dnnclient.py for Python 3

dnnclient.py

#!/usr/bin/python3

# 	Julius Dictation Kit
# 	License Agreement

#  Copyright (c) 1991-2016 Kawahara Lab., Kyoto University
#  Copyright (c) 2005-2016 Lee Lab., Nagoya Institute of Technology

# "Julius Dictation Kit" has been developed at Kawahara Lab., Kyoto
# University and Lee Lab., Nagoya Institute of Technology (collectively
# referred to herein as the "Licensers").

# The Licensers reserve the copyright thereto.  However, as long as you
# accept and remain in strict compliance with the terms and conditions
# of the license set forth herein, you are hereby granted a royalty-free
# license to use this software program including the source code thereof
# and the documentation thereto (collectively referred to herein as the
# "Software").  Use by you of the Software shall constitute acceptance
# by you of all terms and conditions of the license set forth herein.

# 	      TERMS AND CONDITIONS OF LICENSE

# (Grant of License)
# 1. So long as you accept and strictly comply with the terms and
# conditions of the license set forth herein, the Licensers will not
# enforce the copyright or moral rights in respect of the Software, in
# connection with the use, copying, duplication, adaptation,
# modification, preparation of a derivative work, aggregation with
# another program, or insertion into another program of the Software and
# the distribution or transmission of the Software.

# (Copyright Notice)
# 2. In the event you provide to any third party all or any portion of
# the Software, you shall affix the above copyright notice and all terms
# and conditions of this license (both the Japanese original and English
# translation).  Moreover, in the event you revise all or any portion of
# the Software for distribution, a notice shall be affixed indicating
# that the Software has been revised, together with the revision date
# and the name of the person or entity that made the revision.

# (Publication)
# 3. When you publish or present any results by using the Software, you
# must explicitly mention your use of "Julius Dictation Kit".

# 3-1. The language model included in this Software has been developed
# by using the "Balanced Corpus of Contemporary Written Japanese
# (BCCWJ)" compiled and owned by the National Institute for Japanese
# Language and Linguistics.  When you publish or present any results by
# using the Software, you must explicitly mention this fact.

# 3-2. The acoustic model included in this Software has been developed
# by using the "ASJ Continuous Speech Corpus -- Japanese Newspaper
# Article Sentences (JNAS)" and the "Corpus of Spontaneous Japanese
# (CSJ)".

# (Disclaimer of Warranty)
# 4. The Licensers make no warranty of any kind and nature, whether
# express, implied or statutory on the Software and any matters relating
# thereto. THE LICENSERS SPECIFICALLY DISCLAIMS ALL IMPLIED WARRANTIES,
# DUTIES OR CONDITIONS, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
# WARRANTIES OF MERCHANTABILITY, OF FITNESS FOR A PARTICULAR PURPOSE, OF
# ACCURACY OR COMPLETENESS OF RESPONSES AND RESULTS, AND THAT THE USE OF
# THE SOFTWARE WILL NOT INFRINGE ANY PATENT, TRADEMARK, TRADE-SECRETS OR
# OTHER PROPRIETARY RIGHTS OF ANY THIRD PARTY.

# (No Liability)
# 5. The Licensers shall not be liable to you for any general or direct,
# special, incidental, consequential or other indirect damages
# whatsoever (including, but not limited to, damages for loss of profits
# or savings) arising out of or in any way connected with the use of or
# inability to use the Software, or otherwise under or in connection
# with any provisions of this License, even in the event of the fault,
# tort (including negligence), misrepresentation, strict liability
# (including product liability), breach of contract or breach of
# warranty of the Licensers, and even if the Licensers has been advised
# of possibility of such damages.

# (Choice of Law and Exclusive Forum)
# 6. This license of use of the Software shall be governed by the laws
# of Japan, and the Kyoto District Court shall have exclusive primary
# jurisdiction with respect to all disputes arising with respect
# thereto.

# (Contact)
# 7. Inquiries about support or maintenance of the Software, or
# inquiries concerning the license of use besides the conditions above,
# may be sent to Kawahara Lab., Kyoto University or Lee Lab., Nagoya
# Institute of Technology.


import sys
import numpy as np
import socket
import struct

adinserver_host = 'localhost'
adinserver_port = 5532
julius_host = 'localhost'
julius_port = 5531
num_raw = 120
num_input = 1320
num_hid = 2048
num_output = 2004
num_context = 11  # 1320 / 120
batchsize = 32

w_filename = ["dnn_sample/W_l1.npy", "dnn_sample/W_l2.npy",
              "dnn_sample/W_l3.npy", "dnn_sample/W_l4.npy",
              "dnn_sample/W_l5.npy", "dnn_sample/W_l6.npy",
              "dnn_sample/W_l7.npy", "dnn_sample/W_output.npy"]
b_filename = ["dnn_sample/bias_l1.npy", "dnn_sample/bias_l2.npy",
              "dnn_sample/bias_l3.npy", "dnn_sample/bias_l4.npy",
              "dnn_sample/bias_l5.npy", "dnn_sample/bias_l6.npy",
              "dnn_sample/bias_l7.npy", "dnn_sample/bias_output.npy"]

prior_filename = "dnn_sample/seedhmm.cluster.prior"

