ding, logger

BitLit_main.py

@@ -23,7 +23,7 @@ import hashlib
 import tempfile
 from logger import logger
 
-from snowboydecoder import play_audio_file
+from snowboydecoder import play_ding, play_dong
 
 import speech_recognition as sr  ## Packages for voice recognizer
 for index, name in enumerate(sr.Microphone.list_microphone_names()):
@@ -42,11 +42,11 @@ except:
 
 def play_mp3(mp3_file):
     """Play mp3 file with pyglet."""
-    # FIXME (wassname) It currently plays in background but we want to wait untill it's finished
     source = pyglet.media.load(filename=mp3_file, streaming=False)
     source.play()
-    time.sleep(source.duration*2 + 2)  # TODO must be a better way to wait untill the media has played
-
+    print(mp3_file, source.duration)
+    time.sleep(source.duration + 4)  # must be a better way to wait untill the media has played
+    print(mp3_file, source.duration)
 
 def cache_gtts(text, lang="en-nz", cache_file=None):
     """
@@ -71,7 +71,7 @@ def cache_gtts(text, lang="en-nz", cache_file=None):
         en: English
 
     """
-    print('say:', text)
+    logger.debug('say: %s', text)
     if not cache_file:
         hash_filename = hashlib.md5(text.encode()).hexdigest() + '.mp3'
         cache_file = os.path.join(tempfile.gettempdir(), hash_filename)
@@ -85,49 +85,61 @@ def generate_poem():
 
 
     ############ AUDIO CONVERSION TO TEST
-    play_audio_file()
+    play_dong()
     t0 = time.time()
     r = sr.Recognizer()
     with sr.Microphone() as source:
-        outfile1 = cache_gtts(text="Hi! My Name is BIT-LIT. PLEASE SPEAK SOME IDEAS FOR A POEM between the beeps.")
+
+        # print(r.energy_threshold)
+        # r.adjust_for_ambient_noise(source) 
+        # print('energy_threshold', r.energy_threshold)
+        r.energy_threshold=50
+
+
+        print('mic', source)
+        outfile1 = cache_gtts(text="Hi! My Name is BIT-LIT. PLEASE SPEAK SOME IDEAS FOR A POEM AFTER THE BEEP.")
         play_mp3(outfile1)
+        play_ding()
 
-        play_audio_file()
+        print('speak now', time.time())
+        audio = r.listen(source)
+        logger.debug('done recording %s', time.time())
+        logger.info('recorded %s s', len(audio.frame_data)/audio.sample_rate)
 
-        audio = r.listen(source, phrase_time_limit=20)
-        play_audio_file()
-
-        outfile2 = cache_gtts(text="BEEP. THANK YOU! GIVE ME A MINUTE TO GENERATE AND READ YOUR POEM. BEEP")
+        play_dong()
+        outfile2 = cache_gtts(text="BEEP. THANK YOU! GIVE ME A MINUTE TO GENERATE AND READ YOUR POEM")
         play_mp3(outfile2)
 
     t1 = time.time()
-    print('listen took', t1 - t0)
+    logger.debug('listen took %s', t1 - t0)
 
     try:
-        print("using google speech to text...")
+        logger.debug("using google speech to text...")
         USER_INPUT = r.recognize_google_cloud(audio, credentials_json=GOOGLE_CLOUD_SPEECH_CREDENTIALS)
-        print("Google thinks you said: " + USER_INPUT)
+        logger.info("Google thinks you said: " + USER_INPUT)
     except sr.UnknownValueError as e:
-        print("Could not understand audio. {}".format(e))
+        logger.error("Could not understand audio. {}".format(e))
         return 
     except sr.RequestError as e:
-        print("Could not request results; {0}".format(e))
+        logger.error("Could not request results; {0}".format(e))
         return
 
     t1b = time.time()
-    print('transcribe took', t1b - t1)
+    logger.debug('transcribe took %s', t1b - t1)
+
+    return
 
     # Generate poem from user seed
     text_generated = poem(USER_INPUT)
     t2 = time.time()
     logger.info("ML POEM is: %s", text_generated)
-    logger.info('poem and rhyme generation took %s', t2 - t1)
+    logger.debug('poem and rhyme generation took %s', t2 - t1)
 
     # TEXT CONVERSION IN AUDIO
     # FEED POEM TO TRANSCRIBER
     tts = gTTS(text=text_generated)
-    # ts = datetime.datetime.utcnow().strftime('%Y%m%d_%H-%M-%S')
-    poem_mp3 = "outputs/BitLit_poem.mp3"#.format(ts)
+    ts = datetime.datetime.utcnow().strftime('%Y%m%d_%H-%M-%S')
+    poem_mp3 = "outputs/BitLit_{}.mp3".format(ts)
     tts.save(poem_mp3)
     play_mp3(poem_mp3)
 
@@ -136,8 +148,8 @@ def generate_poem():
 
     ######
     t3 = time.time()
-    logger.info('Poem to speech took %s', t3 - t2)
-    logger.info("Time spent is about: %s seconds")
+    logger.debug('Poem to speech took %s', t3 - t2)
+    logger.debug("Total time spent is about: %s seconds", np.round(t3 - t0))
 
 
 if __name__ == "__main__":