init commit

app.py

@@ -0,0 +1,60 @@
+import gradio as gr
+
+from utils import rag_api, agent_api, base64_to_audio
+
+
+# Helper functions
+def assignment_part_1(input_text, language):
+    response = rag_api(input_text, language)
+    if response["audio"]:
+        output_audio = base64_to_audio(response["audio"])
+        return response, output_audio
+    else:
+        return response, None, None
+
+
+def assignment_part_2(input_text, language):
+    response = agent_api(input_text, language)
+    if response["audio"]:
+        output_audio = base64_to_audio(response["audio"])
+        return response, output_audio
+    else:
+        return response, None, None
+
+
+# Create the Gradio interface
+with gr.Blocks() as block:
+
+    gr.Markdown("# <p style='text-align:center;'>Sarvam Intern Assignment</p>")
+
+    # User Query
+    input_text = gr.Textbox(label="Enter your query:")
+
+    # Language selection
+    language = gr.Dropdown(label="Select audio_language_code",
+                           choices=["hi-IN", "bn-IN", "kn-IN", "ml-IN",
+                                    "mr-IN", "od-IN", "pa-IN", "ta-IN",
+                                    "te-IN", "gu-IN", "en-IN"],)
+
+    # Two buttons
+    with gr.Row():
+        button_1 = gr.Button("RAG")
+        button_2 = gr.Button("Agent + RAG")
+
+    # Output field
+    output_text = gr.JSON(label="Output")
+    output_audio = gr.Audio(type="numpy",
+                            label="Output Audio",
+                            interactive=False)
+
+    # Button click actions
+    button_1.click(fn=assignment_part_1,
+                   inputs=[input_text, language],
+                   outputs=[output_text, output_audio])
+
+    button_2.click(fn=assignment_part_2,
+                   inputs=[input_text, language],
+                   outputs=[output_text, output_audio])
+
+# Launch the Gradio app
+block.launch()

utils.py

@@ -0,0 +1,55 @@
+import os
+import base64
+
+import io
+import requests
+import numpy as np
+from scipy.io import wavfile
+from dotenv import load_dotenv
+
+
+# Load environment variables
+load_dotenv()
+
+
+URL = os.getenv("URL")
+
+
+# RAG API
+def rag_api(query, audio_language_code):
+
+    res = requests.post(f"{URL}/api/rag",
+                        json={
+                            'query': query,
+                            'audio_language_code': audio_language_code
+                        })
+
+    return res.json()
+
+
+# Agent + RAG API
+def agent_api(query, audio_language_code):
+
+    res = requests.post(f"{URL}/api/agent",
+                        json={
+                            'query': query,
+                            'audio_language_code': audio_language_code
+                        })
+
+    return res.json()
+
+
+def base64_to_audio(base64_string):
+
+    try:
+        # Decode the base64 string
+        audio_data = base64.b64decode(base64_string)  # Decode
+        audio_file = io.BytesIO(audio_data)  # Convert to BytesIO object
+
+        sample_rate, samples = wavfile.read(audio_file)
+        samples_array = np.array(samples)
+
+        return sample_rate, samples_array
+    except Exception as e:
+        print(e)
+        return None, None