crawlrs 0.2.0

crawlrs - Enterprise-grade web scraping and crawling platform built with Rust
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
// Copyright (c) 2025-2026 Kirky.X🌠
// SPDX-License-Identifier: Apache-2.0

use anyhow::{Context, Result};
use chrono::Utc;
use dashmap::DashMap;
use log::{debug, error, info, warn};
use serde_json::{json, Value};
use std::collections::HashMap;
use std::sync::atomic::{AtomicI64, Ordering};
use std::sync::Arc;
use std::time::Duration;
use tokio::sync::OwnedSemaphorePermit;
use tokio::time::sleep;

use crate::application::dto::crawl_request::CrawlConfigDto;
use crate::application::dto::scrape_request::{ScrapeOptionsDto, ScrapeRequestDto};
use crate::application::use_cases::create_scrape::CreateScrapeUseCaseTrait;
use crate::config::settings::Settings;
use crate::domain::id::Id;
use crate::domain::models::{Task, TaskStatus};
use crate::domain::repositories::crawl_repository::CrawlRepository;
use crate::domain::repositories::credits_repository::CreditsRepository;
use crate::domain::repositories::scrape_result_repository::ScrapeResultRepository;
use crate::domain::repositories::task_repository::TaskRepository;
use crate::domain::services::extraction_service::ExtractionServiceTrait;
use crate::domain::services::retry_handler::RetryHandler;
use crate::domain::services::webhook_service::WebhookService;
// URL 分层去重器(Bloom 预筛 + DB 保权威)
use crate::utils::dedup::Deduplicator;

use crate::common::CacheContext;
use crate::common::CacheMode;
use crate::common::HttpMethod;
use crate::domain::services::team_semaphore::TeamSemaphore;
use crate::engines::engine_client::{
    EngineClient, EngineError, PageAction, ScrapeOptions, ScrapeRequest, ScrapeResponse,
    ScreenshotConfig, ScrollDirection, WaitFor,
};
use crate::infrastructure::oxcache::CacheService;
use crate::infrastructure::security::ssrf::is_internal_url;
use crate::queue::task_queue::TaskQueue;
// 请求合并协调器(替代原 request_coalescer 字段 + try_coalesce 方法)
use crate::workers::coalesce_coordinator::CoalesceCoordinator;
// 优雅退出协调器
use crate::workers::shutdown::ShutdownCoordinator;
// cache key 生成
use crate::workers::cache_utils;
// 请求合并键(租户 + 规范化 URL + 完整请求身份指纹)
use crate::workers::coalesce_coordinator::CoalesceKey;
// URL 日志脱敏
use crate::utils::url_redaction::redact_url_for_log;
// Markdown 后处理器(gated `markdown` 特性,替代原 maybe_generate_markdown 方法)
use crate::utils::retry_policy::RetryPolicy;
#[cfg(feature = "content")]
use crate::workers::markdown_post_processor::MarkdownPostProcessor;
// 重试指令与分类器(消费 RetryTracker + RetryDirective)
use crate::utils::retry::{RetryDirective, RetryTracker};
// 自适应爬取停止条件
use crate::utils::robots::RobotsCheckerTrait;
// 内存感知调度器接入 scrape_worker
// MemoryScheduler 依赖 SystemMonitorTrait(metrics 特性门控),故整块接入由 metrics 门控
#[cfg(feature = "metrics")]
use crate::workers::scheduler::memory_scheduler::MemoryScheduler;

// 提取到独立模块的函数导入
pub(super) use super::crawl_link_extractor::{
    check_robots_txt as check_robots_txt_fn, extract_and_queue_links as extract_and_queue_links_fn,
    update_crawl_completion_status as update_crawl_completion_status_fn,
};
pub(super) use super::scrape_executor::{
    process_text_encoding, save_result, try_read_scrape_cache, try_write_scrape_cache,
};
pub(super) use super::scrape_response_builder::{
    build_crawl_request as build_crawl_request_fn,
    build_extract_request as build_extract_request_fn, parse_crawl_payload, parse_extract_payload,
};

// Test-only imports for cfg(test) wrappers and functions
#[cfg(test)]
use crate::application::dto::extract_request::ExtractRequestDto;

