token-count 0.4.0

Count tokens for LLM models using exact tokenization
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

//! Integration tests for Google Gemini tokenization
//!
//! Tests CLI integration with Google Gemini models using the gemini-tokenizer
//! library for exact, offline token counting.

use assert_cmd::Command;
use predicates::prelude::*;

#[test]
fn test_gemini_default_alias() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini"]).write_stdin("Hello, Gemini!").assert().success().stdout("4\n");
}

#[test]
fn test_gemini_pro_alias() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini-pro"]).write_stdin("test").assert().success().stdout("1\n");
}

#[test]
fn test_all_gemini_models() {
    // Test all 4 supported models
    let models =
        vec!["gemini-2.5-pro", "gemini-2.5-flash", "gemini-2.5-flash-lite", "gemini-3-pro-preview"];

    for model in models {
        let mut cmd = Command::cargo_bin("token-count").unwrap();
        cmd.args(["--model", model]).write_stdin("test").assert().success().stdout("1\n");
    }
}

#[test]
fn test_gemini_case_insensitive() {
    for variant in &["gemini", "GEMINI", "Gemini", "GeMiNi"] {
        let mut cmd = Command::cargo_bin("token-count").unwrap();
        cmd.args(["--model", variant]).write_stdin("test").assert().success().stdout("1\n");
    }
}

#[test]
fn test_gemini_provider_format() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "google/gemini"]).write_stdin("test").assert().success().stdout("1\n");
}

#[test]
fn test_gemini_verbose() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini", "-vv"])
        .write_stdin("test")
        .assert()
        .success()
        .stdout(predicate::str::contains("Model: gemini-2.5-flash"))
        .stdout(predicate::str::contains("gemini-gemma3"))
        .stdout(predicate::str::contains("Tokens: 1"))
        .stdout(predicate::str::contains("Context window: 1000000"));
}

#[test]
fn test_gemini_3_pro_preview() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini-3-pro-preview", "-vv"])
        .write_stdin("test")
        .assert()
        .success()
        .stdout(predicate::str::contains("Context window: 1000000"));
}

#[test]
fn test_gemini_list_models() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.arg("--list-models")
        .assert()
        .success()
        .stdout(predicate::str::contains("gemini-2.5-flash"))
        .stdout(predicate::str::contains("gemini-gemma3"));
}

#[test]
fn test_gemini_unknown_model() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini-4"])
        .write_stdin("test")
        .assert()
        .failure()
        .stderr(predicate::str::contains("Unknown model: 'gemini-4'"));
}

#[test]
fn test_gemini_empty_input() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini"]).write_stdin("").assert().success().stdout("0\n");
}

#[test]
fn test_gemini_multiline() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini"])
        .write_stdin("Line 1\nLine 2\nLine 3")
        .assert()
        .success()
        .stdout("11\n");
}

#[test]
fn test_gemini_unicode() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini"]).write_stdin("Hello δΈ–η•Œ 🌍").assert().success();
    // Note: exact count may vary, just verify it succeeds
}

#[test]
fn test_gemini_long_text() {
    let long_text = "The quick brown fox jumps over the lazy dog. ".repeat(100);
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini"]).write_stdin(long_text.as_str()).assert().success();
    // Should succeed with reasonable token count
}

#[test]
fn test_gemini_flash_alias() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini-flash"]).write_stdin("test").assert().success().stdout("1\n");
}

#[test]
fn test_gemini_lite_alias() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini-lite"]).write_stdin("test").assert().success().stdout("1\n");
}

#[test]
fn test_gemini_3_pro_alias() {
    let mut cmd = Command::cargo_bin("token-count").unwrap();
    cmd.args(["--model", "gemini-3-pro"]).write_stdin("test").assert().success().stdout("1\n");
}

#[test]
fn test_gemini_comparison_openai() {
    // Verify Gemini and OpenAI can both tokenize the same text
    let test_text = "Hello world";

    let mut gemini_cmd = Command::cargo_bin("token-count").unwrap();
    let gemini_output =
        gemini_cmd.args(["--model", "gemini"]).write_stdin(test_text).output().unwrap();

    let mut openai_cmd = Command::cargo_bin("token-count").unwrap();
    let openai_output =
        openai_cmd.args(["--model", "gpt-4"]).write_stdin(test_text).output().unwrap();

    // Both should succeed
    assert!(gemini_output.status.success());
    assert!(openai_output.status.success());

    // Both should produce valid token counts (may differ)
    let gemini_count = String::from_utf8(gemini_output.stdout).unwrap();
    let openai_count = String::from_utf8(openai_output.stdout).unwrap();

    assert!(!gemini_count.trim().is_empty());
    assert!(!openai_count.trim().is_empty());
}