aria2-core 0.3.8

aria2-rust core download engine: multi-protocol segmented downloads, rate limiting, configuration, session persistence, and BitTorrent seeding
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
//! Batched disk writer: coalesces small writes into larger sequential flushes.
//!
//! This writer implements [`SeekableDiskWriter`] by buffering writes in a
//! `BTreeMap<u64, Bytes>` and flushing them when a configurable threshold
//! (total buffered bytes or pending write count) is exceeded. This reduces the
//! number of `pwrite`/`seek_write` syscalls for workloads with many small,
//! non-contiguous writes.

use async_trait::async_trait;
use bytes::{Bytes, BytesMut};
use std::collections::BTreeMap;
use std::path::{Path, PathBuf};
use tracing::debug;

use crate::error::{Aria2Error, Result};
use crate::filesystem::disk_writer::SeekableDiskWriter;
use crate::filesystem::positioned_disk_writer::PositionedDiskWriter;

pub struct BatchedDiskWriter {
    file: Option<PositionedDiskWriter>,
    path: PathBuf,
    buffer: BTreeMap<u64, Bytes>,
    flush_threshold_bytes: usize,
    total_buffered: usize,
    max_pending_writes: usize,
    opened: bool,
}

impl BatchedDiskWriter {
    pub fn new(path: &Path) -> Self {
        Self {
            file: None,
            path: path.to_path_buf(),
            buffer: BTreeMap::new(),
            flush_threshold_bytes: 256 * 1024,
            total_buffered: 0,
            max_pending_writes: 16,
            opened: false,
        }
    }

    pub fn with_threshold(mut self, bytes: usize) -> Self {
        self.flush_threshold_bytes = bytes;
        self
    }

    pub fn with_max_pending(mut self, max: usize) -> Self {
        self.max_pending_writes = max;
        self
    }

    /// Open (or create) the file without truncating existing data.
    async fn ensure_open(&mut self) -> Result<()> {
        if !self.opened {
            let mut writer = PositionedDiskWriter::new(&self.path, None);
            writer.open().await.map_err(|error| {
                Aria2Error::FileOpen(format!("Failed to open {}: {}", self.path.display(), error))
            })?;
            self.file = Some(writer);
            self.opened = true;
        }
        Ok(())
    }

    fn append_buffered(&mut self, offset: u64, data: Bytes) -> Result<()> {
        let end = offset.checked_add(data.len() as u64).ok_or_else(|| {
            Aria2Error::InvalidArgument("buffered write range exceeds u64 address space".into())
        })?;

        // Keep buffered ranges disjoint and merge adjacent ranges. Besides
        // reducing syscalls for sequential small writes, this preserves
        // last-write-wins behavior for overlapping writes.
        let overlapping_keys: Vec<u64> = self
            .buffer
            .range(..=end)
            .filter_map(|(&start, existing)| {
                let existing_end = start.checked_add(existing.len() as u64)?;
                (existing_end >= offset && start <= end).then_some(start)
            })
            .collect();

        let merge_start = overlapping_keys
            .first()
            .copied()
            .unwrap_or(offset)
            .min(offset);
        let merge_end = overlapping_keys
            .iter()
            .filter_map(|start| {
                self.buffer
                    .get(start)
                    .and_then(|existing| start.checked_add(existing.len() as u64))
            })
            .max()
            .unwrap_or(end)
            .max(end);

        let merged_len = usize::try_from(merge_end - merge_start)
            .map_err(|_| Aria2Error::InvalidArgument("buffered write range is too large".into()))?;
        let mut merged = BytesMut::zeroed(merged_len);

        for start in &overlapping_keys {
            let existing = self
                .buffer
                .get(start)
                .expect("buffered range disappeared during merge");
            let relative =
                usize::try_from(*start - merge_start).expect("buffered range offset fits in usize");
            merged[relative..relative + existing.len()].copy_from_slice(existing);
        }

        let relative =
            usize::try_from(offset - merge_start).expect("buffered range offset fits in usize");
        merged[relative..relative + data.len()].copy_from_slice(&data);

        for start in overlapping_keys {
            let existing = self
                .buffer
                .remove(&start)
                .expect("buffered range disappeared during merge");
            self.total_buffered -= existing.len();
        }
        self.buffer.insert(merge_start, merged.freeze());
        self.total_buffered += merged_len;
        Ok(())
    }

