Fix : tts issues

sources/interaction.py

@@ -60,8 +60,8 @@ class Interaction:
         return buffer
     
     def transcription_job(self):
-        self.recorder = AudioRecorder()
-        self.transcriber = AudioTranscriber(self.ai_name, verbose=False)
+        self.recorder = AudioRecorder(verbose=True)
+        self.transcriber = AudioTranscriber(self.ai_name, verbose=True)
         self.transcriber.start()
         self.recorder.start()
         self.recorder.join()
@@ -71,7 +71,7 @@ class Interaction:
 
     def get_user(self):
         if self.stt_enabled:
-            query = self.transcription_job()
+            query = "TTS transcription of user: " + self.transcription_job()
         else:
             query = self.read_stdin()
         if query is None:

sources/memory.py

@@ -20,8 +20,7 @@ class Memory():
                  recover_last_session: bool = False,
                  memory_compression: bool = True):
         self.memory = []
-        self.memory = [{'role': 'user', 'content': system_prompt},
-                        {'role': 'assistant', 'content': f'Hello, How can I help you today ?'}]
+        self.memory = [{'role': 'user', 'content': system_prompt}]
         
         self.session_time = datetime.datetime.now()
         self.session_id = str(uuid.uuid4())

sources/router.py

@@ -22,7 +22,14 @@ class AgentRouter:
             return "cpu"
 
     def classify_text(self, text, threshold=0.5):
-        result = self.pipeline(text, self.labels, threshold=threshold)
+        first_sentence = None
+        for line in text.split("\n"):
+            if line.strip() != "":
+                first_sentence = line.strip()
+                break
+        if first_sentence is None:
+            first_sentence = text
+        result = self.pipeline(first_sentence, self.labels, threshold=threshold)
         return result
     
     def select_agent(self, text: str) -> Agent:

sources/speech_to_text.py

@@ -12,7 +12,7 @@ audio_queue = queue.Queue()
 done = False
 
 class AudioRecorder:
-    def __init__(self, format=pyaudio.paInt16, channels=1, rate=4096, chunk=8192, record_seconds=7, verbose=False):
+    def __init__(self, format=pyaudio.paInt16, channels=1, rate=4096, chunk=8192, record_seconds=5, verbose=False):
         self.format = format
         self.channels = channels
         self.rate = rate
@@ -90,7 +90,7 @@ class Transcript:
     
     def remove_hallucinations(self, text: str):
         # TODO find a better way to do this
-        common_hallucinations = ['Okay.', 'Thank you.', 'Thank you for watching.', 'You\'re', 'Oh', 'you', 'Oh.', 'Uh', 'Oh,', 'Mh-hmm', 'Hmm.']
+        common_hallucinations = ['Okay.', 'Thank you.', 'Thank you for watching.', 'You\'re', 'Oh', 'you', 'Oh.', 'Uh', 'Oh,', 'Mh-hmm', 'Hmm.', 'going to.', 'not.']
         for hallucination in common_hallucinations:
             text = text.replace(hallucination, "")
         return text
@@ -140,15 +140,15 @@ class AudioTranscriber:
         while not done or not audio_queue.empty():
             try:
                 audio_data, sample_rate = audio_queue.get(timeout=1.0)
-                if self.verbose:
-                    print(Fore.BLUE + "AudioTranscriber: Processing audio chunk" + Fore.RESET)
                 
+                start_time = time.time()
                 text = self.transcriptor.transcript_job(audio_data, sample_rate)
+                end_time = time.time()
                 self.recorded += text
-                print(Fore.YELLOW + f"Transcribed: {text}" + Fore.RESET)
+                print(Fore.YELLOW + f"Transcribed: {text} in {end_time - start_time} seconds" + Fore.RESET)
                 for language, words in self.trigger_words.items():
                     if any(word in text.lower() for word in words):
-                        print(Fore.GREEN + f"Start listening..." + Fore.RESET)
+                        print(Fore.GREEN + f"Listening again..." + Fore.RESET)
                         self.recorded = text
                 for language, words in self.confirmation_words.items():
                     if any(word in text.lower() for word in words):

sources/text_to_speech.py

@@ -5,9 +5,10 @@ import subprocess
 import re
 import platform
 
-
-
 class Speech():
+    """
+    Speech is a class for generating speech from text.
+    """
     def __init__(self, language = "english") -> None:
         self.lang_map = {
             "english": 'a',
@@ -24,6 +25,9 @@ class Speech():
         self.speed = 1.2
 
     def speak(self, sentence, voice_number = 1):
+        """
+        Use AI model to generate speech from text after pre-processing the text.
+        """
         sentence = self.clean_sentence(sentence)
         self.voice = self.voice_map["english"][voice_number]
         generator = self.pipeline(
@@ -42,17 +46,39 @@ class Speech():
                 winsound.PlaySound(audio_file, winsound.SND_FILENAME)
 
     def replace_url(self, m):
+        """
+        Replace URL with empty string.
+        """
         domain = m.group(1)
         if re.match(r'\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$', domain):
             return ''
         return domain
 
     def extract_filename(self, m):
+        """
+        Extract filename from path.
+        """
         path = m.group()
         parts = re.split(r'/|\\', path)
         return parts[-1] if parts else path
+    
+    def shorten_paragraph(self, sentence):
+        """
+        Shorten paragraph like **explaination**: <long text> by keeping only the first sentence.
+        """
+        lines = sentence.split('\n')
+        lines_edited = []
+        for line in lines:
+            if line.startswith('**'):
+                lines_edited.append(line.split('.')[0])
+            else:
+                lines_edited.append(line)
+        return '\n'.join(lines_edited)
 
     def clean_sentence(self, sentence):
+        """
+        Clean sentence by removing URLs, filenames, and other non-alphanumeric characters.
+        """
         lines = sentence.split('\n')
         filtered_lines = [line for line in lines if re.match(r'^\s*[a-zA-Z]', line)]
         sentence = ' '.join(filtered_lines)