content_scan 0.1.5

An extensible, plugin-based framework for identifying, analyzing, and recursively extracting structured content from arbitrary bytes.
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
use super::{
    analyzer_list::AnalyzerList, extractor_list::ExtractorList, Content, ContentAnalyzer, ContentExtractor, ContentIdentifier, ContentType, Filter,
    NextAction,
};
use crate::utils;
use crate::ExtractionContext;
use crate::IdentifierSet;
use crate::Object;
use crate::{Context, ScanResult};
use std::collections::HashSet;

/// The engine that drives a scan.
///
/// A `Scanner` bundles together a set of
/// [`ContentIdentifier`](crate::ContentIdentifier)s,
/// [`ContentAnalyzer`](crate::ContentAnalyzer)s,
/// [`ContentExtractor`](crate::ContentExtractor)s, an optional
/// [`Filter`], and a maximum recursion depth. Build one with
/// [`ScannerBuilder`] and drive scans with [`Scanner::scan`].
///
/// Scanners are reusable: the internal [`Context`] is cleared at the
/// start of every [`scan`](Self::scan) call, so a single instance can
/// process many independent inputs sequentially.
pub struct Scanner<T: ContentType> {
    filter: Option<Filter>,
    identifiers: IdentifierSet<T>,
    analyzers: AnalyzerList<Box<dyn ContentAnalyzer<T>>>,
    extractors: ExtractorList<Box<dyn ContentExtractor<T>>, T>,
    context: Context<T>,
    max_depth: u32,
}
impl<T: ContentType> Scanner<T> {
    const EMPTY_VAR_MAP: varmap::VarMap = varmap::VarMap::new();
    /// Scans a single top-level [`Content`] and returns the results.
    ///
    /// The scanner:
    ///
    /// 1. Clears its internal [`Context`] so that no state leaks
    ///    between calls.
    /// 2. Applies the configured [`Filter`], if any, to the top-level
    ///    content — but only when `filter_root` is `true`. If the
    ///    filter rejects it, the returned [`ScanResult`] contains no
    ///    objects.
    /// 3. Recursively identifies, analyzes, and extracts nested
    ///    content up to the configured
    ///    [`max_depth`](ScannerBuilder::max_depth).
    ///
    /// Pass `filter_root = false` when the root is a container the
    /// filter was not written to accept, such as a
    /// [`FolderContent`](crate::FolderContent) scanned with a filter
    /// that only allows certain file extensions. Extracted children
    /// are filtered either way, unless their
    /// [`Entry`](crate::Entry) sets `skip_from_filtering`.
    ///
    /// The returned [`ScanResult`] borrows from `self` and stays
    /// valid until the next call on this scanner. Copy anything you
    /// need to keep out before starting another scan.
    pub fn scan<'a>(&'a mut self, content: &mut dyn Content<T>, filter_root: bool) -> ScanResult<'a, T> {
        self.context.clear();
        if filter_root {
            if let Some(filter) = self.filter.as_mut() {
                if !filter.should_process(content.path(), content.size()) {
                    return ScanResult::new(&self.context);
                }
            }
        }
        self.inner_scan(content, 1, Object::INVALID_INDEX);
        ScanResult::new(&self.context)
    }
    fn inner_scan(&mut self, content: &mut dyn Content<T>, depth: u32, parent_index: u32) -> NextAction {
        self.context.local_varmap_handle = None; // so that next time someone ask for a local varmap, it will get one from the context varmap_pool
        self.context.current_object_index = None;
        let start_req_count = self.context.extraction_requests_stack.len();
        let (ty, my_index) = self.create_object(content, parent_index);

        let mut response = self.run_analyzers(content, ty);
        if response == NextAction::Continue {
            response = self.run_extractors(content, ty, depth, my_index, start_req_count);
        }
        self.restore_extraction_request_stack(start_req_count);
        match response {
            NextAction::Continue | NextAction::Skip => NextAction::Continue,
            NextAction::Exit => NextAction::Exit,
        }
    }
    fn create_object(&mut self, content: &mut dyn Content<T>, parent_index: u32) -> (Option<T>, u32) {
        let ty = self.retrieve_content_type(content);

