+ {detail.status === 'reviewing' && detail.reviewer && (
+
+
+
+ 审核人
+
+
+
+
+ )}
) -> (&'static str, Vec) {
(primary, vec![primary.to_string()])
}
+/// 语言标签前缀
+const LANG_LABEL_PREFIX: &str = "语言";
+
+const LANGUAGE_CN_NAMES: &[(&str, &str)] = &[
+ ("日语", "ja"),
+ ("汉语", "zh"),
+ ("英语", "en"),
+ ("韩语", "ko"),
+ ("其它", "others"),
+ ("其他", "others"),
+];
+
+fn label_to_language(label: &str) -> Option<&'static str> {
+ let normalized = label.trim().replace('：', ":");
+ let mut segments = normalized
+ .split(':')
+ .map(str::trim)
+ .filter(|s| !s.is_empty());
+
+ if segments.next()? != LANG_LABEL_PREFIX {
+ return None;
+ }
+ let name = segments.next()?;
+ LANGUAGE_CN_NAMES
+ .iter()
+ .find(|(alias, _)| *alias == name)
+ .map(|(_, code)| *code)
+}
+
+/// 从 PR 标签提取语言代码：按标签顺序去重；命中具体语言时丢弃 `others`
+pub fn labels_language(labels: &[String]) -> Vec<&'static str> {
+ let mut codes: Vec<&'static str> = Vec::new();
+ for label in labels {
+ if let Some(code) = label_to_language(label)
+ && !codes.contains(&code)
+ {
+ codes.push(code);
+ }
+ }
+ if codes.iter().any(|code| *code != "others") {
+ codes.retain(|code| *code != "others");
+ }
+ codes
+}
+
+/// 过滤掉语言标签，剩下的作为投稿 tags
+pub fn non_language_labels(labels: &[String]) -> Vec {
+ labels
+ .iter()
+ .filter(|label| label_to_language(label).is_none())
+ .map(|label| label.trim().to_string())
+ .filter(|label| !label.is_empty())
+ .collect()
+}
+
+/// 判定投稿语言
+pub fn resolve_language(ttml: Option<&str>, labels: &[String]) -> (String, Vec) {
+ let from_labels = labels_language(labels);
+ if let Some(primary) = from_labels.first() {
+ let languages = from_labels.iter().map(|code| code.to_string()).collect();
+ return (primary.to_string(), languages);
+ }
+ let (primary, languages) = map_language(ttml);
+ (primary.to_string(), languages)
+}
+
/// 解析 TTML，返回站点 submissions 需要的元数据
pub struct ParsedTtmlMeta {
pub title: String,
@@ -202,7 +268,7 @@ pub struct ParsedTtmlMeta {
}
/// 解析 TTML 字节流（复用 ttml_processor，与 /validate 口径一致）
-pub fn parse_ttml_meta(bytes: &[u8]) -> Result {
+pub fn parse_ttml_meta(bytes: &[u8], labels: &[String]) -> Result {
use ttml_processor::model::PlatformId;
let text = String::from_utf8_lossy(bytes);
@@ -229,7 +295,7 @@ pub fn parse_ttml_meta(bytes: &[u8]) -> Result {
}
}
- let (language, languages) = map_language(meta.language.as_deref());
+ let (language, languages) = resolve_language(meta.language.as_deref(), labels);
Ok(ParsedTtmlMeta {
title: join(meta.title).chars().take(200).collect(),
@@ -239,7 +305,7 @@ pub fn parse_ttml_meta(bytes: &[u8]) -> Result {
qq_id,
am_id,
spotify_id,
- language: language.to_string(),
+ language,
languages,
ttml_author_github: meta
.author_ids
@@ -458,4 +524,111 @@ mod tests {
assert_eq!(map_language(None).0, "others");
assert_eq!(map_language(Some("fr")).0, "others");
}
+
+ #[test]
+ fn reads_language_from_pr_label() {
+ let labels = vec!["语言：日语".to_string()];
+ // 标签优先于 TTML metadata
+ let (primary, languages) = resolve_language(Some("zh-Hans"), &labels);
+ assert_eq!(primary, "ja");
+ assert_eq!(languages, vec!["ja".to_string()]);
+ }
+
+ #[test]
+ fn maps_all_current_language_labels() {
+ for (label, code) in [
+ ("语言：其它", "others"),
+ ("语言：日语", "ja"),
+ ("语言：汉语", "zh"),
+ ("语言：英语", "en"),
+ ("语言：韩语", "ko"),
+ // 半角冒号同样识别
+ ("语言:韩语", "ko"),
+ ] {
+ assert_eq!(
+ resolve_language(None, &[label.to_string()]).0,
+ code,
+ "标签 {} 应识别为 {}",
+ label,
+ code
+ );
+ }
+ }
+
+ #[test]
+ fn collects_multiple_label_languages_deduped() {
+ let labels = vec![
+ "语言：日语".to_string(),
+ "待更新".to_string(),
+ "语言：英语".to_string(),
+ "语言：日语".to_string(),
+ ];
+ let (primary, languages) = resolve_language(None, &labels);
+ assert_eq!(primary, "ja");
+ assert_eq!(languages, vec!["ja".to_string(), "en".to_string()]);
+ }
+
+ #[test]
+ fn explicit_others_label_wins_over_ttml() {
+ let labels = vec!["语言：其它".to_string()];
+ let (primary, languages) = resolve_language(Some("ja"), &labels);
+ assert_eq!(primary, "others");
+ assert_eq!(languages, vec!["others".to_string()]);
+ }
+
+ #[test]
+ fn drops_others_when_concrete_language_present() {
+ let labels = vec!["语言：其它".to_string(), "语言：韩语".to_string()];
+ let (primary, languages) = resolve_language(None, &labels);
+ assert_eq!(primary, "ko");
+ assert_eq!(languages, vec!["ko".to_string()]);
+ }
+
+ #[test]
+ fn falls_back_to_ttml_without_language_label() {
+ let labels = vec!["待更新".to_string(), "符合规范".to_string()];
+ let (primary, languages) = resolve_language(Some("en-US"), &labels);
+ assert_eq!(primary, "en");
+ assert_eq!(languages, vec!["en".to_string()]);
+ // 都没有 → others
+ assert_eq!(resolve_language(None, &labels).0, "others");
+ }
+
+ #[test]
+ fn strips_language_labels_from_tags() {
+ let labels = vec![
+ "语言：英语".to_string(),
+ "首次投稿".to_string(),
+ "语言：其它".to_string(),
+ "已完成".to_string(),
+ ];
+ assert_eq!(
+ non_language_labels(&labels),
+ vec!["首次投稿".to_string(), "已完成".to_string()]
+ );
+ // 非语言标签全部保留
+ let only_others = vec!["待更新".to_string(), "MSR".to_string()];
+ assert_eq!(non_language_labels(&only_others), only_others);
+ }
+
+ #[test]
+ fn ignores_labels_outside_whitelist() {
+ // 严格匹配：白名单外的标签一律忽略，不做中文名或子串推断
+ let ignored = [
+ "日语",
+ "语言",
+ "语言：法语",
+ "语言：日语（简体）",
+ "lang｜ja｜日语",
+ "lang｜fr｜法语",
+ ];
+ for label in ignored {
+ assert_eq!(
+ resolve_language(None, &[label.to_string()]).0,
+ "others",
+ "标签 {} 不应被识别",
+ label
+ );
+ }
+ }
}