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
//! Replay recordings payload and processor.
//!
//! # Data Scrubbing
//!
//! Since recordings contain snapshot of the browser's DOM, network traffic, and console logs, they
//! are likely to contain sensitive data. This module provides [`RecordingScrubber`], which applies
//! data scrubbing on the payload of recordings while leaving their structure and required fields
//! intact.
//!
//! Data scrubbing applies to only Sentry event payloads within the recording event stream,
//! identified by `type: 5`. The scrubber skips all other node types and does not perform any
//! validation beyond JSON parsing.

use std::borrow::Cow;
use std::cell::RefCell;
use std::fmt;
use std::io::Read;
use std::rc::Rc;

use flate2::bufread::ZlibDecoder;
use flate2::write::ZlibEncoder;
use flate2::Compression;
use once_cell::sync::Lazy;
use relay_event_schema::processor::{FieldAttrs, Pii, ProcessingState, Processor, ValueType};
use relay_pii::{PiiConfig, PiiProcessor};
use relay_protocol::Meta;
use serde::{de, ser, Deserializer};
use serde_json::value::RawValue;

use crate::transform::Transform;

/// Paths to fields on which datascrubbing rules should be applied.
///
/// This is equivalent to marking a field as `pii = true` in an `Annotated` schema.
static PII_FIELDS: Lazy<[Vec<&str>; 2]> = Lazy::new(|| {
    [
        vec!["data", "payload", "description"],
        vec!["data", "payload", "data"],
    ]
});

/// Returns `True` if the given path should be treated as `pii = true`.
fn scrub_at_path(path: &Vec<String>) -> bool {
    PII_FIELDS.iter().any(|pii_path| {
        path.len() >= pii_path.len() && pii_path.iter().zip(path).all(|(k1, k2)| k1 == k2)
    })
}

/// Static field attributes used for fields in [`PII_FIELDS`].
const FIELD_ATTRS_PII_TRUE: FieldAttrs = FieldAttrs::new().pii(Pii::True);

/// Static field attributes used for fields without PII scrubbing.
const FIELD_ATTRS_PII_FALSE: FieldAttrs = FieldAttrs::new().pii(Pii::False);

/// Error returned from [`RecordingScrubber`].
#[derive(Debug)]
pub enum ParseRecordingError {
    /// An error parsing the JSON payload.
    Parse(serde_json::Error),
    /// Invalid or broken compression.
    Compression(std::io::Error),
    /// Validation of the payload failed.
    ///
    /// The body is empty, is missing the headers, or the body.
    Message(&'static str),
}

impl fmt::Display for ParseRecordingError {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        match self {
            ParseRecordingError::Parse(serde_error) => write!(f, "{serde_error}"),
            ParseRecordingError::Compression(error) => write!(f, "{error}"),
            ParseRecordingError::Message(message) => write!(f, "{message}"),
        }
    }
}

impl std::error::Error for ParseRecordingError {
    fn source(&self) -> Option<&(dyn std::error::Error + 'static)> {
        match self {
            ParseRecordingError::Parse(e) => Some(e),
            ParseRecordingError::Compression(e) => Some(e),
            ParseRecordingError::Message(_) => None,
        }
    }
}

impl From<serde_json::Error> for ParseRecordingError {
    fn from(err: serde_json::Error) -> Self {
        ParseRecordingError::Parse(err)
    }
}

/// The [`Transform`] implementation for data scrubbing.
///
/// This is used by [`EventStreamVisitor`] and [`ScrubbedValue`] to scrub recording events.
struct ScrubberTransform<'a> {
    /// PII processors that are applied one by one on each value.
    processor1: Option<PiiProcessor<'a>>,
    processor2: Option<PiiProcessor<'a>>,
    /// The state encoding the current path, which is fed by `push_path` and `pop_path`.
    state: ProcessingState<'a>,
    /// The current path. This is redundant with `state`, which also contains the full path,
    /// but easier to match on.
    path: Vec<String>,
}

