File size: 11,544 Bytes
43ffb0e
 
 
 
 
 
 
 
 
 
 
 
 
30eacd3
 
 
 
 
 
43ffb0e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30eacd3
 
 
43ffb0e
 
 
 
 
30eacd3
 
 
 
 
 
 
 
43ffb0e
 
 
 
 
 
 
 
30eacd3
 
 
43ffb0e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30eacd3
 
 
43ffb0e
 
 
 
 
30eacd3
 
 
 
 
 
 
 
43ffb0e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30eacd3
43ffb0e
 
 
 
 
30eacd3
 
43ffb0e
 
 
ea4087a
 
 
43ffb0e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30eacd3
 
 
43ffb0e
 
 
 
 
30eacd3
 
 
 
 
 
 
 
ea4087a
 
43ffb0e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
286
287
288
289
290
291
292
293
294
295
import base64
import tempfile
import os
import requests
import gradio as gr
from openai import OpenAI

# Available voices for audio generation
VOICES = ["alloy", "ash", "ballad", "coral", "echo", "fable", "onyx", "nova", "sage", "shimmer", "verse"]

def process_text_input(api_key, text_prompt, selected_voice):
    """Generate audio response from text input"""
    try:
        if not api_key:
            return "Please enter your OpenAI API key first.", None
            
        if not text_prompt:
            return "Please enter a text prompt first.", None
        
        # Initialize OpenAI client with the provided API key
        client = OpenAI(api_key=api_key)
        
        completion = client.chat.completions.create(
            model="gpt-4o-audio-preview",
            modalities=["text", "audio"],
            audio={"voice": selected_voice, "format": "wav"},
            messages=[
                {
                    "role": "user",
                    "content": text_prompt
                }
            ]
        )
        
        # Save the audio to a temporary file
        if not hasattr(completion.choices[0].message, 'audio') or not completion.choices[0].message.audio:
            return "No audio response was generated.", None
            
        wav_bytes = base64.b64decode(completion.choices[0].message.audio.data)
        temp_path = tempfile.mktemp(suffix=".wav")
        with open(temp_path, "wb") as f:
            f.write(wav_bytes)
        
        # Get the text response - with proper error handling
        text_response = "No text response available."
        if hasattr(completion.choices[0].message, 'content') and completion.choices[0].message.content:
            if isinstance(completion.choices[0].message.content, list) and len(completion.choices[0].message.content) > 0:
                if hasattr(completion.choices[0].message.content[0], 'text'):
                    text_response = completion.choices[0].message.content[0].text
            elif isinstance(completion.choices[0].message.content, str):
                text_response = completion.choices[0].message.content
        
        return text_response, temp_path
    except Exception as e:
        return f"Error: {str(e)}", None

def process_audio_input(api_key, audio_path, text_prompt, selected_voice):
    """Process audio input and generate a response"""
    try:
        if not api_key:
            return "Please enter your OpenAI API key first.", None
            
        if not audio_path:
            return "Please upload or record audio first.", None
        
        # Initialize OpenAI client with the provided API key
        client = OpenAI(api_key=api_key)
        
        # Read audio file and encode to base64
        with open(audio_path, "rb") as audio_file:
            audio_data = audio_file.read()
        encoded_audio = base64.b64encode(audio_data).decode('utf-8')
        
        # Create message content with both text and audio
        message_content = []
        
        if text_prompt:
            message_content.append({
                "type": "text",
                "text": text_prompt
            })
        
        message_content.append({
            "type": "input_audio",
            "input_audio": {
                "data": encoded_audio,
                "format": "wav"
            }
        })
        
        # Call OpenAI API
        completion = client.chat.completions.create(
            model="gpt-4o-audio-preview",
            modalities=["text", "audio"],
            audio={"voice": selected_voice, "format": "wav"},
            messages=[
                {
                    "role": "user",
                    "content": message_content
                }
            ]
        )
        
        # Save the audio response
        if not hasattr(completion.choices[0].message, 'audio') or not completion.choices[0].message.audio:
            return "No audio response was generated.", None
            
        wav_bytes = base64.b64decode(completion.choices[0].message.audio.data)
        temp_path = tempfile.mktemp(suffix=".wav")
        with open(temp_path, "wb") as f:
            f.write(wav_bytes)
        
        # Get the text response - with proper error handling
        text_response = "No text response available."
        if hasattr(completion.choices[0].message, 'content') and completion.choices[0].message.content:
            if isinstance(completion.choices[0].message.content, list) and len(completion.choices[0].message.content) > 0:
                if hasattr(completion.choices[0].message.content[0], 'text'):
                    text_response = completion.choices[0].message.content[0].text
            elif isinstance(completion.choices[0].message.content, str):
                text_response = completion.choices[0].message.content
        
        return text_response, temp_path
    except Exception as e:
        return f"Error: {str(e)}", None

def download_example_audio():
    """Download an example audio file for testing"""
    try:
        url = "https://cdn.openai.com/API/docs/audio/alloy.wav"
        response = requests.get(url)
        response.raise_for_status()
        
