This commit is contained in:
lychang
2026-07-20 01:33:44 +08:00
parent 6b2f19ef47
commit 5b803cc178
9 changed files with 821 additions and 184 deletions

View File

@@ -17,6 +17,7 @@ use crate::db::Database;
use crate::engine::title_parser;
use crate::engine::tag_extractor;
use crate::engine::person_merger;
use crate::engine::tag_dict;
use std::collections::HashSet;
// ════════════════════════════════════════════════
@@ -273,6 +274,51 @@ pub async fn phase_clean_persons(db: &Database) -> (DirtySet, usize) {
(dirty, deleted)
}
// ════════════════════════════════════════════════
// Phase A2B:清理 persons 表中与停用词重叠的人物
// ════════════════════════════════════════════════
/// Phase A2B:检查 persons 表中的人名是否在停用词词典中。
///
/// 当用户在词典中添加了新的停用词,而这些词恰好已被作为人物创建时,
/// 需要删除该人物并释放对应的相册,让它们在后续 Phase 1 中重新解析。
/// 这样可以防止停用词(如 "photo"、"model" 等通用词)被错误识别为人物。
pub async fn phase_clean_stop_word_persons(db: &Database) -> (DirtySet, usize) {
let persons = db.get_all_persons_with_counts().await;
let dict = tag_dict::get_tag_dict();
let mut dirty = DirtySet::new();
let mut deleted = 0usize;
for (pid, name, _albums, _images) in &persons {
// 检查人名本身是否在停用词中
let in_stop_words = dict.is_stop_word(name)
|| dict.is_stop_word(&name.to_lowercase());
// 检查中文核心是否在停用词中(例如 "杨晨晨" 不在,但 "photo" 可能在)
let cn_in_stop_words = title_parser::extract_cn(name)
.map(|cn| dict.is_stop_word(&cn))
.unwrap_or(false);
if in_stop_words || cn_in_stop_words {
let album_list = db.get_person_albums(*pid).await;
let album_ids: Vec<String> = album_list.into_iter().map(|a| a.id.to_string()).collect();
for aid in &album_ids { dirty.add(aid); }
println!("[管线·PhaseA2B] 停用词命中人物:{} (id={}),释放 {} 个相册", name, pid, album_ids.len());
// 清空相册的 person 字段
db.update_album_person_by_person_id(*pid).await;
// 彻底删除该人物
db.delete_person_completely(*pid).await;
deleted += 1;
}
}
if deleted > 0 {
println!("[管线·PhaseA2B] 清理完成:移除 {} 个与停用词重叠的人物,标记 {} 个脏相册", deleted, dirty.len());
} else {
println!("[管线·PhaseA2B] 清理完成:无与停用词重叠的人物");
}
(dirty, deleted)
}
// ════════════════════════════════════════════════
// Phase A3:收集新相册 + 半成品
// ════════════════════════════════════════════════
@@ -454,8 +500,9 @@ pub async fn phase_clean_tags_after_parse(db: &Database) -> usize {
pub struct PipelineResult {
pub tags_cleaned: usize, // Phase 0
pub person_like_tags_removed: usize, // Phase A1
pub dirty_albums_detected: usize, // A1+A2+A3
pub dirty_albums_detected: usize, // A1+A2+A2B+A3
pub non_persons_deleted: usize, // Phase A2
pub stop_word_persons_removed: usize, // Phase A2B
pub new_albums_found: usize, // Phase A3
pub persons_updated: usize, // Phase 1
pub tags_updated: usize, // Phase 1
@@ -469,12 +516,14 @@ pub async fn run_pipeline(db: &Database) -> PipelineResult {
let tags_cleaned = phase_clean_tags(db).await;
let (dirty_a1, person_like_tags_removed) = phase_detect_person_in_tags(db).await;
let (dirty_a2, non_persons_deleted) = phase_clean_persons(db).await;
let (dirty_a2b, stop_word_persons_removed) = phase_clean_stop_word_persons(db).await;
let dirty_a3 = phase_collect_new_albums(db).await;
let new_albums_found = dirty_a3.len();
let mut master_dirty = DirtySet::new();
master_dirty.extend(dirty_a1.into_inner());
master_dirty.extend(dirty_a2.into_inner());
master_dirty.extend(dirty_a2b.into_inner());
master_dirty.extend(dirty_a3.into_inner());
let dirty_albums_detected = master_dirty.len();
@@ -484,7 +533,7 @@ pub async fn run_pipeline(db: &Database) -> PipelineResult {
PipelineResult {
tags_cleaned, person_like_tags_removed, dirty_albums_detected,
non_persons_deleted, new_albums_found,
non_persons_deleted, stop_word_persons_removed, new_albums_found,
persons_updated, tags_updated, persons_merged,
aliases_moved, tags_removed_person,
}