/// 抓取工作者
pub struct ScrapeWorker {
    repository: Arc<dyn TaskRepository>,
    result_repository: Arc<dyn ScrapeResultRepository>,
    crawl_repository: Arc<dyn CrawlRepository>,
    webhook_service: Arc<dyn WebhookService>,
    credits_repository: Arc<dyn CreditsRepository>,
    engine_client: Arc<EngineClient>,
    _create_scrape_use_case: Arc<dyn CreateScrapeUseCaseTrait>,
    team_semaphore: Arc<TeamSemaphore>,
    /// 请求合并协调器
    ///
    /// 同 URL 并发请求只允许首个执行实际抓取,其余 worker 等待广播后从
    /// `result_repo` 读取结果,避免重复网络往返。所有 worker 共享同一实例
    /// (由 `WorkerManager` 从 `ServicesComponents.request_coalescer` +
    /// `repository` + `result_repository` 构造注入)。
    coalesce_coordinator: Arc<CoalesceCoordinator>,
    /// Markdown 后处理器
    ///
    /// 无状态服务,根据任务 `formats` 字段判断是否生成 Markdown。
    /// gated `markdown` 特性:关闭时本字段不存在,相关分支也不编译。
    #[cfg(feature = "content")]
    markdown_post_processor: MarkdownPostProcessor,
    token_usage: Arc<DashMap<crate::domain::id::Id, AtomicI64>>,
    robots_checker: Arc<dyn RobotsCheckerTrait>,
    settings: Arc<Settings>,
    worker_id: crate::domain::id::Id,
    default_concurrency_limit: usize,
    retry_handler: RetryHandler,
    extraction_service: Arc<dyn ExtractionServiceTrait>,
    /// 内存感知调度器
    ///
    /// `metrics` 启用时由 `WorkerManager` 注入;`process_task` 在获取并发许可前
    /// 调用 `admit()`,Pressure 时延后、Critical 时重排到 backlog。
    #[cfg(feature = "metrics")]
    memory_scheduler: Arc<MemoryScheduler>,
    /// URL 分层去重器
    ///
    /// UrlNormalizer + Bloom + HashSet 三层组合,用于 `extract_and_queue_links`
    /// 预筛 URL 是否已爬:
    /// - Bloom 阴性 → 绝对新,直接入队并 insert
    /// - Bloom 阳性 → 可能已爬,回落 `find_existing_urls` DB 校验(保权威)
    ///
    /// `RwLock` 因为 `extract_and_queue_links` 是 `&self`,但 bloom insert 需 `&mut`。
    /// `Arc` 是为后续支持跨 worker 共享(当前每 worker 独立实例)。
    deduplicator: Arc<parking_lot::RwLock<Deduplicator>>,
    /// crawl 终态短 TTL 缓存(crawl_id → (is_terminal, checked_at))
    ///
    /// 终态闸每个 crawl 子任务都会查一次父 crawl;活跃 crawl 下这是纯增量
    /// SELECT(审查性能项)。10s TTL 把重复查询折叠,最坏延迟 10s 发现终态。
    crawl_terminal_cache: DashMap<crate::domain::id::Id, (bool, std::time::Instant)>,
    /// 缓存插入计数:每 1024 次触发一次过期项 retain(避免每任务 len 扫描)
    crawl_terminal_cache_inserts: std::sync::atomic::AtomicUsize,
    /// 团队黑名单快照缓存(crawl_id → (归一化条目, checked_at))。
    /// crawl.config 创建后不再修改;300s TTL 折叠逐任务 PK SELECT。
    crawl_blacklist_cache:
        DashMap<crate::domain::id::Id, (std::sync::Arc<Vec<String>>, std::time::Instant)>,
    crawl_blacklist_cache_inserts: std::sync::atomic::AtomicUsize,
    /// 高级缓存服务
    ///
    /// 由 `WorkerManager` 从 `InfrastructureComponents.cache_service` 注入,
    /// 所有 worker 共享同一实例。`process_scrape_task` 在读写缓存前经
    /// `CacheContext` 门控:`is_cacheable && should_read` 查缓存命中直返,
    /// `is_cacheable && should_write` 抓取成功后写回。
    cache_service: Arc<dyn CacheService>,
    /// 优雅退出协调器
    ///
    /// 由 `WorkerManager` 注入,所有 worker 共享同一实例。`run` 循环开头
    /// 检查 `is_shutting_down`,置位后不再 acquire 新任务,完成当前任务即退出。
    shutdown_coordinator: Arc<ShutdownCoordinator>,
}

impl std::fmt::Debug for ScrapeWorker {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        f.debug_struct("ScrapeWorker")
            .field("worker_id", &self.worker_id)
            .field("default_concurrency_limit", &self.default_concurrency_limit)
            .finish_non_exhaustive()
    }
}

impl ScrapeWorker {
    /// 本 worker 的认领身份(即任务 lock_token)
    pub fn worker_id(&self) -> crate::domain::id::Id {
        self.worker_id
    }

