SoXX
20fcad7fbb
FA renders its species/character/artist/type system tags as tag-block anchors with a data-tag-name carrying a single-letter prefix (s_/c_/a_-u_/t_) and a sibling tag-invalid span instead of a /search/ link. The existing keyword pass skips them, so they were lost. Adds a Submission.CategorizedTags field exposing the four buckets with the prefix stripped, plus an examples/categorized_tags runnable demo.
221 lines
8.2 KiB
Go
221 lines
8.2 KiB
Go
package fa
|
||
|
||
import (
|
||
"bytes"
|
||
"fmt"
|
||
"strings"
|
||
"testing"
|
||
|
||
"github.com/PuerkitoBio/goquery"
|
||
)
|
||
|
||
// syntheticSubmissionHTML is a minimal hand-rolled page that exercises every
|
||
// selector the parser cares about. Real FA HTML differs in ways that
|
||
// fixture-driven tests will catch; this synthetic input pins the parser
|
||
// against a stable, deterministic input independent of FA's mood.
|
||
const syntheticSubmissionHTML = `<html><body>
|
||
<meta property="og:url" content="https://www.furaffinity.net/view/1234/"/>
|
||
<section class="submission-description">
|
||
<div class="submission-description-header">
|
||
<div class="submission-description-artist">
|
||
<div><a href="/user/somefurry/"><img class="submission-user-icon avatar" src="//d.example/avatars/somefurry.png"/></a></div>
|
||
<div>
|
||
<div class="submission-title"><h2>My Test Submission</h2></div>
|
||
<div>by <span class="c-usernameBlockSimple"><a href="/user/somefurry/"><span class="c-usernameBlockSimple__displayName" title="somefurry">SomeFurry</span></a></span></div>
|
||
<div>Posted <span class="popup_date" title="March 17, 2026 04:21:21 PM">5 hours ago</span></div>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
<div class="section-body">
|
||
<div class="submission-description-text"><p>Hello <b>world</b>.</p></div>
|
||
</div>
|
||
</section>
|
||
|
||
<img id="submissionImg" src="//d.example/art/somefurry/1234.png" data-fullview-src="//d.example/art/somefurry/1234_full.png" data-preview-src="//d.example/art/somefurry/1234_thumb.png" data-fullview-width="1920" data-fullview-height="1080"/>
|
||
|
||
<div class="submission-page-stats">
|
||
<div title="Views"><div>1,234</div><div class="highlight">Views</div></div>
|
||
<div title="Comments"><div>7</div><div class="highlight">Comments</div></div>
|
||
<div title="Favorites"><div>56</div><div class="highlight">Favorites</div></div>
|
||
<div><div class="font-large inline c-contentRating--general">General</div><div class="highlight">Rating</div></div>
|
||
</div>
|
||
|
||
<div class="submission-content-stats">
|
||
<span class="highlight">
|
||
<span>Category</span><span>Theme</span><span>Species</span><span>Gender</span><span>Resolution</span>
|
||
</span>
|
||
<span>
|
||
<span>Artwork (Digital)</span><span>Digital</span><span>Wolf</span><span>Male</span><span>1920 x 1080</span>
|
||
</span>
|
||
</div>
|
||
|
||
<div class="submission-tags">
|
||
<div class="highlight">Keywords</div>
|
||
<div>
|
||
<span class="tags"><span><a href="javascript:void(0);" class="tag-block"></a><a href="/search/@keywords wolf">wolf</a></span></span>
|
||
<span class="tags"><span><a href="javascript:void(0);" class="tag-block"></a><a href="/search/@keywords art">art</a></span></span>
|
||
<span class="tags"><span><a href="javascript:void(0);" data-tag-name="s_wolf" class="tag-block"></a><span class="tag-invalid">s_wolf</span></span></span>
|
||
<span class="tags"><span><a href="javascript:void(0);" data-tag-name="c_artwork_digital" class="tag-block"></a><span class="tag-invalid">c_artwork_digital</span></span></span>
