage(t *testing.T) {
r := NewRewriter(nil, nil, 0)
input := `正文 ![占位图](data:image/png;base64,iVBORw0KGgo= "") 后文`
got := r.Rewrite(context.Background(), input)
if strings.Contains(got, "data:image") || strings.Contains(got, "占位图") {
t.Fatalf("fail-closed rewrite retained a data image: %s", got)
}
if !strings.Contains(got, "正文") || !strings.Contains(got, "后文") {
t.Fatalf("surrounding text was lost: %s", got)
}
}
func TestRewriteDowngradesPDFImageToDocumentLink(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := `![机械臂示教与应用课件](https://sales-edu.example/06%20robot-teaching.PDF?tenant=demo "课件")`
want := `[机械臂示教与应用课件](https://sales-edu.example/06%20robot-teaching.PDF?tenant=demo "课件")`
if got := r.Rewrite(context.Background(), input); got != want {
t.Fatalf("PDF image rewrite=%q, want %q", got, want)
}
}
func TestFinalAndPartialDowngradeKnownDocumentImages(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
for _, extension := range []string{
".pdf", ".doc", ".docx", ".xls", ".xlsx", ".ppt", ".pptx",
".csv", ".txt", ".zip", ".rar", ".7z",
} {
t.Run(extension, func(t *testing.T) {
url := "https://docs.example/manual" + extension + "?tenant=demo"
input := "![资料](" + url + ")"
want := "[资料](" + url + ")"
if got := r.Rewrite(context.Background(), input); got != want {
t.Fatalf("final rewrite=%q, want %q", got, want)
}
if got := r.SanitizePartial(input); got != want {
t.Fatalf("partial rewrite=%q, want %q", got, want)
}
})
}
html := `<img src="https://docs.example/manual.xlsx" alt="参数表">`
if got, want := r.Rewrite(context.Background(), html), `[参数表](https://docs.example/manual.xlsx)`; got != want {
t.Fatalf("HTML document image rewrite=%q, want %q", got, want)
}
if got, want := r.SanitizePartial(html), `[参数表](https://docs.example/manual.xlsx)`; got != want {
t.Fatalf("partial HTML document image rewrite=%q, want %q", got, want)
}
}
func TestDocumentImageDowngradePreservesImageAndLinkContracts(t *testing.T) {
r := NewRewriter(&fakeFetcher{}, &fakePublisher{}, 5)
input := strings.Join([]string{
`![稳定图片](https://weknora-api.example/i/006d001d-7d0b-542e-b5ab-6ee7d7cdce54)`,
`[已有PDF链接](https://docs.example/manual.pdf)`,
`![危险PDF](https://127.0.0.1/manual.pdf)`,
}, "\n")
got := r.Rewrite(context.Background(), input)
if !strings.Contains(got, `![稳定图片](https://weknora-api.example/i/006d001d-7d0b-542e-b5ab-6ee7d7cdce54)`) {
t.Fatalf("extensionless stable image changed: %s", got)
}
if !strings.Contains(got, `[已有PDF链接](https://docs.example/manual.pdf)`) {
t.Fatalf("existing document link changed: %s", got)
}
if strings.Contains(got, "127.0.0.1") || strings.Contains(got, "危险PDF") {
t.Fatalf("unsafe document image became a link: %s", got)
}
}