        let path_index = self.context.path_arena.alloc(content.path().as_printable_string().as_bytes());
        let obj = Object {
            path: path_index,
            parent_index,
            next_sibling_index: Object::INVALID_INDEX,
            varmap_handle: None,
            first_child_index: Object::INVALID_INDEX,
            last_child_index: Object::INVALID_INDEX,
            type_id: if let Some(ty) = ty { ty.as_u16() } else { u16::MAX },
        };
        let my_index = self.context.objects.len() as u32;
        self.context.objects.push(obj);
        self.context.current_object_index = Some(my_index);
        // links to parent and sibling
        let mut last_sibling_index = Object::INVALID_INDEX;
        if parent_index != Object::INVALID_INDEX {
            if let Some(parent) = self.context.objects.get_mut(parent_index as usize) {
                if parent.first_child_index == Object::INVALID_INDEX {
                    parent.first_child_index = my_index;
                } else {
                    last_sibling_index = parent.last_child_index;
                }
                parent.last_child_index = my_index;
                //println!("Parent: {}, Me: {}, First: {}, Last: {} -> {}", parent_index, my_index, parent.first_child_index, parent.last_child_index, content.path());
            }
        }
        if last_sibling_index != Object::INVALID_INDEX {
            if let Some(last_sibling) = self.context.objects.get_mut(last_sibling_index as usize) {
                last_sibling.next_sibling_index = my_index;
            }
        }
        (ty, my_index)
    }
    fn run_analyzers(&mut self, content: &mut dyn Content<T>, content_type: Option<T>) -> NextAction {
        if let Some(ty) = content_type {
            if let Some((start, end)) = self.analyzers.range(ty) {
                let res = self.scan_range(content, start, end);
                if matches!(res, NextAction::Skip | NextAction::Exit) {
                    return res;
                }
            }
        }
        // generic analyzers
        if let Some((start, end)) = self.analyzers.generic_range() {
            let res = self.scan_range(content, start, end);
            if matches!(res, NextAction::Skip | NextAction::Exit) {
                return res;
            }
        }
        NextAction::Continue
    }
    fn run_extractors(
        &mut self,
        content: &mut dyn Content<T>,
        content_type: Option<T>,
        depth: u32,
        my_index: u32,
        start_req_index: usize,
    ) -> NextAction {
        // type-specific extractors
        if let Some(ty) = content_type {
            if let Some((start, end)) = self.extractors.range(ty) {
                let res = self.extract_range(content, start, end, depth, my_index, None);
                if matches!(res, NextAction::Skip | NextAction::Exit) {
                    return res;
                }
            }
        }
        // run extraction requests
        let req_count = self.context.extraction_requests_stack.len();
        for i in start_req_index..req_count {
            let ty = self.context.extraction_requests_stack[i].content_type;
            if let Some((start, end)) = self.extractors.range(ty) {
                let res = self.extract_range(content, start, end, depth, my_index, Some(i as u32));
                if matches!(res, NextAction::Skip | NextAction::Exit) {
                    return res;
                }
            } else {
                if let Some(handle) = self.context.extraction_requests_stack[i].params_handle {
                    self.context.varmap_pool.release(handle);
                    self.context.extraction_requests_stack[i].params_handle = None;
                }
            }
        }
        NextAction::Continue
    }
    fn restore_extraction_request_stack(&mut self, start_req_index: usize) {
        // first release the params
        for req in self.context.extraction_requests_stack.iter_mut().skip(start_req_index) {
            if let Some(handle) = req.params_handle {
                self.context.varmap_pool.release(handle);
            }
        }
        // restore the stack
        self.context.extraction_requests_stack.truncate(start_req_index);
    }
    fn scan_range(&mut self, content: &mut dyn Content<T>, start: usize, end: usize) -> NextAction {