|
||
<span class="tags"><span><a href="javascript:void(0);" data-tag-name="t_general_furry_art" class="tag-block"></a><span class="tag-invalid">t_general_furry_art</span></span></span>
|
||
<span class="tags"><span><a href="javascript:void(0);" data-tag-name="u_somefurry" class="tag-block"></a><span class="tag-invalid">u_somefurry</span></span></span>
|
||
</div>
|
||
</div>
|
||
|
||
<div class="minigallery-navigation">
|
||
<a href="/view/1235/">« Newer</a>
|
||
<a href="/view/1233/">Older »</a>
|
||
</div>
|
||
</body></html>`
|
||
|
||
func TestParseSubmission_Synthetic(t *testing.T) {
|
||
doc, err := goquery.NewDocumentFromReader(strings.NewReader(syntheticSubmissionHTML))
|
||
if err != nil {
|
||
t.Fatalf("setup: %v", err)
|
||
}
|
||
sub, err := parseSubmission(1234, doc)
|
||
if err != nil {
|
||
t.Fatalf("parseSubmission: %v", err)
|
||
}
|
||
|
||
checks := []struct {
|
||
name string
|
||
got any
|
||
want any
|
||
}{
|
||
{"ID", sub.ID, SubmissionID(1234)},
|
||
{"Title", sub.Title, "My Test Submission"},
|
||
{"Author.Name", sub.Author.Name, "somefurry"},
|
||
{"Author.DisplayName", sub.Author.DisplayName, "SomeFurry"},
|
||
{"Author.AvatarURL", sub.Author.AvatarURL, "https://d.example/avatars/somefurry.png"},
|
||
{"Rating", sub.Rating, RatingGeneral},
|
||
{"Category", sub.Category, Category("Artwork (Digital)")},
|
||
{"Type", sub.Type, Type("Digital")},
|
||
{"Species", sub.Species, Species("Wolf")},
|
||
{"Gender", sub.Gender, Gender("Male")},
|
||
{"FileURL", sub.FileURL, "https://d.example/art/somefurry/1234_full.png"},
|
||
{"ThumbURL", sub.ThumbURL, "https://d.example/art/somefurry/1234_thumb.png"},
|
||
{"Width", sub.Width, 1920},
|
||
{"Height", sub.Height, 1080},
|
||
{"Stats.Views", sub.Stats.Views, 1234},
|
||
{"Stats.Favorites", sub.Stats.Favorites, 56},
|
||
{"Stats.Comments", sub.Stats.Comments, 7},
|
||
{"Prev (Newer)", sub.Prev, SubmissionID(1235)},
|
||
{"Next (Older)", sub.Next, SubmissionID(1233)},
|
||
{"len(Tags)", len(sub.Tags), 2},
|
||
}
|
||
for _, c := range checks {
|
||
if c.got != c.want {
|
||
t.Errorf("%s = %v; want %v", c.name, c.got, c.want)
|
||
}
|
||
}
|
||
if !sub.PostedAt.IsZero() && sub.PostedAt.Year() != 2026 {
|
||
t.Errorf("PostedAt year = %d; want 2026", sub.PostedAt.Year())
|
||
}
|
||
if !strings.Contains(sub.Description, "world") {
|
||
t.Errorf("Description missing expected content: %q", sub.Description)
|
||
}
|
||
|
||
catChecks := []struct {
|
||
name string
|
||
got []string
|
||
want []string
|
||
}{
|
||
{"Species", sub.CategorizedTags.Species, []string{"wolf"}},
|
||
{"Characters", sub.CategorizedTags.Characters, []string{"artwork_digital"}},
|
||
{"Types", sub.CategorizedTags.Types, []string{"general_furry_art"}},
|
||
{"Artists", sub.CategorizedTags.Artists, []string{"somefurry"}},
|
||
}
|
||
for _, c := range catChecks {
|
||
if !equalStrings(c.got, c.want) {
|
||
t.Errorf("CategorizedTags.%s = %v; want %v", c.name, c.got, c.want)
|
||
}
|
||
}
|
||
}
|
||
|
||
func equalStrings(a, b []string) bool {
|
||
if len(a) != len(b) {
|
||
return false
|
||
}
|
||
for i := range a {
|
||
if a[i] != b[i] {
|
||
return false
|
||
}
|
||
}
|
||
return true
|
||
}
|
||
|
||
// TestParseSubmission_FavoritedState verifies parseSubmission reports the
|
||
// authenticated viewer's favorite state. FA renders exactly one of the
|
||
// "+Fav" (/fav/) or "−Fav" (/unfav/) anchors, matching the viewer's current
|
||
// state; an anonymous fetch shows neither.
