stt-cli 0.2.1

Speech to text Cli using Groq API and OpenAI API
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

// providers/async_openai.rs
use crate::providers::TranscriptionProvider;
use anyhow::Result;
use async_trait::async_trait;
use bytes::Bytes;
use tracing::{debug, error, info, instrument};

pub struct OpenAIProvider {
    client: async_openai::Client<async_openai::config::OpenAIConfig>,
}

impl OpenAIProvider {
    pub async fn new() -> Self {
        Self {
            client: async_openai::Client::new(),
        }
    }
}

#[async_trait]
impl TranscriptionProvider for OpenAIProvider {
    fn name(&self) -> &'static str {
        "OpenAI"
    }

    fn min_chunk_duration(&self) -> std::time::Duration {
        std::time::Duration::from_secs(5)
    }

    #[instrument(skip(self, audio_data))]
    async fn transcribe(&self, audio_data: &[u8]) -> Result<String> {
        use async_openai::types::{AudioInput, AudioResponseFormat, CreateTranscriptionRequest};

        info!("Starting OpenAI transcription request");
        debug!("Audio data size: {} bytes", audio_data.len());

        // Validate audio meets minimum duration requirements (5 seconds @ 16kHz)
        let min_required_bytes = 16000 * 2 * 5; // 16kHz * 16bits * 5s
        if audio_data.len() < min_required_bytes {
            return Err(anyhow::anyhow!(
                "Audio chunk too short ({} bytes < {} bytes). Minimum 5 seconds required",
                audio_data.len(),
                min_required_bytes
            ));
        }

        // Create audio input from raw bytes
        let audio_input =
            AudioInput::from_bytes("audio.wav".to_string(), Bytes::copy_from_slice(audio_data));

        let request = CreateTranscriptionRequest {
            file: audio_input,
            model: "whisper-1".to_string(),
            response_format: Some(AudioResponseFormat::Json),
            ..Default::default()
        };

        debug!("Sending request to OpenAI Whisper API");
        match self.client.audio().transcribe(request).await {
            Ok(response) => {
                info!("OpenAI transcription successful");
                info!("Received transcription: {}", response.text);
                Ok(response.text)
            }
            Err(e) => {
                error!("OpenAI transcription failed: {}", e);
                Err(anyhow::anyhow!("OpenAI API error: {}", e))
            }
        }
    }
}