        if (end <= start) || (end > self.analyzers.len()) {
            return NextAction::Continue;
        }
        for i in start..end {
            let result = unsafe { self.analyzers.get(i).analyze(content, &mut self.context) };
            match result {
                NextAction::Continue => continue,
                NextAction::Exit => return NextAction::Exit,
                NextAction::Skip => return NextAction::Skip,
            }
        }
        NextAction::Continue
    }
    fn extract_range(
        &mut self,
        content: &mut dyn Content<T>,
        start: usize,
        end: usize,
        depth: u32,
        parent_index: u32,
        req_index: Option<u32>,
    ) -> NextAction {
        if (end <= start) || (end > self.extractors.len()) {
            return NextAction::Continue;
        }
        let (ec, param_handle) = if let Some(req_index) = req_index {
            let request = &self.context.extraction_requests_stack[req_index as usize];
            let params = if let Some(handle) = request.params_handle {
                if self.context.varmap_pool.get(handle).is_some() {
                    Some(handle)
                } else {
                    None
                }
            } else {
                None
            };
            (
                ExtractionContext {
                    offset: request.start,
                    length: request.len,
                    params: &Self::EMPTY_VAR_MAP,
                },
                params,
            )
        } else {
            (
                ExtractionContext {
                    offset: 0,
                    length: Some(content.size()),
                    params: &Self::EMPTY_VAR_MAP,
                },
                None,
            )
        };
        let next_action = {
            let mut next_action = NextAction::Continue;
            for i in start..end {
                let result = self.extract_content(content, i, depth, parent_index, &ec, param_handle);
                match result {
                    NextAction::Continue => continue,
                    NextAction::Exit | NextAction::Skip => {
                        next_action = result;
                        break;
                    }
                }
            }
            next_action
        };
        if let Some(handle) = param_handle {
            self.context.varmap_pool.release(handle);
            if let Some(req_index) = req_index {
                self.context.extraction_requests_stack[req_index as usize].params_handle = None;
            }
        }
        next_action
    }
    fn extract_content(
        &mut self,
        content: &mut dyn Content<T>,
        index: usize,
        depth: u32,
        parent_index: u32,
        ec: &ExtractionContext,
        ph: Option<varmap::PoolHandle>,
    ) -> NextAction {
        if depth >= self.max_depth {
            return NextAction::Continue;
        }
        let len = self.extractors.len();
        if index >= len {
            return NextAction::Continue;
        }
        let mut extractor = unsafe { self.extractors.get(index) };
        let acquired = {
            let params = ph.and_then(|h| self.context.varmap_pool.get(h)).unwrap_or(ec.params);
            let acquire_ec = ExtractionContext {
                offset: ec.offset,
                length: ec.length,
                params,
            };
            extractor.acquire(content, &acquire_ec)
        };
        if let Some(handle) = acquired {
            while let Some(entry) = unsafe { self.extractors.get(index).advance(handle, content) } {
                if !entry.skip_from_filtering {
                    if let Some(filter) = self.filter.as_mut() {
                        if !filter.should_process(&entry.path, entry.size) {
                            // println!("Skip: {:?}", &entry.path);
                            continue;
                        }
                    }
                }
                extractor = unsafe { self.extractors.get(index) };
                if let Some(mut extracted_content) = extractor.extract(handle, content) {
                    let result = self.inner_scan(&mut *extracted_content, depth + 1, parent_index);
                    match result {
                        NextAction::Continue => continue,
                        NextAction::Exit => {