if len(sys.argv) > 1:
    conffile = sys.argv[1]
    f = open(conffile)
    for line in f:
        linebuf = line.strip().split(' ')
        if linebuf[0] == "--adinserver_host":
            adinserver_host = linebuf[1]
        elif linebuf[0] == "--adinserver_port":
            adinserver_port = int(linebuf[1])
        elif linebuf[0] == "--julius_host":
            julius_host = linebuf[1]
        elif linebuf[0] == "--julius_port":
            julius_port = int(linebuf[1])
        elif linebuf[0] == "--num_raw":
            num_raw = int(linebuf[1])
        elif linebuf[0] == "--num_input":
            num_input = int(linebuf[1])
        elif linebuf[0] == "--num_hid":
            num_hid = int(linebuf[1])
        elif linebuf[0] == "--num_output":
            num_output = int(linebuf[1])
        elif linebuf[0] == "--num_context":
            num_context = int(linebuf[1])
        elif linebuf[0] == "--batchsize":
            batchsize = int(linebuf[1])
        elif linebuf[0] == "--prior_filename":
            prior_filename = linebuf[1]
        elif linebuf[0] == "--w_filename":
            for i in range(1, len(linebuf)):
                w_filename[i - 1] = linebuf[i]
        elif linebuf[0] == "--b_filename":
            for i in range(1, len(linebuf)):
                b_filename[i - 1] = linebuf[i]
        elif linebuf[0] == "#":
            pass
        else:
            print("unkown switch")
            sys.exit()
    f.close()

w1 = np.load(w_filename[0])
w2 = np.load(w_filename[1])
w3 = np.load(w_filename[2])
w4 = np.load(w_filename[3])
w5 = np.load(w_filename[4])
w6 = np.load(w_filename[5])
w7 = np.load(w_filename[6])
wo = np.load(w_filename[7])

b1 = np.load(b_filename[0])
b2 = np.load(b_filename[1])
b3 = np.load(b_filename[2])
b4 = np.load(b_filename[3])
b5 = np.load(b_filename[4])
b6 = np.load(b_filename[5])
b7 = np.load(b_filename[6])
bo = np.load(b_filename[7])

state_prior = np.zeros((bo.shape[0], 1))
prior_factor = 1.0
for line in open(prior_filename):
    state_id, state_p = line[:-1].split(' ')
    state_id = int(state_id)
    state_p = float(state_p) * prior_factor
    state_prior[state_id][0] = state_p


def ff(x0):
    x1 = 1. / (1 + np.exp(-(np.dot(w1.T, x0) + b1)))
    x2 = 1. / (1 + np.exp(-(np.dot(w2.T, x1) + b2)))
    x3 = 1. / (1 + np.exp(-(np.dot(w3.T, x2) + b3)))
    x4 = 1. / (1 + np.exp(-(np.dot(w4.T, x3) + b4)))
    x5 = 1. / (1 + np.exp(-(np.dot(w5.T, x4) + b5)))
    x6 = 1. / (1 + np.exp(-(np.dot(w6.T, x5) + b6)))
    x7 = 1. / (1 + np.exp(-(np.dot(w7.T, x6) + b7)))
    tmp = np.dot(wo.T, x7) + bo
    np.exp(tmp, tmp)
    tmp /= np.sum(tmp, axis=0)
    tmp /= state_prior
    np.log10(tmp, tmp)
    return tmp


adinserversock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
adinserversock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
adinserversock.bind((adinserver_host, adinserver_port))
adinserversock.listen(1)

juliusclientsock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
juliusclientsock.connect((julius_host, julius_port))

sendconf = 0

print('Waiting for connections...')
adinclientsock, adinclient_address = adinserversock.accept()

splice_feature = np.zeros(num_input)
buf_splice_feature = None
fnum = 0

while True:

    rcvmsg = adinclientsock.recv(4)
    nbytes = struct.unpack('=i', rcvmsg)[0]

    if nbytes == 12:

        rcvmsg = adinclientsock.recv(12)
        fbank_vecdim, fbank_shift, fbank_outprob_p = struct.unpack(
            '=iii', rcvmsg)
        c_msg = struct.pack('=iiii', 12, num_output, 10, 1)
        juliusclientsock.sendall(c_msg)
        sendconf = 1

    elif nbytes == num_raw * 4:

        buffer = ''
        while len(buffer) < nbytes:
            tmpdata = adinclientsock.recv(nbytes - len(buffer))
            tmpdata = tmpdata.decode('ISO-8859-1')
            if not tmpdata:
                break
            buffer += tmpdata

        rcvmsg = buffer

        val = struct.unpack("=" + "f" * num_raw, rcvmsg.encode('ISO-8859-1'))
        splice_feature = np.r_[splice_feature[num_raw:num_input], val]
        if fnum >= num_context:
            if buf_splice_feature is not None:
                buf_splice_feature = np.hstack(
                    (buf_splice_feature, splice_feature[:, np.newaxis]))
            else:
                buf_splice_feature = splice_feature[:, np.newaxis]

        if buf_splice_feature is not None and buf_splice_feature.shape[1] == batchsize:
            xo = ff(buf_splice_feature)
            for i in range(xo.shape[1]):
                r_feature = xo[:, i]
                r_msg = struct.pack('=i', num_output * 4)
                juliusclientsock.sendall(r_msg)
                r_msg = struct.pack("=" + "f" * num_output, *r_feature)
                juliusclientsock.sendall(r_msg)
            buf_splice_feature = None
        fnum = fnum + 1

    elif nbytes == 0:

        if buf_splice_feature is not None:
            xo = ff(buf_splice_feature)
            for i in range(xo.shape[1]):
                r_feature = xo[:, i]
                r_msg = struct.pack('=i', num_output * 4)
                juliusclientsock.sendall(r_msg)
                r_msg = struct.pack("=" + "f" * num_output, *r_feature)
                juliusclientsock.sendall(r_msg)

        r_msg = struct.pack('=i', 0)
        juliusclientsock.sendall(r_msg)
        splice_feature = np.zeros(num_input)
        buf_splice_feature = None
        fnum = 0

r_msg = struct.pack('=i', 0)
juliusclientsock.sendall(r_msg)

r_msg = struct.pack('=i', -1)
juliusclientsock.sendall(r_msg)

adinclientsock.close()