impl ScrubberTransform<'_> {
    fn ensure_empty(&mut self) {
        if !self.path.is_empty() || self.state.depth() > 0 {
            debug_assert!(false, "ScrubberTransform not empty");
            relay_log::error!("ScrubberTransform not empty");
        }
        self.state = ProcessingState::new_root(None, None);
        self.path.clear();
    }
}

impl<'de> Transform<'de> for &'_ mut ScrubberTransform<'_> {
    fn push_path(&mut self, key: &'de str) {
        self.path.push(key.to_owned());
        let field_attrs = if scrub_at_path(&self.path) {
            &FIELD_ATTRS_PII_TRUE
        } else {
            &FIELD_ATTRS_PII_FALSE
        };

        self.state = std::mem::take(&mut self.state).enter_owned(
            key.to_owned(),
            Some(Cow::Borrowed(field_attrs)),
            Some(ValueType::String), // Pretend everything is a string.
        )
    }

    fn pop_path(&mut self) {
        if let Ok(Some(parent)) = std::mem::take(&mut self.state).try_into_parent() {
            self.state = parent;
        }
        let popped = self.path.pop();
        debug_assert!(popped.is_some()); // pop_path should never be called on an empty state.
    }

    fn transform_str<'a>(&mut self, v: &'a str) -> Cow<'a, str> {
        self.transform_string(v.to_owned())
    }

    fn transform_string(&mut self, mut value: String) -> Cow<'static, str> {
        if let Some(ref mut processor) = self.processor1 {
            if processor
                .process_string(&mut value, &mut Meta::default(), &self.state)
                .is_err()
            {
                return Cow::Borrowed("");
            }
        }

        if let Some(ref mut processor) = self.processor2 {
            if processor
                .process_string(&mut value, &mut Meta::default(), &self.state)
                .is_err()
            {
                return Cow::Borrowed("");
            }
        }

        Cow::Owned(value)
    }
}

/// Helper that runs data scrubbing on a raw JSON value during serialization.
///
/// This is used by [`EventStreamVisitor`] to serialize recording events on-the-fly from a stream.
/// It uses a [`ScrubberTransform`] holding all state to perform the actual work.
struct ScrubbedValue<'a, 'b>(&'a RawValue, Rc<RefCell<ScrubberTransform<'b>>>);

impl serde::Serialize for ScrubbedValue<'_, '_> {
    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
    where
        S: serde::Serializer,
    {
        let mut transform = self.1.borrow_mut();
        let mut deserializer = serde_json::Deserializer::from_str(self.0.get());
        let scrubber = crate::transform::Deserializer::new(&mut deserializer, &mut *transform);
        serde_transcode::transcode(scrubber, serializer)
    }
}

/// A visitor that deserializes, scrubs, and serializes a stream of recording events.
struct EventStreamVisitor<'a, S> {
    serializer: S,
    scrubber: Rc<RefCell<ScrubberTransform<'a>>>,
}

impl<'a, S> EventStreamVisitor<'a, S> {
    /// The proprietary rrweb node type that identifies Sentry payloads.
    ///
    /// The visitor exclusively scrubs these nodes.
    const SENTRY_EVENT_TYPE: u8 = 5;

    /// Creates a new visitor wrapping a `serializer`.
    fn new(serializer: S, scrubber: Rc<RefCell<ScrubberTransform<'a>>>) -> Self {
        Self {
            serializer,
            scrubber,
        }
    }
}

impl<'de, 'a, S> de::Visitor<'de> for EventStreamVisitor<'a, S>
where
    S: ser::Serializer,
{
    type Value = S::Ok;

    fn expecting(&self, formatter: &mut std::fmt::Formatter) -> std::fmt::Result {
        write!(formatter, "a replay recording event stream")
    }

    fn visit_seq<A>(self, mut v: A) -> Result<Self::Value, A::Error>
    where
        A: de::SeqAccess<'de>,
    {
        /// Efficiently deserializes the type discriminator of event nodes.
        #[derive(Clone, Copy, serde::Deserialize)]
        struct TypeHelper {
            #[serde(rename = "type")]
            ty: u8,
        }

        use serde::ser::SerializeSeq;
        let mut seq = self.serializer.serialize_seq(v.size_hint()).map_err(s2d)?;