    fn should_flush(&self) -> bool {
        self.total_buffered >= self.flush_threshold_bytes
            || self.buffer.len() >= self.max_pending_writes
    }

    pub fn buffered_count(&self) -> usize {
        self.buffer.len()
    }

    pub fn buffered_bytes(&self) -> usize {
        self.total_buffered
    }
}

#[async_trait]
impl SeekableDiskWriter for BatchedDiskWriter {
    async fn open(&mut self) -> Result<()> {
        self.ensure_open().await
    }

    async fn write_at(&mut self, offset: u64, data: &[u8]) -> Result<()> {
        self.ensure_open().await?;

        if data.is_empty() {
            return Ok(());
        }

        self.append_buffered(offset, Bytes::copy_from_slice(data))?;

        if self.should_flush() {
            self.flush().await?;
        }

        Ok(())
    }

    async fn write_bytes_at(&mut self, offset: u64, data: bytes::Bytes) -> Result<()> {
        self.ensure_open().await?;

        if data.is_empty() {
            return Ok(());
        }

        self.append_buffered(offset, data)?;

        if self.should_flush() {
            self.flush().await?;
        }

        Ok(())
    }

    async fn read_at(&mut self, offset: u64, buf: &mut [u8]) -> Result<usize> {
        self.flush().await?;
        self.ensure_open().await?;
        let file = self.file.as_mut().ok_or_else(|| {
            Aria2Error::Io("file not open after ensure_open — invariant violated".into())
        })?;
        file.read_at(offset, buf).await
    }

    async fn truncate(&mut self, length: u64) -> Result<()> {
        self.flush().await?;
        self.ensure_open().await?;
        let file = self.file.as_mut().ok_or_else(|| {
            Aria2Error::Io("file not open after ensure_open — invariant violated".into())
        })?;
        file.truncate(length).await
    }

    async fn flush(&mut self) -> Result<()> {
        if self.buffer.is_empty() {
            return Ok(());
        }

        self.ensure_open().await?;
        let writer = self.file.as_mut().ok_or_else(|| {
            Aria2Error::Io("file not open after ensure_open — invariant violated".into())
        })?;

        debug!(
            "[BatchedDiskWriter] Flushing {} writes ({} bytes)",
            self.buffer.len(),
            self.total_buffered
        );

        for (&offset, data) in &self.buffer {
            writer.write_bytes_at(offset, data.clone()).await?;
        }

        writer.flush().await?;

        self.buffer.clear();
        self.total_buffered = 0;
        Ok(())
    }

    async fn len(&self) -> Result<u64> {
        match self.file.as_ref() {
            Some(writer) => writer.len().await,
            None => Ok(0),
        }
    }

    fn path(&self) -> &Path {
        &self.path
    }

    async fn close(&mut self) -> Result<()> {
        self.flush().await?;
        if let Some(mut writer) = self.file.take() {
            writer.close().await?;
        }
        self.opened = false;
        Ok(())
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use tempfile::TempDir;

    #[tokio::test]
    async fn test_new_writer() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let writer = BatchedDiskWriter::new(&path);

        assert!(!writer.opened);
        assert!(writer.file.is_none());
        assert_eq!(writer.buffered_count(), 0);
        assert_eq!(writer.buffered_bytes(), 0);
        assert_eq!(writer.path(), path.as_path());
    }

    #[tokio::test]
    async fn test_write_at_buffers() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path);

        writer.write_at(0, b"hello").await.unwrap();
        writer.write_at(100, b"world").await.unwrap();
        writer.write_at(200, b"!").await.unwrap();

        assert_eq!(writer.buffered_count(), 3);
        assert_eq!(writer.buffered_bytes(), 11);
        assert!(writer.opened);
    }

    #[tokio::test]
    async fn test_adjacent_writes_are_coalesced() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path).with_threshold(1024 * 1024);

        writer.write_at(100, b"world").await.unwrap();
        writer.write_at(105, b"!").await.unwrap();
        writer.write_at(94, b"hello ").await.unwrap();

        assert_eq!(writer.buffered_count(), 1);
        assert_eq!(writer.buffered_bytes(), 12);
        writer.flush().await.unwrap();

        let contents = tokio::fs::read(&path).await.unwrap();
        assert_eq!(&contents[94..106], b"hello world!");
    }

