mirror of
https://github.com/FAUSheppy/ths-speech
synced 2025-12-06 07:21:39 +01:00
refactor
This commit is contained in:
Yannik Schmidt
@@ -1,45 +0,0 @@
|
||||
#!/usr/bin/python3
|
||||
|
||||
import socketserver
|
||||
import server_interface
|
||||
from configparse_wrapper.cpwrap import CFG, parse_config, change_cfg
|
||||
|
||||
class AppRequestHandler(socketserver.BaseRequestHandler):
|
||||
def __init__(self, request, client_address, server):
|
||||
socketserver.BaseRequestHandler.__init__(self, request, client_address, server)
|
||||
return
|
||||
|
||||
def setup(self):
|
||||
return socketserver.BaseRequestHandler.setup(self)
|
||||
|
||||
def handle(self):
|
||||
data = b""
|
||||
print("Recieved request")
|
||||
while True:
|
||||
tmp = self.request.recv(1024)
|
||||
if not tmp or b"\nterminate\n" in tmp:
|
||||
if len(tmp) > len("\nterminate\n"):
|
||||
data += tmp[:-len("\nterminate\n")]
|
||||
break
|
||||
data += tmp
|
||||
retval = server_interface.parse_request(data)
|
||||
try:
|
||||
self.request.send(retval)
|
||||
self.request.send(b"terminate\n")
|
||||
print("Replied with: {}".format(retval))
|
||||
except BrokenPipeError:
|
||||
print("Connection was closed before reply could be sent.")
|
||||
return
|
||||
|
||||
def finish(self):
|
||||
return socketserver.BaseRequestHandler.finish(self)
|
||||
|
||||
def start_server():
|
||||
socketserver.TCPServer.allow_reuse_address = True
|
||||
server = socketserver.TCPServer((CFG("interface"),CFG("port")),AppRequestHandler)
|
||||
server.serve_forever()
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parse_config()
|
||||
start_server()
|
||||
@@ -3,89 +3,42 @@ import os.path
|
||||
import audiosegment_wrapper as AudioSegment
|
||||
from configparse_wrapper.cpwrap import CFG
|
||||
|
||||
audiofiles = []
|
||||
FTP_DIR = CFG("shareDir")
|
||||
|
||||
def save_audio(filename, base64_string):
|
||||
global audiofiles
|
||||
audiofiles += [filename[:-4]]
|
||||
decoded = None
|
||||
orig_filename = filename[:-4]+"_orig"
|
||||
try:
|
||||
decoded = base64.b64decode(base64_string)
|
||||
except TypeError:
|
||||
return b"ERROR_INVALID_ENCODING_64"
|
||||
def saveAudio(filename, data):
|
||||
'''Save a single Audiofile to disk'''
|
||||
|
||||
if os.path.isfile(filename):
|
||||
return b"ERROR_FILE_EXISTS"
|
||||
write_file(orig_filename, decoded)
|
||||
raise ValueError("File: {} already exists.".format(filename))
|
||||
|
||||
seg = AudioSegment.from_file(orig_filename)
|
||||
seg = seg.resample(sample_rate_Hz=32000, sample_width=2, channels=1)
|
||||
with open(filename,"wb") as f:
|
||||
f.write(data)
|
||||
|
||||
seg.export(filename,format="wav")
|
||||
seg.export(FTP_DIR + filename,format="wav")
|
||||
if False:
|
||||
seg = AudioSegment.from_file(orig_filename)
|
||||
seg = seg.resample(sample_rate_Hz=32000, sample_width=2, channels=1)
|
||||
seg.export("{}.wav".format(filename), format="wav")
|
||||
|
||||
return b"SUCCESS"
|
||||
def saveAudioChain(filename, dataArray):
|
||||
'''Combine multiple received audio files into one and save it'''
|
||||
|
||||
def save_audio_chain(file_str_tupels):
|
||||
completeAudio = None
|
||||
for fname, base64_string in file_str_tupels:
|
||||
print("Filename: {}".format(fname))
|
||||
decoded = None
|
||||
orig_filename = fname[:-4]+"_orig"
|
||||
try:
|
||||
decoded = base64.b64decode(base64_string)
|
||||
except TypeError:
|
||||
return b"ERROR_INVALID_ENCODING_64"
|
||||
with open(orig_filename,"wb") as f:
|
||||
f.write(decoded)
|
||||
if completeAudio == None:
|
||||
completeAudio = AudioSegment.from_file(orig_filename)
|
||||
for index, data in enumerate(dataArray):
|
||||
tmpFile = "{}.{}.tmp".format(filename, index)
|
||||
with open(filename, "wb") as f:
|
||||
f.write(data)
|
||||
if not completeAudio:
|
||||
completeAudio = AudioSegment.from_file(tmpFile)
|
||||
else:
|
||||
completeAudio += AudioSegment.from_file(orig_filename)
|
||||
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")
|
||||
completeAudio.export(FTP_DIR + file_str_tupels[0][0],format="wav")
|
||||
return b"SUCCESS"
|
||||
completeAudio += AudioSegment.from_file(tmpFile)
|
||||
|
||||