|
||
func TestParseSubmission_FavoritedState(t *testing.T) {
|
||
const tmpl = `<html><body>
|
||
<meta property="og:url" content="https://www.furaffinity.net/view/1234/"/>
|
||
<section class="submission-description"><div class="submission-description-header">
|
||
<div class="submission-description-artist"><div></div>
|
||
<div><div class="submission-title"><h2>T</h2></div></div></div></div></section>
|
||
<div id="submission-options">%s</div>
|
||
</body></html>`
|
||
|
||
cases := []struct {
|
||
name string
|
||
link string
|
||
want bool
|
||
}{
|
||
{"favorited shows unfav link", `<a href="/unfav/1234/?key=abc">−Fav</a>`, true},
|
||
{"not favorited shows fav link", `<a href="/fav/1234/?key=abc">+Fav</a>`, false},
|
||
{"anonymous shows neither", ``, false},
|
||
}
|
||
for _, c := range cases {
|
||
t.Run(c.name, func(t *testing.T) {
|
||
doc, err := goquery.NewDocumentFromReader(strings.NewReader(fmt.Sprintf(tmpl, c.link)))
|
||
if err != nil {
|
||
t.Fatalf("setup: %v", err)
|
||
}
|
||
sub, err := parseSubmission(1234, doc)
|
||
if err != nil {
|
||
t.Fatalf("parseSubmission: %v", err)
|
||
}
|
||
if sub.Favorited != c.want {
|
||
t.Errorf("Favorited = %v; want %v", sub.Favorited, c.want)
|
||
}
|
||
})
|
||
}
|
||
}
|
||
|
||
func TestParseSubmission_MissingTitleErrors(t *testing.T) {
|
||
doc, err := goquery.NewDocumentFromReader(strings.NewReader("<html><body></body></html>"))
|
||
if err != nil {
|
||
t.Fatalf("setup: %v", err)
|
||
}
|
||
if _, err := parseSubmission(1, doc); err == nil {
|
||
t.Fatal("expected parse error for missing title")
|
||
}
|
||
}
|
||
|
||
// TestParseSubmission_RealFixture runs the parser against a real FA HTML
|
||
// dump captured by the `fixtures` build tag. Skips cleanly if no fixture
|
||
// has been recorded.
|
||
func TestParseSubmission_RealFixture(t *testing.T) {
|
||
raw := loadFixture(t, "submission.html")
|
||
doc, err := goquery.NewDocumentFromReader(bytes.NewReader(raw))
|
||
if err != nil {
|
||
t.Fatalf("read doc: %v", err)
|
||
}
|
||
sub, err := parseSubmission(0, doc)
|
||
if err != nil {
|
||
t.Fatalf("parseSubmission(real): %v", err)
|
||
}
|
||
// We can't assert exact values against a fixture whose contents we don't
|
||
// pin in this repo. Instead assert that the load-bearing fields populated.
|
||
if sub.Title == "" {
|
||
t.Error("real fixture: Title is empty")
|
||
}
|
||
if sub.Author.Name == "" {
|
||
t.Error("real fixture: Author.Name is empty")
|
||
}
|
||
if sub.FileURL == "" {
|
||
t.Error("real fixture: FileURL is empty")
|
||
}
|
||
}
|