nab 0.7.1

Token-optimized HTTP client for LLMs — fetches any URL as clean markdown
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

//! JSON-file persistence for [`Watch`] metadata.
//!
//! Each watch is stored as `<storage_dir>/<id>.json`.
//! Snapshot body bytes are stored content-addressed at `<snapshot_dir>/<sha256>`.

use std::collections::HashSet;
use std::fs;
use std::path::{Path, PathBuf};

use anyhow::{Context, Result};

use super::types::{Watch, WatchId, WatchSnapshot};

// ─── Watch persistence ────────────────────────────────────────────────────────

/// Persist a watch's metadata to `<dir>/<id>.json`.
pub fn save_watch(dir: &Path, watch: &Watch) -> Result<()> {
    fs::create_dir_all(dir)
        .with_context(|| format!("create watch storage dir {}", dir.display()))?;
    let path = watch_path(dir, &watch.id);
    let bytes = serde_json::to_vec_pretty(watch).context("serialize watch")?;
    atomic_write(&path, &bytes)
}

/// Load a watch from `<dir>/<id>.json`, returning `None` if it does not exist.
pub fn load_watch(dir: &Path, id: &WatchId) -> Option<Watch> {
    let bytes = fs::read(watch_path(dir, id)).ok()?;
    serde_json::from_slice(&bytes).ok()
}

/// Delete a watch file from `<dir>/<id>.json`.
pub fn delete_watch(dir: &Path, id: &WatchId) -> Result<()> {
    let path = watch_path(dir, id);
    fs::remove_file(&path).with_context(|| format!("remove watch file {}", path.display()))
}

/// Load all watches from `dir`, silently skipping corrupt files.
pub fn load_all_watches(dir: &Path) -> Vec<Watch> {
    let Ok(entries) = fs::read_dir(dir) else {
        return vec![];
    };
    entries
        .flatten()
        .filter(|e| {
            e.path()
                .extension()
                .is_some_and(|ext| ext.eq_ignore_ascii_case("json"))
        })
        .filter_map(|e| {
            let bytes = fs::read(e.path()).ok()?;
            serde_json::from_slice(&bytes).ok()
        })
        .collect()
}

fn watch_path(dir: &Path, id: &WatchId) -> PathBuf {
    dir.join(format!("{id}.json"))
}

// ─── Snapshot body storage ────────────────────────────────────────────────────

/// Save body bytes to `<snapshot_dir>/<sha256>` (content-addressed, idempotent).
pub fn save_snapshot_body(snapshot_dir: &Path, sha256: &str, body: &[u8]) -> Result<()> {
    fs::create_dir_all(snapshot_dir)
        .with_context(|| format!("create snapshot dir {}", snapshot_dir.display()))?;
    let path = snapshot_dir.join(sha256);
    if path.exists() {
        // Deduplicated — identical content already stored.
        return Ok(());
    }
    atomic_write(&path, body)
}

/// Load snapshot body bytes from content-addressed storage.
pub fn load_snapshot_body(snapshot_dir: &Path, sha256: &str) -> Option<Vec<u8>> {
    fs::read(snapshot_dir.join(sha256)).ok()
}

/// Garbage-collect snapshot files that are no longer referenced by any watch.
///
/// Called after a watch is removed or its snapshot list is pruned.
pub fn gc_snapshots<S: std::hash::BuildHasher>(
    snapshot_dir: &Path,
    referenced: &HashSet<String, S>,
) {
    let Ok(entries) = fs::read_dir(snapshot_dir) else {
        return;
    };
    for entry in entries.flatten() {
        if let Some(name) = entry.file_name().to_str()
            && !referenced.contains(name)
        {
            let _ = fs::remove_file(entry.path()); // best-effort
        }
    }
}

