Add application file

app.py

@@ -0,0 +1,117 @@
+import gradio as gr
+
+
+import os
+
+HF_TOKEN = os.getenv('HF_TOKEN')
+callback = gr.HuggingFaceDatasetSaver(HF_TOKEN, "paulbauriegel/voice-coe-demo")
+
+sentences = \
+    {'en':[
+        "In winter, the dry leaves fly around in the air.",
+        "The good old man broke through the ice with his horse and fell into the cold water.",
+        "He always eats the eggs without salt and pepper.",
+        "Good to know. Now I can fix the appointment.",
+        "We find this approach is particularly effective at learning speech to text translation.",
+        "Ever wonder what your Representative has been up to?",
+        "The Wenker sheets are the data basis for Georg Wenker's language atlases",
+        "At least she gets 7000 dollars in damages"
+    ],
+    'de':[
+        "Im Winter fliegen die trocknen Blätter durch die Luft herum.",
+        "Der gute alte Mann ist mit dem Pferde durch´s Eis gebrochen und in das kalte Wasser gefallen.",
+        "Er isst die Eier immer ohne Salz und Pfeffer.",
+        "Gut zu wissen. Jetzt kann ich den Termin vereinbaren.",
+        "Wir haben festgestellt, dass dieser Ansatz besonders effektiv beim Erlernen der Sprache-zu-Text Übersetzung 
+ist.",
+        "Haben Sie sich jemals gefragt, was Ihr Abgeordneter so treibt?",
+        "Die Wenkerbogen stellen die Datengrundlage für Georg Wenkers Sprachatlanten dar",
+        "Zumindest bekommt sie 7000 Dollar Schmerzensgeld",
+    ]}
+with gr.Blocks(title='Voice CoE Data Collection') as demo:
+    _ = gr.HTML('<center><h1>CoE Voice Data Collection</h1><center>')
+    lang = gr.Dropdown(
+            ["de", "en"], 
+            value='en',
+            interactive=True,
+            label="Choose your language", 
+    )
+    client_ip = gr.Label("", label="User-IP", visible=False)
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_0 = gr.Label(sentences['en'][0], label="")
+        audio_0 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_1 = gr.Label(sentences['en'][1], label="")
+        audio_1 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_2 = gr.Label(sentences['en'][2], label="")
+        audio_2 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_3 = gr.Label(sentences['en'][3], label="")
+        audio_3 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_4 = gr.Label(sentences['en'][4], label="")
+        audio_4 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_5 = gr.Label(sentences['en'][5], label="")
+        audio_5 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_6 = gr.Label(sentences['en'][6], label="")
+        audio_6 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        #outputs = gr.components.Textbox(label=)
+        label_7 = gr.Label(sentences['en'][7], label="")
+        audio_7 = gr.Audio(source="microphone", type="filepath", label="Record sample")
+    with gr.Row():
+        acc = gr.Dropdown(
+            ["yes", "no", "maybe"], 
+            label="Do you have an accent in the spoken language", 
+        )
+    with gr.Row():
+        agree = gr.Checkbox(value=False, label='I agree that my data is stored and analysed by the iHub CoE Voice Team')
+    with gr.Row():
+        btn = gr.Button("Submit data")
+        thx = gr.HTML('') # 
+
+    
+    lang.change(lambda x: {label_0: sentences[x][0], 
+                           label_1: sentences[x][1],
+                           label_2: sentences[x][2], 
+                           label_3: sentences[x][3], 
+                           label_4: sentences[x][4], 
+                           label_5: sentences[x][5], 
+                           label_6: sentences[x][6], 
+                           label_7: sentences[x][7], }, 
+                lang, 
+                [label_0, label_1, label_2, label_3,
+                 label_4, label_5, label_6, label_7])
+
+    # This needs to be called at some point prior to the first call to callback.flag()
+    callback.setup([client_ip, lang, audio_0, audio_1, audio_2, audio_3, audio_4, audio_5, audio_6, audio_7, acc], 
+"flagged_data_points")
+
+    # We can choose which components to flag -- in this case, we'll flag all of them
+    def submit_data(client_ip, lang, audio_0, audio_1, audio_2, audio_3, audio_4, audio_5, audio_6, audio_7, acc, agree, 
+request: gr.Request):
+        if not agree:
+            return '<h3>No data has been submitted</h3>'
+        else:
+            client_ip_d = {'ip': request.client.host}
+            callback.flag([client_ip_d, lang, audio_0, audio_1, audio_2, audio_3, audio_4, audio_5, audio_6, audio_7, 
+acc])
+            return '<h3>Thank you for submitting you data</h3>'
+
+    btn.click(submit_data, 
+              [client_ip, lang, audio_0, audio_1, audio_2, audio_3, audio_4, audio_5, audio_6, audio_7, acc, agree], 
+              thx, 
+              preprocess=False)
+
+demo.launch()
+