    /// 创建新的抓取工作器实例
    ///
    /// 接受 `ScrapeWorkerDeps` 参数对象,聚合所有外部依赖。
    /// 由 `ScrapeWorkerBuilder::build()` 或 `WorkerManager` 构造传入。
    pub fn new(deps: ScrapeWorkerDeps) -> Self {
        // 根据任务类型选择合适的重试策略
        let retry_policy = RetryPolicy::slow(); // 网络请求适合慢速重试策略
        let retry_handler = RetryHandler::new(deps.repository.clone(), retry_policy.clone());

        Self {
            repository: deps.repository,
            result_repository: deps.result_repository,
            crawl_repository: deps.crawl_repository,
            webhook_service: deps.webhook_service,
            credits_repository: deps.credits_repository,
            engine_client: deps.engine_client,
            _create_scrape_use_case: deps.create_scrape_use_case,
            team_semaphore: deps.team_semaphore,
            coalesce_coordinator: deps.coalesce_coordinator,
            #[cfg(feature = "content")]
            markdown_post_processor: MarkdownPostProcessor::new(
                // 按 [markdown].converter 装配(moli 链默认 / htmd 直通诊断)
                crate::domain::services::markdown::build_converter(&deps.settings.markdown),
                // bootstrap 注入的带质量门 Facade([content_extraction] 四键 +
                // llm_service 回退);不再本地 new(None) 另立 disabled 实例
                Some(Arc::clone(&deps.content_extractor)),
            ),
            token_usage: Arc::new(DashMap::new()),
            robots_checker: deps.robots_checker,
            settings: deps.settings,
            worker_id: crate::domain::id::Id::generate(),
            default_concurrency_limit: deps.default_concurrency_limit,
            retry_handler,
            extraction_service: deps.extraction_service,
            #[cfg(feature = "metrics")]
            memory_scheduler: deps.memory_scheduler,
            // 默认每 worker 独立 Deduplicator
            // 后续可由 WorkerManager 通过 Builder 注入共享实例优化 DB 查询量
            deduplicator: Arc::new(parking_lot::RwLock::new(Deduplicator::new())),
            crawl_terminal_cache: DashMap::new(),
            crawl_terminal_cache_inserts: std::sync::atomic::AtomicUsize::new(0),
            crawl_blacklist_cache: DashMap::with_capacity(1024),
            crawl_blacklist_cache_inserts: std::sync::atomic::AtomicUsize::new(0),
            cache_service: deps.cache_service,
            // 默认独立协调器,由 WorkerManager 通过
            // `with_shutdown_coordinator` 注入共享实例。
            shutdown_coordinator: Arc::new(ShutdownCoordinator::default()),
        }
    }

    /// 注入优雅退出协调器
    ///
    /// 由 `WorkerManager` 在构造后调用,使所有 worker 共享同一实例;
    /// 关闭信号到达后 `run()` 循环检查 `is_shutting_down()` 协同退出。
    pub fn with_shutdown_coordinator(mut self, coordinator: Arc<ShutdownCoordinator>) -> Self {
        self.shutdown_coordinator = coordinator;
        self
    }

    /// 运行抓取工作器
    pub async fn run(&self, queue: Arc<dyn TaskQueue>) {
        info!("Scrape worker {} started", self.worker_id);

        loop {
            // 优雅退出检查
            //
            // flag 置位后不再 acquire 新任务;若当前有任务正在执行,
            // 该任务完成返回后本循环即退出(不抢占正在执行的任务)。
            if self.shutdown_coordinator.is_shutting_down() {
                info!(
                    "Scrape worker {} received shutdown, exiting after current task",
                    self.worker_id
                );
                break;
            }

            match self.process_next_task(&queue).await {
                Ok(processed) => {
                    if !processed {
                        sleep(Duration::from_secs(1)).await;
                    }
                }
                Err(e) => {
                    error!("Error processing task: {}", e);
                    sleep(Duration::from_secs(1)).await;
                }
            }
        }
    }

    /// Builder 内部使用:替换 deduplicator 字段
    ///
    /// 用于 `ScrapeWorkerBuilder::build` 在调用 `ScrapeWorker::new`(内部默认
    /// 初始化 deduplicator)后,注入外部共享实例。`None` 保留默认实例。
    pub(crate) fn with_deduplicator_opt(
        mut self,
        dedup: Option<Arc<parking_lot::RwLock<Deduplicator>>>,
    ) -> Self {
        if let Some(d) = dedup {
            self.deduplicator = d;
        }
        self
    }

