Update app.py

app.py

@@ -5,24 +5,11 @@ from transformers import pipeline
 from gtts import gTTS
 from pydub import AudioSegment
 from pydub.silence import detect_nonsilent
-from simple_salesforce import Salesforce  # Import Salesforce connection
-import re
 from waitress import serve
+from simple_salesforce import Salesforce
 
 app = Flask(__name__)
 
-# Salesforce connection using provided credentials
-sf_username = '[email protected]'
-sf_password = 'Sati@1020'
-sf_token = 'sSSjyhInIsUohKpG8sHzty2q'
-
-# Establish Salesforce connection
-try:
-    sf = Salesforce(username=sf_username, password=sf_password, security_token=sf_token)
-    print("Connected to Salesforce successfully!")
-except Exception as e:
-    print(f"Failed to connect to Salesforce: {str(e)}")
-
 # Use whisper-small for faster processing and better speed
 device = "cuda" if torch.cuda.is_available() else "cpu"
 asr_model = pipeline("automatic-speech-recognition", model="openai/whisper-small", device=0 if device == "cuda" else -1)
@@ -62,47 +49,49 @@ def convert_to_wav(input_path, output_path):
         audio = AudioSegment.from_file(input_path)
         audio = audio.set_frame_rate(16000).set_channels(1)  # Convert to 16kHz, mono
         audio.export(output_path, format="wav")
+        print(f"Converted audio to {output_path}")
     except Exception as e:
+        print(f"Error: {str(e)}")
         raise Exception(f"Audio conversion failed: {str(e)}")
 
 # Function to check if audio contains actual speech
 def is_silent_audio(audio_path):
     audio = AudioSegment.from_wav(audio_path)
     nonsilent_parts = detect_nonsilent(audio, min_silence_len=500, silence_thresh=audio.dBFS-16)  # Reduced silence duration
+    print(f"Detected nonsilent parts: {nonsilent_parts}")
     return len(nonsilent_parts) == 0  # If no speech detected
 
-# Extract name, email, and phone number from transcribed text
-def extract_name_email_phone(text):
-    # Regex for basic email and phone number
-    email = re.search(r'\S+@\S+', text)
-    phone = re.search(r'\+?\d{10,15}', text)  # Consider different formats for phone numbers
-
-    name = text.split(' ')[0]  # Simplified assumption that name is the first word
-    email = email.group(0) if email else "[email protected]"
-    phone = phone.group(0) if phone else "0000000000"
-
-    return name, email, phone
+# Salesforce connection details
+try:
+    print("Attempting to connect to Salesforce...")
+    sf = Salesforce(username='[email protected]', password='Sati@1020', security_token='sSSjyhInIsUohKpG8sHzty2q')
+    print("Connected to Salesforce successfully!")
+    print("User Info:", sf.UserInfo)  # Log the user info to verify the connection
+except Exception as e:
+    print(f"Failed to connect to Salesforce: {str(e)}")
 
 # Function to create Salesforce record
 def create_salesforce_record(name, email, phone_number):
     try:
-        # Create the record in Salesforce
+        # Attempt to create a record in Salesforce
         customer_login = sf.Customer_Login__c.create({
             'Name': name,
             'Email__c': email,
             'Phone_Number__c': phone_number
         })
-        
-        # Log the response from Salesforce
+        print(f"Salesforce response: {customer_login}")
+
         if customer_login.get('id'):
             print(f"Record created successfully with ID: {customer_login['id']}")
             return customer_login
         else:
-            print("Record creation failed: No ID returned")
+            print("No ID returned, record creation may have failed.")
             return {"error": "Record creation failed: No ID returned"}
     except Exception as e:
-        print(f"Error creating Salesforce record: {str(e)}")
-        return {"error": f"Failed to create record in Salesforce: {str(e)}"}
+        # Catch and log any exceptions during record creation
+        error_message = str(e)
+        print(f"Error creating Salesforce record: {error_message}")
+        return {"error": f"Failed to create record in Salesforce: {error_message}"}
 
 @app.route("/")
 def index():
@@ -111,6 +100,7 @@ def index():
 @app.route("/transcribe", methods=["POST"])
 def transcribe():
     if "audio" not in request.files:
+        print("No audio file provided")
         return jsonify({"error": "No audio file provided"}), 400
 
     audio_file = request.files["audio"]
@@ -125,13 +115,20 @@ def transcribe():
         # Check for silence
         if is_silent_audio(output_audio_path):
             return jsonify({"error": "No speech detected. Please try again."}), 400
+        else:
+            print("Audio contains speech, proceeding with transcription.")
 
         # Use Whisper ASR model for transcription
         result = asr_model(output_audio_path, generate_kwargs={"language": "en"})
         transcribed_text = result["text"].strip().capitalize()
+        print(f"Transcribed text: {transcribed_text}")
 
         # Extract name, email, and phone number from the transcribed text
-        name, email, phone_number = extract_name_email_phone(transcribed_text)
+        parts = transcribed_text.split()
+        name = parts[0] if len(parts) > 0 else "Unknown Name"
+        email = parts[1] if '@' in parts[1] else "[email protected]"
+        phone_number = parts[2] if len(parts) > 2 else "0000000000"
+        print(f"Parsed data - Name: {name}, Email: {email}, Phone Number: {phone_number}")
 
         # Create record in Salesforce
         salesforce_response = create_salesforce_record(name, email, phone_number)
@@ -140,7 +137,7 @@ def transcribe():
         if "error" in salesforce_response:
             print(f"Error creating record in Salesforce: {salesforce_response['error']}")
             return jsonify(salesforce_response), 500
-        
+
         print(f"Salesforce Response: {salesforce_response}")
         return jsonify({"text": transcribed_text, "salesforce_record": salesforce_response})