                            extractor = unsafe { self.extractors.get(index) };
                            extractor.release(handle);
                            return NextAction::Exit;
                        }
                        NextAction::Skip => {
                            extractor = unsafe { self.extractors.get(index) };
                            extractor.release(handle);
                            return NextAction::Continue;
                        }
                    }
                }
            }
            extractor = unsafe { self.extractors.get(index) };
            extractor.release(handle);
        }
        NextAction::Continue
    }
    fn retrieve_content_type(&mut self, content: &mut dyn Content<T>) -> Option<T> {
        if let Some(ty) = content.content_type() {
            return Some(ty);
        }
        let p = content.path().as_bytes();
        // type from file name
        let file_name = utils::get_file_name(p);
        let type_from_file_name = self.identifiers.type_from_file_name(file_name);
        let extension = utils::get_extension(file_name);
        let type_from_extension = self.identifiers.type_from_extension(extension);
        // type from magic
        let type_from_magic = {
            if let Some(buf) = content.read(0, 16) {
                self.identifiers.type_from_magic(buf)
            } else {
                None
            }
        };
        if let Some(ty) = type_from_magic {
            if self.validate_content_type(content, ty) {
                return Some(ty);
            }
        }
        if let Some(ty) = type_from_file_name {
            if self.validate_content_type(content, ty) {
                return Some(ty);
            }
        }
        if let Some(ty) = type_from_extension {
            if self.validate_content_type(content, ty) {
                return Some(ty);
            }
        }
        for &ty in self.identifiers.identifiers_without_prefilter() {
            if self.validate_content_type(content, ty) {
                return Some(ty);
            }
        }
        None
    }
    #[inline(always)]
    fn validate_content_type(&self, content: &mut dyn Content<T>, content_type: T) -> bool {
        self.identifiers
            .get(content_type)
            .map(|identifier| identifier.validate(content))
            .unwrap_or(false)
    }
}
/// Fluent builder for [`Scanner`].
///
/// Register identifiers, analyzers, extractors, and (optionally) a
/// [`Filter`] with the `add_*` / [`filter`](Self::filter) methods,
/// then call [`build`](Self::build) to obtain a ready-to-use
/// [`Scanner`].
///
/// # Priorities
///
/// Analyzers are registered with a `priority` byte. Within the same
/// [`ContentType`] bucket (or within the generic bucket), they run in
/// **ascending** priority order — lower numbers first.
///
/// Extractors have no priority: multiple extractors for the same type
/// run in registration order.
///
/// # Generic vs. typed plugins
///
/// - `add_analyzer` registers a plugin against a specific
///   [`ContentType`]. It only runs when the scanner has positively
///   identified content of that type.
/// - `add_extractor` registers a plugin against a specific
///   [`ContentType`]. It runs when the current object is that type,
///   and also when an analyzer requested that type via
///   [`Context::request_extract`](crate::Context::request_extract).
/// - `add_generic_analyzer` registers an analyzer that runs for every
///   content object, regardless of type (including unidentified
///   content). Generic analyzers run after the type-specific ones.
pub struct ScannerBuilder<T: ContentType> {
    filter: Option<Filter>,
    analyzers: Vec<(u32, Box<dyn ContentAnalyzer<T>>)>,
    extractors: Vec<(T, Box<dyn ContentExtractor<T>>)>,
    identifiers: Vec<(T, Box<dyn ContentIdentifier<T>>)>,
    max_depth: u32,
}
impl<T: ContentType> ScannerBuilder<T> {
    /// Creates a new, empty builder with sensible defaults.
    ///
    /// The default maximum recursion depth is `8`. Change it with
    /// [`max_depth`](Self::max_depth).
    pub fn new() -> Self {
        Self {
            filter: None,
            analyzers: Vec::with_capacity(16),
            extractors: Vec::with_capacity(4),
            identifiers: Vec::with_capacity(4),
            max_depth: 8,
        }
    }