    /// 测试 helper:获取 deduplicator 引用(仅 `#[cfg(test)]` 可用)
    ///
    /// 用于单元测试预填充 Bloom,模拟"URL 已爬"场景:
    /// ```ignore
    /// let worker = build_mock_worker.await;
    /// {
    ///     let mut dedup = worker.deduplicator_for_test.write;
    ///     dedup.insert("https://example.com/page1");
    /// }
    /// // 现在 page1 在 Bloom 中阳性,调用 extract_and_queue_links 时
    /// // 会走 find_existing_urls DB 校验路径
    /// ```
    #[cfg(test)]
    pub(crate) fn deduplicator_for_test(&self) -> Arc<parking_lot::RwLock<Deduplicator>> {
        self.deduplicator.clone()
    }

    async fn process_next_task(&self, queue: &dyn TaskQueue) -> Result<bool> {
        let task_opt = queue.dequeue(self.worker_id).await?;

        if let Some(task) = task_opt {
            self.process_task(task).await?;
            return Ok(true);
        }

        Ok(false)
    }

    fn acquire_concurrency_permit(&self, task: &Task) -> Option<OwnedSemaphorePermit> {
        self.team_semaphore.try_acquire(task.team_id)
    }

    async fn process_task(&self, task: Task) -> Result<()> {
        debug!(
            "process_task: task_id={}, url={}, task_type={}",
            task.id, task.url, task.task_type
        );
        // 锁续期心跳执行期间按锁时长的 1/3 周期续期,
        // 防止长任务锁过期被其他 worker 抢占。RAII guard 在任意返回路径自动停止。
        let _lock_heartbeat = crate::workers::lock_heartbeat::LockHeartbeat::start(
            self.repository.clone(),
            task.id,
            self.worker_id,
            self.settings.concurrency.task_lock_duration_seconds,
        );
        info!("Processing task");

        if self.handle_expired_task(&task).await? {
            return Ok(());
        }

        // 内存感知准入检查(策略与 AIMD 联动详见 [`ScrapeWorker::admit_task_by_memory`])
        #[cfg(feature = "metrics")]
        if self.admit_task_by_memory(&task).await? {
            return Ok(());
        }

        // Concurrency Check (Layer 2: Team Semaphore)
        // The permit is held for the duration of task processing and auto-releases on drop.
        let _permit = match self.acquire_permit_or_reschedule(&task).await? {
            Some(p) => p,
            None => return Ok(()),
        };

        let task_type = task.task_type;

        // Take task by value only for the specific branch that needs it
        // This avoids 3 unnecessary clones in the match
        let result = match task_type.as_str() {
            "scrape" => self.process_scrape_task(task).await,
            "crawl" => self.process_crawl_task(task).await,
            "extract" => self.process_extract_task(task).await,
            _ => return Err(anyhow::anyhow!("Unknown task type: {}", task_type)),
        };

        // _permit auto-releases here when it goes out of scope

        if let Err(ref e) = result {
            debug!("error: {}", e);
        } else {
            debug!("Task processing completed successfully");
        }

        result
    }

    /// 从 Task payload 解析出 [`ScrapeRequestDto`](PERF- 拆分两步式)。
    ///
    /// 该方法仅负责反序列化,调用方可继续调 [`Self::build_scrape_request_from_dto`]
    /// 构造 [`ScrapeRequest`],或直接复用 dto 引用避免二次解析。
    ///
    /// # 性能要点
    ///
    /// `serde_json::from_value` 需要 owned `Value`,因此 `task.payload` 必须 clone 一次。
    /// 调用方拿到 dto 后应在所有后续路径(构造 [`ScrapeRequest`]、生成 markdown、
    /// 检查 extraction_rules)复用同一引用,禁止再次 `from_value(task.payload.clone())`。
    pub(crate) fn parse_scrape_request_dto(task: &Task) -> Result<ScrapeRequestDto> {
        serde_json::from_value(task.payload.clone()).context("Failed to parse task payload")
    }

    /// 从已解析的 [`ScrapeRequestDto`] 构造 [`ScrapeRequest`](PERF- 拆分两步式)。
    ///
    /// 不再读取 `task.payload`,避免重复解析与 clone。
    pub(crate) fn build_scrape_request_from_dto(dto: &ScrapeRequestDto) -> Result<ScrapeRequest> {
        let options = dto.options.as_ref();
        let headers = build_dto_headers(options);
        let needs_js = dto_needs_js(dto, options);
        let screenshot_config = build_dto_screenshot_config(options);