def save_transcript(filename, transcript):
|
||||
global audiofiles
|
||||
try:
|
||||
audiofiles.remove(filename[:4])
|
||||
except:
|
||||
print("Audiofile not in working list")
|
||||
if os.path.isfile(filename):
|
||||
pass
|
||||
with open(filename + "_transcript","w") as f:
|
||||
f.write(transcript)
|
||||
with open(FTP_DIR + filename + "_transcript","w") as f:
|
||||
completeAudio = completeAudio.resample(sample_rate_Hz=32000, sample_width=2, channels=1)
|
||||
completeAudio.export("{}.wav".format(filename), format="wav")
|
||||
|
||||
def saveTranscript(filename, transcript):
|
||||
with open("{}_transcript.txt".format(filename), "w") as f:
|
||||
f.write(transcript)
|
||||
|
||||
def get_transcript(filename):
|
||||
global audiofiles
|
||||
try:
|
||||
with open("data/" + filename + ".wav_transcript","r") as f:
|
||||
with open("{}_transcript.txt".format(filename), "r") as f:
|
||||
return f.read()
|
||||
except FileNotFoundError:
|
||||
if filename in audiofiles:
|
||||
return "ERROR_FILE_STILL_IN_WORK"
|
||||
return "ERROR_FILE_NOT_AVAILIABLE"
|
||||
|
||||
def filelist():
|
||||
return ""
|
||||
|
||||
def fileinfo(filename):
|
||||
return ""
|
||||
|
||||
def copy_to_output(filename):
|
||||
return ""
|
||||
|
||||
def write_file(filename, data):
|
||||
global FTP_DIR
|
||||
with open(FTP_DIR + filename,"wb") as f:
|
||||
f.write(data)
|
||||
with open(filename,"wb") as f:
|
||||
f.write(data)
|
||||
|
||||
46
python-server/server.py
Executable file
46
python-server/server.py
Executable file
@@ -0,0 +1,46 @@
|
||||
#!/usr/bin/python3
|
||||
|
||||
import flask
|
||||
|
||||
@app.route('/submit')
|
||||
def transferToServer():
|
||||
# TODO receive data
|
||||
data = None
|
||||
decoded = decoded = base64.b64decode(data)
|
||||
filesystem.saveAudio(filename, data)
|
||||
return "", 204
|
||||
|
||||
@app.route('/submit-chain')
|
||||
def transferToServerChain():
|
||||
# TODO receive data
|
||||
data = []
|
||||
dataDecoded = [ base64.b64decode(d) for d in data ]
|
||||
filesystem.saveAudioChain(filename, dataDecoded)
|
||||
return "", 204
|
||||
|
||||
@app.route('/transcript')
|
||||
def queryTranscript():
|
||||
filename = flask.request.args.filename
|
||||
transcript = filesystem.getTranscript(filename)
|
||||
return transcript
|
||||
|
||||
@app.route('/fileinfo')
|
||||
def singleFileInfo():
|
||||
filename = flask.request.args.filename
|
||||
return filesystem.fileinfo(filename)
|
||||
|
||||
@app.route('/filelist')
|
||||
def getMultiFileInfo():
|
||||
return filesystem.filelist()
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = argparse.ArgumentParser(description='Start speech app server', \
|
||||
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
|
||||
parser.add_argument('--interface', default="localhost", \
|
||||
help='Interface on which flask (this server) will take requests on')
|
||||
parser.add_argument('--port', default="5002", \
|
||||
help='Port on which flask (this server) will take requests on')
|
||||
args = parser.parse_args()
|
||||
|
||||
app.run(host=args.interface, port=args.port)
|
||||
@@ -1,111 +0,0 @@
|
||||
import speech
|
||||
import filesystem
|
||||
from configparse_wrapper.cpwrap import CFG
|
||||
|
||||
MAIN_DIR = bytes(CFG("mainDataDir"),CFG("encoding"))
|
||||
|
||||
|
||||
def parse_request(data):
|
||||
''' parse request and call correct function '''
|
||||
#return b"DUMMY"
|
||||
print(data.split(b",")[0])
|
||||
# echo/test connection #
|
||||
cleared_data = is_data_type(b"ECHOREQUEST,",data)
|
||||
if cleared_data:
|
||||
return cleared_data
|
||||
|
||||
# reply transcript #
|
||||
cleared_data = is_data_type(b"GET_TRANSCRIPT,",data)
|
||||
if cleared_data:
|
||||
filename = cleared_data.decode("utf-8")
|
||||
return filesystem.get_transcript(filename).encode("utf-8")
|
||||
|
||||
# get single file info #
|
||||
cleared_data = is_data_type(b"GET_FILEINFO,",data)
|
||||
if cleared_data:
|
||||
filename = data.decode("utf-8")
|
||||
return filesystem.fileinfo(filename).encode("utf-8")
|
||||
|
||||
# get single file info #
|
||||
cleared_data = is_data_type(b"GET_FILEINFO_ALL,",data)
|
||||
if cleared_data:
|
||||
