llmshim 0.1.26

Blazing fast LLM API translation layer in pure Rust
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

use chrono::Utc;
use serde::Serialize;
use serde_json::Value;
use std::io::Write;
use std::sync::{Arc, Mutex};
use std::time::Instant;

/// A single log entry for an LLM request.
#[derive(Debug, Clone, Serialize)]
pub struct LogEntry {
    pub ts: String,
    pub model: String,
    pub provider: String,
    pub latency_ms: u64,
    pub input_tokens: u64,
    pub output_tokens: u64,
    pub reasoning_tokens: u64,
    pub total_tokens: u64,
    pub status: String,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub error: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub request_id: Option<String>,
}

impl LogEntry {
    /// Extract token counts from a normalized OpenAI-format response.
    pub fn from_response(
        provider: &str,
        model: &str,
        response: &Value,
        latency: std::time::Duration,
    ) -> Self {
        let usage = response
            .get("usage")
            .cloned()
            .unwrap_or(serde_json::json!({}));
        Self {
            ts: Utc::now().to_rfc3339_opts(chrono::SecondsFormat::Millis, true),
            model: model.to_string(),
            provider: provider.to_string(),
            latency_ms: latency.as_millis() as u64,
            input_tokens: usage
                .get("prompt_tokens")
                .and_then(|v| v.as_u64())
                .unwrap_or(0),
            output_tokens: usage
                .get("completion_tokens")
                .and_then(|v| v.as_u64())
                .unwrap_or(0),
            reasoning_tokens: usage
                .get("reasoning_tokens")
                .and_then(|v| v.as_u64())
                .unwrap_or(0),
            total_tokens: usage
                .get("total_tokens")
                .and_then(|v| v.as_u64())
                .unwrap_or(0),
            status: "ok".to_string(),
            error: None,
            request_id: response
                .get("id")
                .and_then(|v| v.as_str())
                .map(String::from),
        }
    }

    pub fn from_error(
        provider: &str,
        model: &str,
        error: &str,
        latency: std::time::Duration,
    ) -> Self {
        Self {
            ts: Utc::now().to_rfc3339_opts(chrono::SecondsFormat::Millis, true),
            model: model.to_string(),
            provider: provider.to_string(),
            latency_ms: latency.as_millis() as u64,
            input_tokens: 0,
            output_tokens: 0,
            reasoning_tokens: 0,
            total_tokens: 0,
            status: "error".to_string(),
            error: Some(error.to_string()),
            request_id: None,
        }
    }
}

/// Logger that writes JSONL to a writer (file or stdout).
pub struct Logger {
    writer: Arc<Mutex<Box<dyn Write + Send>>>,
}

impl Logger {
    /// Create a logger that writes to a file.
    pub fn to_file(path: &str) -> std::io::Result<Self> {
        let file = std::fs::OpenOptions::new()
            .create(true)
            .append(true)
            .open(path)?;
        Ok(Self {
            writer: Arc::new(Mutex::new(Box::new(file))),
        })
    }

    /// Create a logger that writes to stderr (won't interfere with CLI output).
    pub fn to_stderr() -> Self {
        Self {
            writer: Arc::new(Mutex::new(Box::new(std::io::stderr()))),
        }
    }

    pub fn log(&self, entry: &LogEntry) {
        if let Ok(json) = serde_json::to_string(entry) {
            if let Ok(mut writer) = self.writer.lock() {
                let _ = writeln!(writer, "{}", json);
            }
        }
    }
}

/// Timer helper — start before request, finish after.
pub struct RequestTimer {
    start: Instant,
}

impl RequestTimer {
    pub fn start() -> Self {
        Self {
            start: Instant::now(),
        }
    }

    pub fn elapsed(&self) -> std::time::Duration {
        self.start.elapsed()
    }
}