        Ok(ScrapeRequest {
            url: dto.url.clone(),
            options: ScrapeOptions {
                // NOTE: ScrapeOptionsDto 当前未暴露 method/body 字段,默认 Get + None。
                // 若未来 DTO 增加这些字段,需同步更新此处。
                method: HttpMethod::Get,
                body: None,
                headers,
                timeout: Duration::from_secs(options.and_then(|o| o.timeout).unwrap_or(30)),
                needs_js,
                needs_screenshot: options.and_then(|o| o.screenshot).unwrap_or(false),
                screenshot_config,
                mobile: options.and_then(|o| o.mobile).unwrap_or(false),
                proxy: options.and_then(|o| o.proxy.clone()),
                skip_tls_verification: options
                    .and_then(|o| o.skip_tls_verification)
                    .unwrap_or(false),
                needs_tls_fingerprint: options
                    .and_then(|o| o.needs_tls_fingerprint)
                    .unwrap_or(false),
                use_fire_engine: options.and_then(|o| o.use_fire_engine).unwrap_or(false),
                actions: map_dto_actions(dto),
                sync_wait_ms: dto.sync_wait_ms.unwrap_or(0),
                // needs_mllm 在 engine-mllm 特性关闭时显性失败:任务已在队列中,静默降级
                // 到普通引擎会让用户以为做了视觉分析——失败必须显性化。
                // (API 入口的 create_scrape use-case 在提交时已先行拒绝,此处
                // 兜底覆盖直写队列/旧版本任务等绕过 use-case 的路径。)
                needs_mllm: resolve_needs_mllm(options)?,
                block_ads: options.and_then(|o| o.block_ads).unwrap_or(false),
                block_media: options.and_then(|o| o.block_media).unwrap_or(false),
                session_id: None,
                // cache_mode 桥接(bypass_cache 优先级处理)
                //
                // bypass_cache=Some(true) 覆盖 cache_mode 为 Bypass(应急绕过读,正常写回)。
                // 其余情况按 cache_mode 走;两者皆 None 时等价于 Enabled(默认)。
                cache_mode: map_dto_cache_mode(options),
                // P22 接线:DTO 的 wait_for(毫秒)映射为 DomStable 等待策略——
                // 与 create_scrape use-case 层的映射保持一致(此前此处硬编码
                // None,调用方传入的 wait_for 被静默丢弃)
                wait_for: map_wait_for(options.and_then(|o| o.wait_for)),
            },
        })
    }

    /// 从 [`Task`] 一次性解析并构造 [`ScrapeRequest`](便利入口)。
    ///
    /// 调用方仅需 `&Task`,无需关心 dto 解析细节。内部委托
    /// `Self::parse_scrape_request_dto` + `Self::build_scrape_request_from_dto`。
    ///
    /// **若调用方需要同时使用 [`ScrapeRequestDto`] 与 [`ScrapeRequest`]**,
    /// 应直接调 `Self::parse_scrape_request_dto` 拿到 dto 后再调
    /// `Self::build_scrape_request_from_dto`,避免双解析双 clone payload。
    pub fn build_scrape_request(task: &Task) -> Result<ScrapeRequest> {
        let dto = Self::parse_scrape_request_dto(task)?;
        Self::build_scrape_request_from_dto(&dto)
    }
}

/// DTO options → 请求头 map(仅保留字符串值)
fn build_dto_headers(options: Option<&ScrapeOptionsDto>) -> HashMap<String, String> {
    let mut headers = HashMap::with_capacity(16);
    if let Some(opts) = options {
        if let Some(h) = &opts.headers {
            if let Some(obj) = h.as_object() {
                for (k, v) in obj {
                    // 与 crawl 路径同规则:Host 由目标 URL 固定,租户头不得
                    // 覆盖,否则可指向同 IP 的内网虚拟主机(SSRF 旁路)。
                    if k.eq_ignore_ascii_case("host") {
                        log::warn!(
                            "ignoring custom Host header in scrape options (target host is fixed by URL)"
                        );
                        continue;
                    }
                    if let Some(s) = v.as_str() {
                        headers.insert(k.clone(), s.to_string());
                    }
                }
            }
        }
    }
    headers
}

/// JS 渲染需求:存在动作或显式开启
fn dto_needs_js(dto: &ScrapeRequestDto, options: Option<&ScrapeOptionsDto>) -> bool {
    dto.actions.as_ref().map(|a| !a.is_empty()).unwrap_or(false)
        || options.and_then(|o| o.js_rendering).unwrap_or(false)
}