return filesystem.filelist().encode("utf-8")
|
||||
|
||||
# handle audio transmission #
|
||||
cleared_data = is_data_type(b"AUDIO_TRANSMISSION,",data)
|
||||
if cleared_data:
|
||||
print("Handling audio transmission")
|
||||
filename = None
|
||||
try:
|
||||
filename, base64_string = cleared_data.split(b',')
|
||||
filename = MAIN_DIR + filename.split(b"/")[-1] + b".wav"
|
||||
filename = filename.decode("utf-8")
|
||||
except ValueError:
|
||||
return b"ERROR_MALFORMED_REQUEST"
|
||||
ret = filesystem.save_audio(filename, base64_string)
|
||||
if b"ERROR" in ret:
|
||||
return ret
|
||||
speech.async_create_transcript(filename)
|
||||
return ret
|
||||
|
||||
# handle a chain of audiotransmissions #
|
||||
if data.startswith(b"CHAIN_AUDIO_TRANSMISSION"):
|
||||
file_str_tuples = []
|
||||
arr = data.split(b"|")
|
||||
for el in arr[1:-1]:
|
||||
filename, base64_string = el.split(b',')
|
||||
filename = MAIN_DIR + filename.split(b"/")[-1] + b".wav"
|
||||
filename = filename.decode("utf-8")
|
||||
file_str_tuples += [(filename,base64_string)]
|
||||
|
||||
if len(file_str_tuples) < 2: # a chain has 2 or more elements
|
||||
return bytes("ERROR_INVALID_NUMBER_FILES_{}".format(len(file_str_tuples)),"utf-8")
|
||||
|
||||
ret = filesystem.save_audio_chain(file_str_tuples);
|
||||
if b"ERROR" in ret:
|
||||
return ret
|
||||
speech.async_create_transcript(file_str_tuples[0][0])
|
||||
return ret
|
||||
|
||||
|
||||
# other shit
|
||||
return b"UNRECOGNIZED_SERVER_OPTION\n"
|
||||
|
||||
|
||||
|
||||
def is_data_type(tag,data):
|
||||
data = data.strip(b"\n")
|
||||
if data.startswith(tag):
|
||||
ret = data.split(tag)[1]
|
||||
if not ret:
|
||||
ret = b"NULL\n"
|
||||
return ret
|
||||
return None
|
||||
|
||||
|
||||
def reply_logs(loglevel=0,lines=100):
|
||||
''' replies with recent logs '''
|
||||
pass
|
||||
|
||||
def reply_backend(backend=None):
|
||||
''' replies with current backend if None or sets the new one '''
|
||||
pass
|
||||
|
||||
def reply_files(after=None, before=None, regex=None):
|
||||
''' replies with a list of transcribed or to be transcribed files '''
|
||||
|
||||
def reply_transcript(filename=None):
|
||||
''' replies with the latest transcript or the transcript of the chosen name '''
|
||||
pass
|
||||
|
||||
def reply_audio(filename=None):
|
||||
''' replies with the chosen audio as a file '''
|
||||
pass
|
||||
|
||||
def recive_transcribe_request(audiofile):
|
||||
''' saves and transcribes and audiofile '''
|
||||
pass
|
||||
|
||||
def android_unittest_transcribe_request(audiofile):
|
||||
''' the android unittests append a special keyword, requests are dummy handled '''
|
||||
pass
|
||||
@@ -1,43 +1,23 @@
|
||||
import speech_recognition as spr
|
||||
import multiprocessing as mp
|
||||
import os.path
|
||||
import filesystem
|
||||
import log
|
||||
import transcribe_async
|
||||
import os
|
||||
import transcribe
|
||||
|
||||
USE_FREE=False
|
||||
USE_PAID=True
|
||||
|
||||
def async_create_transcript(filename):
|
||||
print("Creating transcript..")
|
||||
mp.Process(target=create_and_save_transcript,args=(filename,)).start()
|
||||
def asyncCreateTranscript(filename):
|
||||
mp.Process(target=createAndSaveTranscript, args=(filename,)).start()
|
||||
|
||||
def create_and_save_transcript(filename):
|
||||
transcript = analyse(filename)
|
||||
filesystem.save_transcript(filename, transcript)
|
||||
#os.system("../permissions.sh")
|
||||
|
||||
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)
|
||||
string = free_google_backend(recognizer, audio)
|
||||
elif USE_PAID:
|
||||
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(filename):
|
||||
return transcribe_async.transcribe_file(filename)
|
||||
def createAndSaveTranscript(filename):
|
||||
if no hasTranscript(filename):
|
||||
try:
|
||||
string = transcribe.transcribeFile(filename)
|
||||
except spr.UnknownValueError:
|
||||
log.log("Audio file is broken or not an audio file.")
|
||||
raise e
|
||||
except spr.RequestError as e:
|
||||
log.log("Could not connect to google API: {}".format(e))
|
||||
raise e
|
||||
filesystem.saveTranscript(filename, string)
|
||||
|
||||
Reference in New Issue
Block a user