aither 0.2.0

Providing unified trait abstractions for AI models
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
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285

use alloc::{string::String, vec::Vec};
use futures_core::Stream;

/// Audio data as bytes.
///
/// Type alias for [`Vec<u8>`] representing raw audio data.
pub type Data = Vec<u8>;

/// Generates audio from text prompts.
/// # Example
///
/// ```rust
/// use aither::AudioGenerator;
/// use futures_core::Stream;
///
/// struct MyAudioGen;
///
/// impl AudioGenerator for MyAudioGen {
///     fn generate(&self, prompt: &str) -> impl Stream<Item = aither::audio::Data> + Send {
///         futures_lite::stream::iter(Some(vec![0u8; 1024]))
///     }
/// }
/// ```
pub trait AudioGenerator {
    /// Generates audio from text prompt.
    ///
    /// Returns a [`Stream`] of [`Data`] chunks.
    fn generate(&self, prompt: &str) -> impl Stream<Item = Data> + Send;
}

/// Transcribes audio to text.
///
/// # Example
///
/// ```rust
/// use aither::AudioTranscriber;
/// use futures_core::Stream;
///
/// struct MyTranscriber;
///
/// impl AudioTranscriber for MyTranscriber {
///     fn transcribe(&self, audio: &[u8]) -> impl Stream<Item = String> + Send {
///         futures_lite::stream::iter(vec!["Hello world".to_string()])
///     }
/// }
/// ```
pub trait AudioTranscriber {
    /// Transcribes audio data to text.
    ///
    /// Returns a [`Stream`] of transcribed text chunks.
    fn transcribe(&self, audio: &[u8]) -> impl Stream<Item = String> + Send;
}

#[cfg(test)]
mod tests {
    use super::*;
    use alloc::{string::ToString, vec};
    use futures_lite::StreamExt;

    struct MockAudioGenerator;

    impl AudioGenerator for MockAudioGenerator {
        fn generate(&self, prompt: &str) -> impl Stream<Item = Data> + Send {
            // Generate mock audio data based on prompt length
            let chunks = if prompt.is_empty() {
                vec![]
            } else if prompt.len() < 10 {
                vec![vec![0x01; 512]] // Short audio for short prompts
            } else {
                vec![
                    vec![0x01; 512],  // First chunk
                    vec![0x02; 1024], // Second chunk
                    vec![0x03; 256],  // Final chunk
                ]
            };

            futures_lite::stream::iter(chunks)
        }
    }

    struct MockAudioTranscriber;

    impl AudioTranscriber for MockAudioTranscriber {
        fn transcribe(&self, audio: &[u8]) -> impl Stream<Item = String> + Send {
            // Generate mock transcription based on audio length
            let text_chunks = if audio.is_empty() {
                vec![]
            } else if audio.len() < 100 {
                vec!["Short".to_string()]
            } else if audio.len() < 1000 {
                vec!["Hello".to_string(), " world".to_string()]
            } else {
                vec![
                    "This".to_string(),
                    " is".to_string(),
                    " a".to_string(),
                    " longer".to_string(),
                    " transcription".to_string(),
                ]
            };

            futures_lite::stream::iter(text_chunks)
        }
    }

    #[tokio::test]
    async fn audio_generator_short_prompt() {
        let generator = MockAudioGenerator;
        let mut stream = generator.generate("Hi");

        let mut chunks = Vec::new();
        while let Some(chunk) = stream.next().await {
            chunks.push(chunk);
        }

        assert_eq!(chunks.len(), 1);
        assert_eq!(chunks[0].len(), 512);
        assert_eq!(chunks[0][0], 0x01);
    }

    #[tokio::test]
    async fn audio_generator_long_prompt() {
        let generator = MockAudioGenerator;
        let mut stream = generator
            .generate("This is a longer prompt that should generate multiple audio chunks");

        let mut chunks = Vec::new();
        while let Some(chunk) = stream.next().await {
            chunks.push(chunk);
        }

        assert_eq!(chunks.len(), 3);
        assert_eq!(chunks[0].len(), 512);
        assert_eq!(chunks[1].len(), 1024);
        assert_eq!(chunks[2].len(), 256);

        assert_eq!(chunks[0][0], 0x01);
        assert_eq!(chunks[1][0], 0x02);
        assert_eq!(chunks[2][0], 0x03);
    }