    /// Attaches a pre-built [`Filter`] to the scanner.
    ///
    /// The filter is consulted for every content object (top-level
    /// and extracted) before any plugin runs. If a filter was
    /// previously set, it is replaced.
    pub fn filter(mut self, filter: Filter) -> Self {
        self.filter = Some(filter);
        self
    }

    /// Registers an analyzer for a specific `content_type`.
    ///
    /// `priority` (`0..=255`) orders analyzers registered for the
    /// same type; lower values run first. Multiple analyzers for the
    /// same `(content_type, priority)` are allowed and their relative
    /// order is unspecified.
    pub fn add_analyzer<A>(mut self, content_type: T, priority: u8, analyzer: A) -> Self
    where
        A: ContentAnalyzer<T> + 'static,
    {
        let hash = (content_type.as_u16() as u32) << 16 | priority as u32;
        self.analyzers.push((hash, Box::new(analyzer)));
        self
    }

    /// Registers a generic analyzer that runs on every content object.
    ///
    /// Generic analyzers run after all type-specific analyzers for a
    /// given object. `priority` (`0..=255`) orders generic analyzers
    /// among themselves; lower values run first.
    pub fn add_generic_analyzer<A>(mut self, priority: u8, analyzer: A) -> Self
    where
        A: ContentAnalyzer<T> + 'static,
    {
        let hash = 0xFFFF0000 | priority as u32;
        self.analyzers.push((hash, Box::new(analyzer)));
        self
    }

    /// Registers an extractor for a specific `content_type`.
    ///
    /// The extractor runs when the scanner has identified content of
    /// that type, and also when an analyzer requested that type via
    /// [`Context::request_extract`](crate::Context::request_extract).
    /// Multiple extractors for the same type run in registration
    /// order. Extracted children are scanned recursively as long as
    /// [`max_depth`](Self::max_depth) allows.
    pub fn add_extractor<E>(mut self, content_type: T, extractor: E) -> Self
    where
        E: ContentExtractor<T> + 'static,
    {
        self.extractors.push((content_type, Box::new(extractor)));
        self
    }

    /// Registers an identifier for `content_type`.
    ///
    /// Only one identifier is allowed per content type; registering
    /// two identifiers for the same type will cause
    /// [`build`](Self::build) to panic. An [`IdentifyMethod::Magic`]
    /// or [`IdentifyMethod::MultipleMagic`] pattern longer than 16
    /// bytes also panics at build.
    pub fn add_identifier<I>(mut self, content_type: T, identifier: I) -> Self
    where
        I: ContentIdentifier<T> + 'static,
    {
        self.identifiers.push((content_type, Box::new(identifier)));
        self
    }

    /// Sets the maximum recursion depth for the scanner.
    ///
    /// The value is clamped to `1..=u32::MAX - 2`. The top-level
    /// content is at depth `1`; children extracted from it are at
    /// depth `2`, and so on. Extraction stops when the next child
    /// would exceed `max_depth`.
    pub fn max_depth(mut self, max_depth: u32) -> Self {
        self.max_depth = max_depth.clamp(1, u32::MAX - 2);
        self
    }

    fn check_unique_identifiers(&self) {
        let mut m = HashSet::new();
        for (content_type, _) in &self.identifiers {
            if m.contains(&content_type.as_u16()) {
                panic!(
                    "There can only be one identifier for type ! Type {:?} has multiple identifiers !",
                    content_type
                );
            }
            m.insert(content_type.as_u16());
        }
    }
    /// Consumes the builder and produces a ready-to-use [`Scanner`].
    ///
    /// # Panics
    ///
    /// Panics if two identifiers are registered for the same
    /// [`ContentType`], or if an [`IdentifyMethod::Magic`] /
    /// [`IdentifyMethod::MultipleMagic`] pattern is longer than 16
    /// bytes.
    pub fn build(self) -> Scanner<T> {
        self.check_unique_identifiers();
        let analyzers = AnalyzerList::new(self.analyzers, T::COUNT);
        let extractors = ExtractorList::new(self.extractors);
        let identifiers = IdentifierSet::new(self.identifiers);
        Scanner {
            filter: self.filter,
            identifiers,
            analyzers,
            extractors,
            context: Context::new(),
            max_depth: self.max_depth,
        }
    }
}