        // NOTE on performance: This loop parses every element 2-3 times:
        //  1. In the call to next_element() to obtain the raw value. This pass is highly efficient
        //     and cannot be removed.
        //  2. Deserializing `TypeHelper` to obtain the node type. It skips most of the content, but
        //     does require a full pass through the raw value.
        //  3. If scrubbing is required, `ScrubbedValue` internally parses the value and runs it
        //     through a transforming deserializer.
        //
        // These redundant passes come with slight performance overhead compared to an
        // implementation that directly parses the `type` and raw `data` into a structure. Together
        // with decompression and data scrubbing, the difference in benchmarks was small. In case
        // this becomes a performance bottleneck, it is worth to first focus on data scrubbing, and
        // then at the redundant parsing.

        while let Some(raw) = v.next_element::<&'de RawValue>()? {
            let helper = serde_json::from_str::<TypeHelper>(raw.get()).map_err(s2d)?;
            // Scrub only sentry-specific events and serialize all others without modification.
            if helper.ty == Self::SENTRY_EVENT_TYPE {
                seq.serialize_element(&ScrubbedValue(raw, self.scrubber.clone()))
                    .map_err(s2d)?;
                // `pop_path` calls should have reset the scrubber's state, but force a
                // reset here just to be sure:
                self.scrubber.borrow_mut().ensure_empty();
            } else {
                seq.serialize_element(raw).map_err(s2d)?;
            }
        }

        seq.end().map_err(s2d)
    }
}

/// Maps a serialization error to a deserialization error.
fn s2d<S, D>(s: S) -> D
where
    S: ser::Error,
    D: de::Error,
{
    D::custom(s.to_string())
}

/// A utility that performs data scrubbing on compressed Replay recording payloads.
///
/// ### Example
///
/// ```
/// use relay_replays::recording::RecordingScrubber;
/// use relay_pii::PiiConfig;
///
/// // Obtain a PII config from the project state or create one on-demand.
/// let pii_config = PiiConfig::default();
/// let mut scrubber = RecordingScrubber::new(1_000_000, Some(&pii_config), None);
///
/// let payload = b"{}\n[]";
/// let result = scrubber.process_recording(payload.as_slice());
/// ```
pub struct RecordingScrubber<'a> {
    limit: usize,
    transform: Rc<RefCell<ScrubberTransform<'a>>>,
}

impl<'a> RecordingScrubber<'a> {
    /// Creates a new `RecordingScrubber` from PII configs.
    ///
    /// `limit` controls the maximum size in bytes during decompression. This function returns an
    /// `Err` if decompressed contents exceed the limit. The two optional configs to be passed here
    /// are from data scrubbing settings and from the dedicated PII config.
    ///
    /// # Performance
    ///
    /// The passed PII configs are [compiled](PiiConfig::compiled) by this constructor if their
    /// compiled version is not yet cached. This can be a CPU-intensive process and should be called
    /// from a blocking context.
    pub fn new(
        limit: usize,
        config1: Option<&'a PiiConfig>,
        config2: Option<&'a PiiConfig>,
    ) -> Self {
        Self {
            limit,
            transform: Rc::new(RefCell::new(ScrubberTransform {
                processor1: config1.map(|c| PiiProcessor::new(c.compiled())),
                processor2: config2.map(|c| PiiProcessor::new(c.compiled())),
                state: ProcessingState::new_root(None, None),
                path: vec![],
            })),
        }
    }

    /// Returns `true` if both configs are empty and no scrubbing would occur.
    pub fn is_empty(&self) -> bool {
        let tmp = self.transform.borrow();
        tmp.processor1.is_none() && tmp.processor2.is_none()
    }

    fn scrub_replay<W>(&mut self, json: &[u8], write: W) -> Result<(), ParseRecordingError>
    where
        W: std::io::Write,
    {
        let mut deserializer = serde_json::Deserializer::from_slice(json);
        let mut serializer = serde_json::Serializer::new(write);

        deserializer.deserialize_seq(EventStreamVisitor::new(
            &mut serializer,
            self.transform.clone(),
        ))?;

