hlrc

#!/usr/bin/python

#PYTHONPATH="/opt/ros/groovy/lib/python2.7/dist-packages:/vol/csra/releases/nightly/lib/python2.7/:/vol/csra/releases/nightly/lib/python2.7/site-packages/

import sys

import wave

from textgrid_hlc import *

import logging

import errno

import os.path

try:

    import rsb

    import rsb.converter

    import rst

    import rstsandbox

    from rst.audition.Utterance_pb2 import Utterance

    from rst.audition.SoundChunk_pb2 import SoundChunk

except ImportError as exception:

    sys.stderr.write("ImportError: {}\n> HINT: try to export PYTHONPATH=$PYTHONPATH:$YOUR_PREFIX/lib/python2.7/site-packages/\n\n".format(exception))

    sys.exit(errno.ENOPKG)

class hlrc_utterance():

    def __init__(self, _base_scope):

        #print "> registering rst converter"

        converter = rsb.converter.ProtocolBufferConverter(messageClass = Utterance)

        rsb.converter.registerGlobalConverter(converter)

        self.set_scope(_base_scope)

    def set_scope(self, scope):

        self.base_scope = str(scope) #NOTE: str() is important here, scope is a qstring (?) and gets deleted during call

        print "> setting scope to '%s'" % self.base_scope

        try:

            self.server = rsb.createRemoteServer(self.base_scope + '/set')

        except ValueError:

            print "> invalid scope given. server deactivated"

            self.server.deactivate()

    def trigger_utterance(self, filename_praat, filename_wav, blocking):

        if (self.server is None):

            print("> invalid server")

            return

        if (not os.path.isfile(filename_praat)):

            print "can not open file '%s'" % (filename_praat)

            return 0

        if (not os.path.isfile(filename_wav)):

            print "can not open file '%s'" % (filename_wav)

            return 0

        print "> reading wave file '%s'" % (filename_wav)

        wav = wave.open(filename_wav, "r")

        print "> parsing praat file '%s'" % (filename_praat)

        tgrid = TextGrid.load(filename_praat)

        #create utterance & fill it with values:

        ut = Utterance()

        #textual description of audio file

        ut.text = filename_praat

        ut.audio.data = wav.readframes(-1)

        ut.audio.sample_count = wav.getnframes()

        ut.audio.channels = wav.getnchannels()

        ut.audio.rate = wav.getframerate()

        if (wav.getsampwidth() == 1):

            ut.audio.sample_type = SoundChunk.SAMPLE_U8

        elif (wav.getsampwidth() == 2):

            ut.audio.sample_type = SoundChunk.SAMPLE_S16

        else:

            print "> invalid sample type. py doc says wave files are either u8 or s16"

            exit(0)

        #wave spec says always little endian

        ut.audio.endianness = SoundChunk.ENDIAN_LITTLE

        print "> filling phones with data from praat"

        for tier in tgrid.tiers:

            idx = (tgrid.tiers.index(tier)) + 1

            transcript = tier.simple_transcript

            for (xmin, xmax, utt) in transcript:

                phoneme = ut.phonemes.add()

                phoneme.symbol = utt

                phoneme.duration = int(1000.0*(float(xmax)-float(xmin)))

        with rsb.createRemoteServer(self.base_scope + '/set') as server:

            if (blocking):

                #blocking:

                print "> calling the utterance rpc (blocking until we finished talking)..."

                print '> server reply: "%s"' % server.utterance(ut)

            else:

               print "> calling the utterance rpc (NON-BLOCKING)..."

               future = server.utterance.async(ut)

               #we can block here for a incoming result with a timeout in seconds

               #print '> server reply: "%s"' % future.get(timeout = 10);

        print "> done"

        return 1

def main():

    # Pacify logger.

    #logging.basicConfig()

    if (len(sys.argv) != 4):

        print "> usage: %s <base_scope> file.praat file.wav\n>     example: %s /flobi1 hello.praat hello.wav" % (sys.argv[0] , sys.argv[0])

        sys.exit(0)

    filename_praat = sys.argv[2]

    filename_wav = sys.argv[3]

    base = sys.argv[1]

    hlc = hlrc_utterance(base)

    hlc.trigger_utterance(filename_praat, filename_wav, 1)

if __name__ == '__main__':

    main()