    #[tokio::test]
    async fn test_overlapping_writes_use_last_write() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path).with_threshold(1024 * 1024);

        writer.write_at(0, b"abcdef").await.unwrap();
        writer.write_at(2, b"XY").await.unwrap();

        assert_eq!(writer.buffered_count(), 1);
        assert_eq!(writer.buffered_bytes(), 6);
        writer.flush().await.unwrap();

        assert_eq!(tokio::fs::read(&path).await.unwrap(), b"abXYef");
    }

    #[tokio::test]
    async fn test_auto_flush_on_threshold() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path).with_threshold(64);

        let large_data = vec![0xABu8; 128];
        writer.write_at(0, &large_data).await.unwrap();

        assert_eq!(writer.buffered_count(), 0);
        assert_eq!(writer.buffered_bytes(), 0);
    }

    #[tokio::test]
    async fn test_auto_flush_on_max_pending() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path).with_max_pending(4);

        for i in 0..6u64 {
            writer.write_at(i * 1000, &[i as u8]).await.unwrap();
        }

        assert_eq!(writer.buffered_count(), 2);
    }

    #[tokio::test]
    async fn test_explicit_flush_writes_to_disk() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path).with_threshold(1024 * 1024);

        writer.write_at(0, b"hello ").await.unwrap();
        writer.write_at(6, b"world").await.unwrap();

        assert_eq!(writer.buffered_count(), 1);

        writer.flush().await.unwrap();
        assert_eq!(writer.buffered_count(), 0);

        let mut file = tokio::fs::File::open(&path).await.unwrap();
        use tokio::io::AsyncReadExt;
        let mut buf = Vec::new();
        file.read_to_end(&mut buf).await.unwrap();
        assert_eq!(&buf, b"hello world");
    }

    #[tokio::test]
    async fn test_close_finalizes() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path);

        writer.write_at(0, b"data").await.unwrap();
        writer.close().await.unwrap();

        assert!(!writer.opened);
        assert!(writer.file.is_none());
        assert_eq!(writer.buffered_count(), 0);

        use tokio::io::AsyncReadExt;
        let mut file = tokio::fs::File::open(&path).await.unwrap();
        let mut buf = Vec::new();
        file.read_to_end(&mut buf).await.unwrap();
        assert_eq!(&buf, b"data");
    }

    #[tokio::test]
    async fn test_sequential_ordering() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("test.bin");
        let mut writer = BatchedDiskWriter::new(&path).with_threshold(1024 * 1024);

        writer.write_at(100, b"B").await.unwrap();
        writer.write_at(50, b"A").await.unwrap();
        writer.write_at(200, b"C").await.unwrap();

        let offsets: Vec<u64> = writer.buffer.keys().copied().collect();
        assert_eq!(offsets, vec![50, 100, 200]);
    }

    #[tokio::test]
    async fn test_open_trait_method() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("trait_open.bin");
        let mut writer = BatchedDiskWriter::new(&path);

        // Use the trait method
        SeekableDiskWriter::open(&mut writer).await.unwrap();
        assert!(writer.opened);
    }

    #[tokio::test]
    async fn test_truncate_seekable_trait() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("truncate_test.bin");
        let mut writer = BatchedDiskWriter::new(&path);

        writer.write_at(0, b"hello world").await.unwrap();
        writer.flush().await.unwrap();

        // Use the trait method
        SeekableDiskWriter::truncate(&mut writer, 5).await.unwrap();

        let metadata = tokio::fs::metadata(&path).await.unwrap();
        assert_eq!(metadata.len(), 5);
    }

    #[tokio::test]
    async fn test_read_at_seekable_trait() {
        let dir = TempDir::new().unwrap();
        let path = dir.path().join("read_test.bin");
        let mut writer = BatchedDiskWriter::new(&path);

        writer.write_at(0, b"hello world").await.unwrap();
        writer.flush().await.unwrap();

        let mut buf = vec![0u8; 5];
        let n = SeekableDiskWriter::read_at(&mut writer, 6, &mut buf)
            .await
            .unwrap();
        assert_eq!(n, 5);
        assert_eq!(&buf, b"world");
    }
}