/// DTO 截图选项 → 引擎截图配置
fn build_dto_screenshot_config(options: Option<&ScrapeOptionsDto>) -> Option<ScreenshotConfig> {
    options.and_then(|o| {
        o.screenshot_options.as_ref().map(|so| ScreenshotConfig {
            full_page: so.full_page.unwrap_or(false),
            selector: so.selector.clone(),
            quality: so.quality,
            format: so.format.clone(),
        })
    })
}

/// DTO 动作 → 引擎页面动作(Screenshot 由全局 needs_screenshot 承载)
fn map_dto_actions(dto: &ScrapeRequestDto) -> Vec<PageAction> {
    dto.actions
        .clone()
        .unwrap_or_default()
        .into_iter()
        .filter_map(|a| match a {
            crate::application::dto::scrape_request::ScrapeActionDto::Wait { milliseconds } => {
                Some(PageAction::Wait { milliseconds })
            }
            crate::application::dto::scrape_request::ScrapeActionDto::Click { selector } => {
                Some(PageAction::Click { selector })
            }
            crate::application::dto::scrape_request::ScrapeActionDto::Scroll { direction } => {
                // Map string direction to ScrollDirection enum
                let dir = map_scroll_direction(&direction);
                Some(PageAction::Scroll { direction: dir })
            }
            crate::application::dto::scrape_request::ScrapeActionDto::Screenshot { .. } => {
                // Screenshot action is handled by global needs_screenshot option
                None
            }
            crate::application::dto::scrape_request::ScrapeActionDto::Input { selector, text } => {
                Some(PageAction::Input { selector, text })
            }
        })
        .collect()
}

/// 滚动方向字符串 → 枚举(未知值回落 Down)
fn map_scroll_direction(direction: &str) -> ScrollDirection {
    match direction.to_lowercase().as_str() {
        "up" => ScrollDirection::Up,
        "top" => ScrollDirection::Top,
        "bottom" => ScrollDirection::Bottom,
        _ => ScrollDirection::Down,
    }
}

/// needs_mllm 解析:特性缺失时显性失败
fn resolve_needs_mllm(options: Option<&ScrapeOptionsDto>) -> Result<bool> {
    let requested = options.and_then(|o| o.needs_mllm).unwrap_or(false);
    if requested {
        #[cfg(not(feature = "engine-mllm"))]
        anyhow::bail!("task requested needs_mllm but this build lacks the `engine-mllm` feature");
        #[cfg(feature = "engine-mllm")]
        log::debug!("task requests MLLM analysis");
    }
    Ok(requested)
}

/// cache_mode 桥接:bypass_cache 覆盖为 Bypass,否则沿用 DTO
fn map_dto_cache_mode(options: Option<&ScrapeOptionsDto>) -> Option<CacheMode> {
    let bypass = options.and_then(|o| o.bypass_cache).unwrap_or(false);
    if bypass {
        Some(CacheMode::Bypass)
    } else {
        options.and_then(|o| o.cache_mode)
    }
}

/// P22:DTO `wait_for`(毫秒)→ 引擎等待策略映射
///
/// 与 create_scrape use-case 层保持一致:毫秒数映射为 `DomStable`(DOM
/// 连续该时长无变化视为就绪)。`None` → `None`(引擎回落 NetworkIdle 默认)。
/// 增量抓取(scoped 切片):从历史结果 headers JSON 提取缓存验证器
///
/// 返回 `(etag, last_modified)`——均为响应头原文,大小写不敏感匹配键名
/// (上游服务器发送 `ETag`/`etag` 均有可能)。两个验证器都缺失时返回
/// `None`(无法发起条件请求)。
fn extract_cache_validators(headers: &serde_json::Value) -> Option<(String, Option<String>)> {
    let map = headers.as_object()?;
    let lookup = |name: &str| {
        map.iter()
            .find(|(k, _)| k.eq_ignore_ascii_case(name))
            .and_then(|(_, v)| v.as_str())
            .map(str::to_string)
    };
    let etag = lookup("etag");
    let last_modified = lookup("last-modified");
    if etag.is_some() || last_modified.is_some() {
        Some((etag.unwrap_or_default(), last_modified))
    } else {
        None
    }
}