/// Prune snapshot metadata on a watch to at most `max` entries (newest retained).
/// Returns the SHA-256 hashes that were removed.
pub fn prune_snapshots(snapshots: &mut Vec<WatchSnapshot>, max: usize) -> Vec<String> {
    if snapshots.len() <= max {
        return vec![];
    }
    // Sort newest-first so we can drain the tail.
    snapshots.sort_by(|a, b| b.captured_at.cmp(&a.captured_at));
    let removed: Vec<String> = snapshots.drain(max..).map(|s| s.sha256).collect();
    removed
}

// ─── Atomic write ─────────────────────────────────────────────────────────────

fn atomic_write(path: &Path, data: &[u8]) -> Result<()> {
    let tmp = path.with_extension("tmp");
    fs::write(&tmp, data).with_context(|| format!("write tmp file {}", tmp.display()))?;
    fs::rename(&tmp, path)
        .with_context(|| format!("rename {} -> {}", tmp.display(), path.display()))
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::watch::types::WatchOptions;
    use chrono::Utc;
    use tempfile::TempDir;

    fn tmp() -> TempDir {
        tempfile::tempdir().expect("tmp dir")
    }

    fn make_watch(id: &str, url: &str) -> Watch {
        Watch {
            id: id.into(),
            url: url.into(),
            selector: None,
            interval_secs: 3600,
            created_at: Utc::now(),
            last_check_at: None,
            last_change_at: None,
            last_etag: None,
            last_last_modified: None,
            snapshots: vec![],
            consecutive_errors: 0,
            options: WatchOptions::default(),
        }
    }

    #[test]
    fn save_and_load_roundtrips() {
        // GIVEN
        let dir = tmp();
        let w = make_watch("abc12345", "https://example.com");
        // WHEN
        save_watch(dir.path(), &w).unwrap();
        let loaded = load_watch(dir.path(), &w.id).unwrap();
        // THEN
        assert_eq!(loaded.url, w.url);
        assert_eq!(loaded.id, w.id);
    }

    #[test]
    fn delete_removes_file() {
        // GIVEN
        let dir = tmp();
        let w = make_watch("del00001", "https://del.com");
        save_watch(dir.path(), &w).unwrap();
        // WHEN
        delete_watch(dir.path(), &w.id).unwrap();
        // THEN
        assert!(load_watch(dir.path(), &w.id).is_none());
    }

    #[test]
    fn load_all_watches_finds_saved() {
        // GIVEN
        let dir = tmp();
        for i in 0..3usize {
            save_watch(
                dir.path(),
                &make_watch(&format!("watch{i:05}"), &format!("https://{i}.com")),
            )
            .unwrap();
        }
        // WHEN
        let watches = load_all_watches(dir.path());
        // THEN
        assert_eq!(watches.len(), 3);
    }

    #[test]
    fn snapshot_body_dedup_shares_file() {
        // GIVEN
        let dir = tmp();
        let body = b"identical content";
        let sha = "deadbeef";
        save_snapshot_body(dir.path(), sha, body).unwrap();
        // WHEN: save same hash again
        save_snapshot_body(dir.path(), sha, body).unwrap();
        // THEN: only one file exists
        let count = fs::read_dir(dir.path()).unwrap().count();
        assert_eq!(count, 1);
    }

    #[test]
    fn prune_snapshots_keeps_newest() {
        // GIVEN
        let mut snaps: Vec<WatchSnapshot> = (0..5u64)
            .map(|i| WatchSnapshot {
                sha256: format!("hash{i}"),
                captured_at: chrono::DateTime::from_timestamp(i as i64, 0).unwrap(),
                size: 100,
            })
            .collect();
        // WHEN
        let removed = prune_snapshots(&mut snaps, 3);
        // THEN: 3 retained, 2 removed
        assert_eq!(snaps.len(), 3);
        assert_eq!(removed.len(), 2);
        // Newest (ts=4,3,2) retained
        assert!(snaps.iter().any(|s| s.sha256 == "hash4"));
    }
}