        Ok(())
    }

    #[doc(hidden)] // Public for benchmarks.
    pub fn transcode_replay(
        &mut self,
        body: &[u8],
        output: &mut Vec<u8>,
    ) -> Result<(), ParseRecordingError> {
        let encoder = ZlibEncoder::new(output, Compression::default());

        if body.first() == Some(&b'[') {
            self.scrub_replay(body, encoder)
        } else {
            let mut decompressed = Vec::with_capacity(8 * 1024);
            let mut decoder = ZlibDecoder::new(body).take(self.limit as u64);
            decoder
                .read_to_end(&mut decompressed)
                .map_err(ParseRecordingError::Compression)?;

            self.scrub_replay(&decompressed, encoder)
        }
    }

    /// Parses a replay recording payload and applies data scrubbers.
    ///
    /// # Compression
    ///
    /// The recording `bytes` passed to this function can be a raw recording payload or compressed
    /// with zlib. The result is always compressed, regardless of the input.
    ///
    /// During decompression, the scrubber applies a `limit`. If the decompressed buffer exceeds the
    /// configured size, an `Err` is returned. This does not apply to decompressed payloads.
    ///
    /// # Errors
    ///
    /// This function requires a full recording payload including headers and body. This function
    /// will return errors if:
    ///  - Headers or the body are missing.
    ///  - Headers and the body are separated by exactly one UNIX newline (`\n`).
    ///  - The payload size exceeds the configured `limit` of the scrubber after decompression.
    ///  - On errors during decompression or JSON parsing.
    pub fn process_recording(&mut self, bytes: &[u8]) -> Result<Vec<u8>, ParseRecordingError> {
        // Check for null byte condition.
        if bytes.is_empty() {
            return Err(ParseRecordingError::Message("no data found"));
        }

        let mut split = bytes.splitn(2, |b| b == &b'\n');
        let header = split
            .next()
            .ok_or(ParseRecordingError::Message("no headers found"))?;

        let body = match split.next() {
            Some(b"") | None => return Err(ParseRecordingError::Message("no body found")),
            Some(body) => body,
        };

        let mut output = header.to_owned();
        output.push(b'\n');
        // Data scrubbing usually does not change the size of the output by much. We can preallocate
        // enough space for the scrubbed output to avoid resizing the output buffer serveral times.
        // Benchmarks have NOT shown a big difference, however.
        output.reserve(body.len());
        self.transcode_replay(body, &mut output)?;

        Ok(output)
    }
}

#[cfg(test)]
mod tests {
    // End to end test coverage.

    use relay_pii::{DataScrubbingConfig, PiiConfig};

    use crate::recording::scrub_at_path;

    use super::RecordingScrubber;

    fn default_pii_config() -> PiiConfig {
        let mut scrubbing_config = DataScrubbingConfig::default();
        scrubbing_config.scrub_data = true;
        scrubbing_config.scrub_defaults = true;
        scrubbing_config.scrub_ip_addresses = true;
        scrubbing_config.pii_config_uncached().unwrap().unwrap()
    }

    fn scrubber(config: &PiiConfig) -> RecordingScrubber {
        RecordingScrubber::new(usize::MAX, Some(config), None)
    }