/// 增量抓取(scoped 切片):304 响应合成——用历史结果内容回填 200 响应,
/// 附加 `x-crawlrs-not-modified: true` 标记头。下游(提取/落库/webhook)
/// 走常规路径,无感知 304 语义。
fn not_modified_response(
    prior: &crate::domain::models::ScrapeResult,
    response: &ScrapeResponse,
) -> ScrapeResponse {
    let mut headers = response.headers.clone();
    headers.insert("x-crawlrs-not-modified".to_string(), "true".to_string());
    // 历史结果的响应头一并回填(客户端可读到原始 ETag 等)
    for (k, v) in prior.headers.as_object().into_iter().flatten() {
        if let Some(v_str) = v.as_str() {
            headers
                .entry(k.clone())
                .or_insert_with(|| v_str.to_string());
        }
    }
    ScrapeResponse {
        js_degraded: None,
        content: prior.content.clone(),
        status_code: 200,
        screenshot: None,
        content_type: prior.content_type.clone(),
        headers,
        response_time_ms: response.response_time_ms,
        final_url: response.final_url.clone(),
        markdown: None,
        // 304 合成响应无 XHR 捕获语义(历史结果不含捕获数据)
        captured_responses: Vec::new(),
    }
}

/// 单测锚点:纯函数模块级可测(无 worker 依赖)
#[cfg(test)]
mod incremental_scrape_tests {
    use super::*;
    use serde_json::json;

    #[test]
    fn extract_validators_reads_etag_and_last_modified_case_insensitive() {
        let headers =
            json!({"ETag": "\"abc123\"", "Last-Modified": "Wed, 24 Sep 2026 08:00:00 GMT"});
        let (etag, lm) = extract_cache_validators(&headers).expect("validators present");
        assert_eq!(etag, "\"abc123\"");
        assert_eq!(lm.as_deref(), Some("Wed, 24 Sep 2026 08:00:00 GMT"));

        // 小写键名(部分代理/引擎会规范化)
        let headers = json!({"etag": "\"x\"", "last-modified": "Wed, 24 Sep 2026 09:00:00 GMT"});
        let (etag, _) = extract_cache_validators(&headers).expect("lowercase keys");
        assert_eq!(etag, "\"x\"");
    }

    #[test]
    fn extract_validators_none_when_missing() {
        assert!(extract_cache_validators(&json!({"content-type": "text/html"})).is_none());
        assert!(extract_cache_validators(&serde_json::Value::Null).is_none());
        assert!(extract_cache_validators(&json!("not an object")).is_none());
    }

    #[test]
    fn not_modified_response_reuses_prior_content_with_marker() {
        use crate::domain::models::ScrapeResult;
        let prior = ScrapeResult {
            id: crate::domain::id::Id::generate(),
            task_id: crate::domain::id::Id::generate(),
            url: "https://example.com/a".to_string(),
            status_code: 200,
            content: "<html>original body</html>".to_string(),
            content_type: "text/html".to_string(),
            headers: json!({"etag": "\"v1\"", "content-type": "text/html"}),
            meta_data: serde_json::Value::Null,
            screenshot: None,
            response_time_ms: 42,
            created_at: chrono::Utc::now(),
        };
        let not_modified = ScrapeResponse {
            js_degraded: None,
            captured_responses: Vec::new(),
            status_code: 304,
            content: String::new(),
            content_type: String::new(),
            headers: {
                let mut h = std::collections::HashMap::new();
                h.insert("etag".to_string(), "\"v1\"".to_string());
                h
            },
            screenshot: None,
            response_time_ms: 5,
            final_url: None,
            markdown: None,
        };

        let reused = not_modified_response(&prior, &not_modified);
        assert_eq!(reused.status_code, 200, "304 must be rewritten as 200");
        assert_eq!(reused.content, "<html>original body</html>");
        assert_eq!(reused.content_type, "text/html");
        assert_eq!(
            reused
                .headers
                .get("x-crawlrs-not-modified")
                .map(String::as_str),
            Some("true"),
            "not-modified marker must be present"
        );
        // 历史响应头回填(客户端可读原始 ETag)
        assert_eq!(
            reused.headers.get("etag").map(String::as_str),
            Some("\"v1\"")
        );
    }

    #[test]
    fn conditional_request_dto_deserializes_all_branches() {
        use crate::application::dto::scrape_request::ScrapeRequestDto;
        let on: ScrapeRequestDto = serde_json::from_value(
            json!({"url": "https://x.com", "options": {"conditional_request": true}}),
        )
        .expect("valid");
        assert_eq!(
            on.options.as_ref().and_then(|o| o.conditional_request),
            Some(true)
        );
        let off: ScrapeRequestDto = serde_json::from_value(
            json!({"url": "https://x.com", "options": {"conditional_request": false}}),
        )
        .expect("valid");
        assert_eq!(
            off.options.as_ref().and_then(|o| o.conditional_request),
            Some(false)
        );
        let absent: ScrapeRequestDto =
            serde_json::from_value(json!({"url": "https://x.com"})).expect("valid");
        assert_eq!(
            absent.options.as_ref().and_then(|o| o.conditional_request),
            None,
            "field absent must default to None (disabled)"
        );
    }
}

fn map_wait_for(ms: Option<u64>) -> Option<WaitFor> {
    // 复查 H3:钳制到 10s(引擎侧 WAIT_TIMEOUT_CAP)——DTO 允许 ≤60000ms,
    // 但 stable_duration 超过等待预算时 DomStable 数学上不可能满足,
    // 会把 API 合法输入变成确定性失败。超预算部分属"尽力等待"语义。
    ms.map(|ms| WaitFor::DomStable(std::time::Duration::from_millis(ms.min(10_000))))
}