    #[tokio::test]
    async fn audio_generator_empty_prompt() {
        let generator = MockAudioGenerator;
        let mut stream = generator.generate("");

        let mut chunks = Vec::new();
        while let Some(chunk) = stream.next().await {
            chunks.push(chunk);
        }

        assert!(chunks.is_empty());
    }

    #[tokio::test]
    async fn audio_transcriber_short_audio() {
        let transcriber = MockAudioTranscriber;
        let audio_data = vec![0x01; 50]; // Short audio
        let mut stream = transcriber.transcribe(&audio_data);

        let mut text_chunks = Vec::new();
        while let Some(chunk) = stream.next().await {
            text_chunks.push(chunk);
        }

        assert_eq!(text_chunks.len(), 1);
        assert_eq!(text_chunks[0], "Short");
    }

    #[tokio::test]
    async fn audio_transcriber_medium_audio() {
        let transcriber = MockAudioTranscriber;
        let audio_data = vec![0x01; 500]; // Medium audio
        let mut stream = transcriber.transcribe(&audio_data);

        let mut text_chunks = Vec::new();
        while let Some(chunk) = stream.next().await {
            text_chunks.push(chunk);
        }

        assert_eq!(text_chunks.len(), 2);
        assert_eq!(text_chunks[0], "Hello");
        assert_eq!(text_chunks[1], " world");
    }

    #[tokio::test]
    async fn audio_transcriber_long_audio() {
        let transcriber = MockAudioTranscriber;
        let audio_data = vec![0x01; 2000]; // Long audio
        let mut stream = transcriber.transcribe(&audio_data);

        let mut text_chunks = Vec::new();
        while let Some(chunk) = stream.next().await {
            text_chunks.push(chunk);
        }

        assert_eq!(text_chunks.len(), 5);
        let full_text: String = text_chunks.join("");
        assert_eq!(full_text, "This is a longer transcription");
    }

    #[tokio::test]
    async fn audio_transcriber_empty_audio() {
        let transcriber = MockAudioTranscriber;
        let audio_data = vec![]; // Empty audio
        let mut stream = transcriber.transcribe(&audio_data);

        let mut text_chunks = Vec::new();
        while let Some(chunk) = stream.next().await {
            text_chunks.push(chunk);
        }

        assert!(text_chunks.is_empty());
    }

    #[test]
    fn data_type_alias() {
        let data: Data = vec![1, 2, 3, 4, 5];
        assert_eq!(data.len(), 5);
        assert_eq!(data[0], 1);
        assert_eq!(data[4], 5);
    }

    #[test]
    fn data_operations() {
        let mut data: Data = vec![0xFF; 1024];
        assert_eq!(data.len(), 1024);

        // Test push
        data.push(0x00);
        assert_eq!(data.len(), 1025);
        assert_eq!(data[1024], 0x00);

        // Test extend
        data.extend_from_slice(&[0x01, 0x02, 0x03]);
        assert_eq!(data.len(), 1028);
        assert_eq!(data[1025], 0x01);
        assert_eq!(data[1026], 0x02);
        assert_eq!(data[1027], 0x03);

        // Test clear
        data.clear();
        assert!(data.is_empty());
    }

    #[test]
    fn data_creation() {
        let empty_data: Data = Vec::new();
        assert!(empty_data.is_empty());

        let filled_data: Data = vec![42; 100];
        assert_eq!(filled_data.len(), 100);
        assert!(filled_data.iter().all(|&x| x == 42));
    }

    #[tokio::test]
    async fn audio_workflow() {
        let generator = MockAudioGenerator;
        let transcriber = MockAudioTranscriber;

        // Generate audio from text
        let prompt = "Hello world";
        let mut audio_stream = generator.generate(prompt);

        let mut all_audio_data = Vec::new();
        while let Some(chunk) = audio_stream.next().await {
            all_audio_data.extend_from_slice(&chunk);
        }

        // Transcribe the generated audio back to text
        let mut transcription_stream = transcriber.transcribe(&all_audio_data);

        let mut transcription_chunks = Vec::new();
        while let Some(chunk) = transcription_stream.next().await {
            transcription_chunks.push(chunk);
        }

        // Verify the workflow
        assert!(!all_audio_data.is_empty());
        assert!(!transcription_chunks.is_empty());

        let full_transcription: String = transcription_chunks.join("");
        assert_eq!(full_transcription, "This is a longer transcription");
    }
}