working google recog

python-server/filesystem.py

@@ -1,6 +1,6 @@
 import base64
 import os.path
-from pydub import AudioSegment
+import audiosegment_wrapper as AudioSegment
 
 def save_audio(filename, base64_string):
     decoded = None
@@ -15,7 +15,9 @@ def save_audio(filename, base64_string):
     #    return b"ERROR_FILE_EXISTS"
     with open(orig_filename,"wb") as f:
         f.write(decoded)
-    AudioSegment.from_file(orig_filename).export(filename,format="wav")
+    seg = AudioSegment.from_file(orig_filename)
+    seg = seg.resample(sample_rate_Hz=32000, sample_width=2, channels=1)
+    seg.export(filename,format="wav")
     return b"SUCCESS"
 
 def save_audio_chain(file_str_tupels):
@@ -37,6 +39,7 @@ def save_audio_chain(file_str_tupels):
     if not completeAudio:
         return b"ERROR_AUDIO_CONCAT_FAILED"
     else:
+        completeAudio = completeAudio.resample(sample_rate_Hz=32000, sample_width=2, channels=1)
         completeAudio.export(file_str_tupels[0][0],format="wav")
         return b"SUCCESS"
 

python-server/speech.py

@@ -3,6 +3,7 @@ import multiprocessing as mp
 import os.path
 import filesystem
 import log
+import transcribe_async
 
 USE_FREE=False
 USE_PAID=True
@@ -17,26 +18,24 @@ def create_and_save_transcript(filename):
 
 def analyse(filename):
     ''' returns the transcripted audio, or None if the analysis fails '''
+    try:
+        if USE_FREE:
             recognizer = spr.Recognizer()
             with spr.AudioFile(filename) as source:
                  audio = recognizer.record(source)
-
-    try:
-        if USE_FREE:
             string = free_google_backend(recognizer, audio)
         elif USE_PAID:
-            string = paid_google_backend(recognizer,audio)
+            string = paid_google_backend(filename)
     except spr.UnknownValueError:
         log.log("Audio file is broken or not an audio file")
         return "ERROR_AUDIO_FILE_INVALID"
     except spr.RequestError as e:
         log.log("Could not connect to google API: {}".format(e))
         return "ERROR_API_FAILURE"
-
     return string
 
 def free_google_backend(recognizer, audio):
     return recognizer.recognize_google(audio,language="de-DE")
 
-def paid_google_backend(recognizer, audio):
-    pass
+def paid_google_backend(filename):
+    return transcribe_async.transcribe_file(filename)

python-server/transcribe_async.py

@@ -24,68 +24,50 @@ Example usage:
 
 import argparse
 import io
+from gcloud import storage
+from google.cloud import speech
+from google.cloud.speech import enums
+from google.cloud.speech import types
 
 
 # [START speech_transcribe_async]
 def transcribe_file(speech_file):
-    """Transcribe the given audio file asynchronously."""
-    from google.cloud import speech
-    from google.cloud.speech import enums
-    from google.cloud.speech import types
-    client = speech.SpeechClient()
+    url = upload_file(speech_file)
+    print(url)
+    return transcribe_gcs("gs://"+url)
 
-    # [START speech_python_migration_async_request]
-    with io.open(speech_file, 'rb') as audio_file:
-        content = audio_file.read()
+def upload_file(filename):
+		bukket = "ths-speech-audio/"
+		client = storage.Client()
+		cb = client.get_bucket("ths-speech-audio")
+		blob = cb.blob(filename)
+		blob.upload_from_filename(filename)
+		return bukket + filename
 
-    audio = types.RecognitionAudio(content=content)
-    config = types.RecognitionConfig(
-        encoding=enums.RecognitionConfig.AudioEncoding.LINEAR16,
-        sample_rate_hertz=16000,
-        language_code='en-US')
-
-    # [START speech_python_migration_async_response]
-    operation = client.long_running_recognize(config, audio)
-    # [END speech_python_migration_async_request]
-
-    print('Waiting for operation to complete...')
-    response = operation.result(timeout=90)
-
-    # Each result is for a consecutive portion of the audio. Iterate through
-    # them to get the transcripts for the entire audio file.
-    for result in response.results:
-        # The first alternative is the most likely one for this portion.
-        print(u'Transcript: {}'.format(result.alternatives[0].transcript))
-        print('Confidence: {}'.format(result.alternatives[0].confidence))
-    # [END speech_python_migration_async_response]
-# [END speech_transcribe_async]
-
-
-# [START speech_transcribe_async_gcs]
 def transcribe_gcs(gcs_uri):
     """Asynchronously transcribes the audio file specified by the gcs_uri."""
-    from google.cloud import speech
-    from google.cloud.speech import enums
-    from google.cloud.speech import types
     client = speech.SpeechClient()
 
     audio = types.RecognitionAudio(uri=gcs_uri)
     config = types.RecognitionConfig(
-        #encoding=enums.RecognitionConfig.AudioEncoding.FLAC,
-        #sample_rate_hertz=16000,
+        encoding=enums.RecognitionConfig.AudioEncoding.LINEAR16,
+        sample_rate_hertz=32000,
         language_code='de-DE')
 
     operation = client.long_running_recognize(config, audio)
 
     print('Waiting for operation to complete...')
-    response = operation.result(timeout=90)
+    response = operation.result(timeout=900)
 
     # Each result is for a consecutive portion of the audio. Iterate through
     # them to get the transcripts for the entire audio file.
+    ret = ""
     for result in response.results:
         # The first alternative is the most likely one for this portion.
+        ret += result.alternatives[0].transcript
         print(u'Transcript: {}'.format(result.alternatives[0].transcript))
         print('Confidence: {}'.format(result.alternatives[0].confidence))
+    return ret
 # [END speech_transcribe_async_gcs]