// ===== Test-only wrappers for extracted free functions =====
// 测试代码通过 `worker.xxx()` 调用这些包装方法,
// 生产代码直接调用提取后的自由函数。
#[cfg(test)]
impl ScrapeWorker {
    async fn parse_crawl_payload(
        &self,
        task: &Task,
    ) -> Result<(crate::domain::id::Id, u32, CrawlConfigDto)> {
        parse_crawl_payload(task)
    }

    async fn check_robots_txt(&self, task: &Task) -> bool {
        check_robots_txt_fn(
            task,
            self.robots_checker.as_ref(),
            self.settings.robots.user_agent.as_str(),
        )
        .await
    }

    fn build_crawl_request(&self, task: &Task, config: &CrawlConfigDto) -> ScrapeRequest {
        build_crawl_request_fn(
            task,
            config,
            self.settings.timeouts.engines.default_timeout_seconds,
        )
    }

    async fn update_crawl_completion_status(&self, crawl_id: crate::domain::id::Id) {
        update_crawl_completion_status_fn(crawl_id, self.crawl_repository.as_ref()).await
    }

    async fn parse_extract_payload(&self, task: &Task) -> Result<(ExtractRequestDto, String)> {
        parse_extract_payload(task)
    }

    fn build_extract_request(&self, url: &str) -> ScrapeRequest {
        build_extract_request_fn(url, self.settings.timeouts.engines.default_timeout_seconds)
    }

    async fn try_read_scrape_cache(
        &self,
        ctx: &CacheContext,
        key: &str,
    ) -> Result<Option<ScrapeResponse>> {
        try_read_scrape_cache(ctx, key, self.cache_service.as_ref()).await
    }

    async fn try_write_scrape_cache(
        &self,
        ctx: &CacheContext,
        key: &str,
        response: &ScrapeResponse,
    ) -> Result<()> {
        try_write_scrape_cache(
            ctx,
            key,
            response,
            self.cache_service.as_ref(),
            self.settings.cache.types.search.ttl_seconds,
        )
        .await
    }

    async fn save_result(
        &self,
        task: &Task,
        response: &ScrapeResponse,
        extra_data: Option<Value>,
    ) -> Result<()> {
        save_result(task, response, extra_data, self.result_repository.as_ref()).await
    }

    /// 测试专用包装:固定传空黑名单,**禁止**生产代码调用
    /// (生产路径必须从 `crawl.config` 读取后传入,否则静默绕过策略)。
    #[cfg(test)]
    async fn extract_and_queue_links(
        &self,
        task: &Task,
        response: &ScrapeResponse,
        crawl_id: crate::domain::id::Id,
        current_depth: u32,
        config: &CrawlConfigDto,
    ) -> Result<()> {
        extract_and_queue_links_fn(
            task,
            response,
            crawl_id,
            current_depth,
            config,
            &[],
            self.repository.as_ref(),
            self.crawl_repository.as_ref(),
            &self.deduplicator,
        )
        .await
    }

    async fn process_text_encoding<'a>(
        &self,
        task: &Task,
        response: &'a ScrapeResponse,
    ) -> Result<std::borrow::Cow<'a, str>> {
        crate::workers::scrape_executor::process_text_encoding(task, response).await
    }
}

// 任务准入 / 结果提取辅助方法(partial impl block)
mod task_lifecycle;
// 抓取管线(partial impl block)
mod scrape_pipeline;
// 提取与计费(partial impl block)
mod extraction_billing;
// 成功收尾 / webhook / 失败处理(partial impl block)
mod success_finalization;
// Crawl 任务处理方法(partial impl block)
mod crawl_task;
// Extract 任务处理方法(partial impl block)
mod extract_task;

// Builder 子模块
mod builder;
pub use builder::ScrapeWorkerBuilder;
mod deps;
pub use deps::ScrapeWorkerDeps;

#[cfg(test)]
#[path = "../tests/scrape_worker_test.rs"]
mod tests;