squeez 1.2.0

End-to-end token optimizer for Claude Code, Copilot CLI, and OpenCode. Compresses bash output up to 95%, collapses redundant calls, injects caveman persona, and compresses memory files. Zero new dependencies.
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

use squeez::strategies::grouping::group_files_by_dir;

#[test]
fn groups_many_files_in_same_dir() {
    let files: Vec<String> = (0..6)
        .map(|i| format!("\tmodified:   src/components/Comp{}.tsx", i))
        .collect();
    let result = group_files_by_dir(files, 4);
    assert_eq!(result.len(), 1);
    assert!(result[0].contains("src/components"));
    assert!(result[0].contains("6 modified"));
}

#[test]
fn keeps_files_below_threshold() {
    let files = vec![
        "modified:   src/a.ts".to_string(),
        "modified:   src/b.ts".to_string(),
    ];
    assert_eq!(group_files_by_dir(files, 4).len(), 2);
}

#[test]
fn mixes_grouped_and_ungrouped() {
    let mut files: Vec<String> = (0..5)
        .map(|i| format!("modified:   src/components/C{}.tsx", i))
        .collect();
    files.push("modified:   README.md".to_string());
    let result = group_files_by_dir(files, 4);
    assert_eq!(result.len(), 2); // components/ grouped + README.md kept
}