    #[test]
    fn test_process_recording_end_to_end() {
        // Valid compressed rrweb payload.  Contains a 16 byte header followed by a new line
        // character and concludes with a gzipped rrweb payload.
        let payload: &[u8] = &[
            123, 34, 115, 101, 103, 109, 101, 110, 116, 95, 105, 100, 34, 58, 51, 125, 10, 120,
            156, 149, 144, 91, 106, 196, 32, 20, 64, 247, 114, 191, 237, 160, 241, 145, 234, 38,
            102, 1, 195, 124, 152, 104, 6, 33, 169, 193, 40, 52, 4, 247, 94, 91, 103, 40, 20, 108,
            59, 191, 247, 30, 207, 225, 122, 57, 32, 238, 171, 5, 69, 17, 24, 29, 53, 168, 3, 54,
            159, 194, 88, 70, 4, 193, 234, 55, 23, 157, 127, 219, 64, 93, 14, 120, 7, 37, 100, 1,
            119, 80, 29, 102, 8, 156, 1, 213, 11, 4, 209, 45, 246, 60, 77, 155, 141, 160, 94, 232,
            43, 206, 232, 206, 118, 127, 176, 132, 177, 7, 203, 42, 75, 36, 175, 44, 231, 63, 88,
            217, 229, 107, 174, 179, 45, 234, 101, 45, 172, 232, 49, 163, 84, 22, 191, 232, 63, 61,
            207, 93, 130, 229, 189, 216, 53, 138, 84, 182, 139, 178, 199, 191, 22, 139, 179, 238,
            196, 227, 244, 134, 137, 240, 158, 60, 101, 34, 255, 18, 241, 6, 116, 42, 212, 119, 35,
            234, 27, 40, 24, 130, 213, 102, 12, 105, 25, 160, 252, 147, 222, 103, 175, 205, 215,
            182, 45, 168, 17, 48, 118, 210, 105, 142, 229, 217, 168, 163, 189, 249, 80, 254, 19,
            146, 59, 13, 115, 10, 144, 115, 190, 126, 0, 2, 68, 180, 16,
        ];

        let config = default_pii_config();
        let result = scrubber(&config).process_recording(payload);
        assert!(!result.unwrap().is_empty());
    }

    #[test]
    fn test_process_recording_no_body_data() {
        // Empty bodies can not be decompressed and fail.
        let payload: &[u8] = &[
            123, 34, 115, 101, 103, 109, 101, 110, 116, 95, 105, 100, 34, 58, 51, 125, 10,
        ];

        let config = default_pii_config();
        let result = scrubber(&config).process_recording(payload);
        assert!(matches!(
            result.unwrap_err(),
            super::ParseRecordingError::Message("no body found"),
        ));
    }

    #[test]
    fn test_process_recording_bad_body_data() {
        // Invalid gzip body contents.  Can not deflate.
        let payload: &[u8] = &[
            123, 34, 115, 101, 103, 109, 101, 110, 116, 95, 105, 100, 34, 58, 51, 125, 10, 22,
        ];

        let config = default_pii_config();
        let result = scrubber(&config).process_recording(payload);
        assert!(matches!(
            result.unwrap_err(),
            super::ParseRecordingError::Compression(_),
        ));
    }

    #[test]
    fn test_process_recording_no_headers() {
        // No header delimiter.  Entire payload is consumed as headers.  The empty body fails.
        let payload: &[u8] = &[
            123, 34, 115, 101, 103, 109, 101, 110, 116, 95, 105, 100, 34, 58, 51, 125,
        ];

        let config = default_pii_config();
        let result = scrubber(&config).process_recording(payload);
        assert!(matches!(
            result.unwrap_err(),
            super::ParseRecordingError::Message("no body found"),
        ));
    }

    #[test]
    fn test_process_recording_no_contents() {
        // Empty payload can not be decompressed.  Header check never fails.
        let payload: &[u8] = &[];

        let config = default_pii_config();
        let result = scrubber(&config).process_recording(payload);
        assert!(matches!(
            result.unwrap_err(),
            super::ParseRecordingError::Message("no data found"),
        ));
    }

    // RRWeb Payload Coverage

    #[ignore = "type 3 nodes are not supported"]
    #[test]
    fn test_pii_credit_card_removal() {
        let payload = include_bytes!("../tests/fixtures/rrweb-pii.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();
        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let parsed = std::str::from_utf8(&transcoded).unwrap();
        assert!(parsed.contains(r#"{"type":3,"textContent":"[Filtered]","id":284}"#));
    }

    #[test]
    fn test_scrub_pii_navigation() {
        let payload = include_bytes!("../tests/fixtures/rrweb-performance-navigation.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();
        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let parsed = std::str::from_utf8(&transcoded).unwrap();
        assert!(parsed.contains("https://sentry.io?credit-card=[Filtered]"));
    }

    #[test]
    fn test_scrub_pii_resource() {
        let payload = include_bytes!("../tests/fixtures/rrweb-performance-resource.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();
        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let parsed = std::str::from_utf8(&transcoded).unwrap();
        assert!(parsed.contains("https://sentry.io?credit-card=[Filtered]"));
    }