        # Save to a temporary file
        temp_path = tempfile.mktemp(suffix=".wav")
        with open(temp_path, "wb") as f:
            f.write(response.content)
        
        return temp_path
    except Exception as e:
        print(f"Error downloading example audio: {str(e)}")
        return None

def use_example_audio():
    """Load example audio for the interface"""
    audio_path = download_example_audio()
    if not audio_path:
        return None
    return audio_path

# Create Gradio Interface
with gr.Blocks(title="VoxTalk") as app:
    gr.Markdown("# VoxTalk")
    gr.Markdown("*Voice-to-voice AI interaction platform*")
    
    # API Key input (used across all tabs)
    api_key = gr.Textbox(
        label="OpenAI API Key", 
        placeholder="Enter your OpenAI API key here",
        type="password"
    )
    
    with gr.Tab("Text to Audio"):
        with gr.Row():
            with gr.Column():
                text_input = gr.Textbox(
                    label="Text Prompt", 
                    placeholder="Enter your question or prompt here...",
                    lines=3
                )
                text_voice = gr.Dropdown(
                    choices=VOICES,
                    value="alloy",
                    label="Voice"
                )
                text_submit = gr.Button("Generate Response")
            
            with gr.Column():
                text_output = gr.Textbox(label="AI Response (Text)", lines=5)
                audio_output = gr.Audio(label="AI Response (Audio)")
        
        text_submit.click(
            fn=process_text_input,
            inputs=[api_key, text_input, text_voice],
            outputs=[text_output, audio_output]
        )
    
    with gr.Tab("Audio Input to Audio Response"):
        with gr.Row():
            with gr.Column():
                audio_input = gr.Audio(
                    label="Audio Input", 
                    type="filepath",
                    sources=["microphone", "upload"]
                )
                example_btn = gr.Button("Use Example Audio")
                
                accompanying_text = gr.Textbox(
                    label="Accompanying Text (Optional)", 
                    placeholder="Add any text context or question about the audio...",
                    lines=2
                )
                audio_voice = gr.Dropdown(
                    choices=VOICES,
                    value="alloy",
                    label="Response Voice"
                )
                audio_submit = gr.Button("Process Audio & Generate Response")
            
            with gr.Column():
                audio_text_output = gr.Textbox(label="AI Response (Text)", lines=5)
                audio_audio_output = gr.Audio(label="AI Response (Audio)")
        
        audio_submit.click(
            fn=process_audio_input,
            inputs=[api_key, audio_input, accompanying_text, audio_voice],
            outputs=[audio_text_output, audio_audio_output]
        )
        
        example_btn.click(
            fn=use_example_audio,
            inputs=[],
            outputs=[audio_input]
        )
    
    with gr.Tab("Voice Samples"):
        gr.Markdown("## Listen to samples of each voice")
        
        def generate_voice_sample(api_key, voice_type):
            try:
                if not api_key:
                    return "Please enter your OpenAI API key first.", None
                
                client = OpenAI(api_key=api_key)
                completion = client.chat.completions.create(
                    model="gpt-4o-audio-preview",
                    modalities=["text", "audio"],
                    audio={"voice": voice_type, "format": "wav"},
                    messages=[
                        {
                            "role": "user",
                            "content": f"This is a sample of the {voice_type} voice. It has its own unique tone and character."
                        }
                    ]
                )
                
                # Save the audio to a temporary file
                if not hasattr(completion.choices[0].message, 'audio') or not completion.choices[0].message.audio:
                    return f"No audio response was generated for {voice_type} voice.", None
                    
                wav_bytes = base64.b64decode(completion.choices[0].message.audio.data)
                temp_path = tempfile.mktemp(suffix=".wav")
                with open(temp_path, "wb") as f:
                    f.write(wav_bytes)
                
                # Get the text response - with proper error handling
                text_response = "Sample generated"
                if hasattr(completion.choices[0].message, 'content') and completion.choices[0].message.content:
                    if isinstance(completion.choices[0].message.content, list) and len(completion.choices[0].message.content) > 0:
                        if hasattr(completion.choices[0].message.content[0], 'text'):
                            text_response = completion.choices[0].message.content[0].text
                    elif isinstance(completion.choices[0].message.content, str):
                        text_response = completion.choices[0].message.content
                
                return f"Sample: {text_response}", temp_path
            except Exception as e:
                return f"Error: {str(e)}", None
        
        with gr.Row():
            sample_voice = gr.Dropdown(
                choices=VOICES,
                value="alloy",
                label="Select Voice Sample"
            )
            sample_btn = gr.Button("Generate Sample")
        
        with gr.Row():
            sample_text = gr.Textbox(label="Status")
            sample_audio = gr.Audio(label="Voice Sample")
        
        sample_btn.click(
            fn=generate_voice_sample,
            inputs=[api_key, sample_voice],
            outputs=[sample_text, sample_audio]
        )
    
    gr.Markdown("""
    ## Notes:
    - You must provide your OpenAI API key in the field above
    - The model used is `gpt-4o-audio-preview`
    - Audio inputs should be in WAV format
    - Available voices: alloy, ash, ballad, coral, echo, fable, onyx, nova, sage, shimmer, and verse
    """)

if __name__ == "__main__":
    app.launch()