    #[ignore = "type 3 nodes are not supported"]
    #[test]
    fn test_pii_ip_address_removal() {
        let payload = include_bytes!("../tests/fixtures/rrweb-pii-ip-address.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();
        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let parsed = std::str::from_utf8(&transcoded).unwrap();
        assert!(parsed.contains("\"value\":\"[ip]\"")); // Assert texts were mutated.
        assert!(parsed.contains("\"textContent\":\"[ip]\"")) // Assert text node was mutated.
    }

    // Event Parsing and Scrubbing.

    #[ignore = "type 2 nodes are not supported"]
    #[test]
    fn test_scrub_pii_full_snapshot_event() {
        let payload = include_bytes!("../tests/fixtures/rrweb-event-2.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();
        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let scrubbed_result = std::str::from_utf8(&transcoded).unwrap();
        // NOTE: The normalization below was removed
        // assert!(scrubbed_result.contains("\"attributes\":{\"src\":\"#\"}"));
        assert!(scrubbed_result.contains("\"textContent\":\"my ssn is [Filtered]\""));
    }

    #[ignore = "type 3 nodes are not supported"]
    #[test]
    fn test_scrub_pii_incremental_snapshot_event() {
        let payload = include_bytes!("../tests/fixtures/rrweb-event-3.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();
        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let scrubbed_result = std::str::from_utf8(&transcoded).unwrap();
        assert!(scrubbed_result.contains("\"textContent\":\"[Filtered]\""));
        assert!(scrubbed_result.contains("\"value\":\"[Filtered]\""));
    }

    #[test]
    fn test_scrub_pii_custom_event() {
        let payload = include_bytes!("../tests/fixtures/rrweb-event-5.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();
        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let scrubbed_result = std::str::from_utf8(&transcoded).unwrap();
        assert!(scrubbed_result.contains("\"description\":\"[Filtered]\""));
        assert!(scrubbed_result.contains("\"description\":\"https://sentry.io?ip-address=[ip]\""));
        // NOTE: default scrubbers do not remove email address
        // assert!(scrubbed_result.contains("\"message\":\"[email]\""));
    }

    #[test]
    fn test_scrub_pii_key_based() {
        let payload = include_bytes!("../tests/fixtures/rrweb-request.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();

        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let scrubbed_result = std::str::from_utf8(&transcoded).unwrap();
        let scrubbed: serde_json::Value = serde_json::from_str(scrubbed_result).unwrap();

        // Normal fields are not scrubbed:
        assert_eq!(scrubbed[0]["data"]["payload"]["data"]["method"], "POST");

        assert_eq!(
            scrubbed[0]["data"]["payload"]["data"]["request"]["body"]["api_key"],
            "[Filtered]"
        );
    }

    #[test]
    fn test_scrub_pii_key_based_edge_cases() {
        let payload = include_bytes!("../tests/fixtures/rrweb-request-edge-cases.json");

        let mut transcoded = Vec::new();
        let config = default_pii_config();

        scrubber(&config)
            .scrub_replay(payload.as_slice(), &mut transcoded)
            .unwrap();

        let scrubbed_result = std::str::from_utf8(&transcoded).unwrap();
        let scrubbed: serde_json::Value = serde_json::from_str(scrubbed_result).unwrap();

        insta::assert_ron_snapshot!(scrubbed);
    }

    #[test]
    fn test_scrub_at_path() {
        for (should_scrub, path) in [
            (false, vec![]),
            (false, vec!["data"]),
            (false, vec!["data", "payload"]),
            (false, vec!["data", "payload", "foo"]),
            (false, vec!["foo", "payload", "data"]),
            (true, vec!["data", "payload", "data"]),
            (true, vec!["data", "payload", "data", "request"]),
            (true, vec!["data", "payload", "data", "request", "body"]),
        ] {
            let path = path.into_iter().map(|p| p.to_owned()).collect::<Vec<_>>();
            assert_eq!(should_scrub, scrub_at_path(&path));
        }
    }
}