From 3500a8cdf03da6c2c940e187b566dbd399a6037e Mon Sep 17 00:00:00 2001
From: yemaozi88 <428968@gmail.com>
Date: Sun, 30 Dec 2018 21:27:20 +0100
Subject: [PATCH] load novo phoneset.

---
 .gitignore                      |   3 +++
 .vs/acoustic_model/v15/.suo     | Bin 79360 -> 69632 bytes
 acoustic_model/check_novoapi.py |  14 ++++++++------
 acoustic_model/defaultfiles.py  |  11 ++++++++---
 acoustic_model/htk_vs_kaldi.py  |  12 ++++++------
 5 files changed, 25 insertions(+), 15 deletions(-)

diff --git a/.gitignore b/.gitignore
index 3c4efe2..b1d3894 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,6 +1,9 @@
 ## Ignore Visual Studio temporary files, build results, and
 ## files generated by popular Visual Studio add-ons.
 
+## important ##
+.acoustic_model/forced_alignment_novo.py
+
 # User-specific files
 *.suo
 *.user
diff --git a/.vs/acoustic_model/v15/.suo b/.vs/acoustic_model/v15/.suo
index b0dbc237cecd5fcbc4b9412f889b0c8ae84ae80c..2dca09361c976aa5e05ca38403fd7de672a53037 100644
GIT binary patch
delta 3954
zcmc(idu&tZ6~OPgwu9q1iIW(|c{Dg7FA^O5O5(U7ae_yAw16RW=^7v*gjZ=2z@Z}~
z#8A3Pq(ZR|;4(Cg-c_jVQ9)dYy0X?hgt|`KI*F<OF|At=pl+0wPILmib8fCf#^{PP
zY5S6&&-tGB`_A{B^W7W#R2@8~9<ddQf|3k^kWH$JxF3(lxs>(7Q~o?L?c}L1P7fZt
z49`<ZY%%tW{ZNHpiZ6@ZBKE8LH32HWfnTZehMc4=CuoRxk|KFL2kA?RYNC{Qme3Ki
ziEYG8BA3V`xGyIc>E%Qb@gk8<Ow>&L<}3N>Br^$mj!v%8as~cSn;mK~rdlYIN*IZi
zA^(izMpC$$H}p)SO--2;rJz-jT)B+&N+nOGnS3s8NiK#i3?*l$ZlU60!c63Z{8_FS
zly-e2cM#uI^8Je3MRI`Ht>l9w@h<eH8Md2YH<{N%1+gWvc|B9iYwx7;RulbD;>mGJ
zCjH)oD*xIk&xU5BSI}&R*vg4|wsJyGSh<$Uo*<@L*{`&lMzV(BXL3J&5L-Dh#thQg
z$R)%qB7?9JX~bNjKoAxQb)?t`o|T(qqBCy<>s2J##B+!;;z^>OXdrlxng}nk0wKMu
zX)$R_h-ZlJ5N!3th&hVhLh{>F%AZs8r%8S<<ev%MA%A6i59NA^t;BYMM|nw+i4k`y
z`N4<Ehm`z2k^mEHQ;}3FdKyVI>nuZ=q!@?{r6P-@k(f@Hh-|`4<Pf<;9+6KxL0E_x
z1RJi9WDzlwC?-k>E-xclPRt@Ilrk<?5aahxbvy-OeE_W3rymk$1O8pVTT@C+J-9P<
zR|>Cw7atZuh{kjJ0Q#{nZ`k~QTNKai;YC#{3$v1}0>L57A;m`62?sG{#q!E)smx8(
zCF)LfWP_9`vK^{rJGdt7Fq;~jWIH6W9gb@8ANGaGu#)r@1Re;LW5BUgBZo2!uAH6n
zdnz6#C}BKjCKhED;svo9-bTN3G5VbK;k6gFeE$60g8K)1)zLd^wScZPL)4Jg3b-*@
zuT9Jt2hz)*+e~fwG;$>20M|`$13sNBIWY1G<v@}MDtuD8j!)=2!~x=6rHlg$pU{1H
zukn%wo+XthAHD6>0|re+(V~JYFzOM-sJNgB@Is~?0}GcX%O1cXlINWwl4*`ohEk-d
z44#a7iZ21i+LmQ>N?(k6u3Z}_I}{j<?YX;ApfICv>*(z1+aB1|x#q<l^bc0z`v=`l
zjnF)2MT^JoaC>Z?U3PoTI)|gqVeN9(uCu!Ao$IU}w%S^2*V>NGj!svnyR*|Zi0UJ2
zux)%e+!ll#k<UtCM}}`uP~G|X(z}Dgy_3DGUAg=`%v)80*{{u@riB<>6^t~OfhX#z
zpecGwy-}|@2BW3`XliH`<G3^X`(JPVw*S`r9M9SOybZi1ljf(qv|f~4N#T5-nq#EA
z3xWJsA#6N`^=Vps(^zljT{}y>NBolDAv84Y=v|{4qz3V!&Tj#}Vamsbs#lVMTE_WS
z<PD47uF^xO#gP1pG6s46k08{$K$<TeP68?8?XF+7eSWC%gLC&A)^{&?>KdCdaw!RZ
zfrI1Tu*Hj`79(E1d0c08Ioxgz;{y6`ROWD;P9(l4gx25jjZL-f<L*ds8kD2`gxPTA
zo$}_l=3hP!f64jqy#4yqIC85uTy^0Qla!W-VOv0S0`9Q=hP;PPpZ!&7xjI~b8`j{L
z^+T{4`{oW$oYdAUD#@!JT`(_kPM2wU)$*C9Rcqv9ieK3mYajC3map;GBDsI!m}}C>
z(v!|BEs|$eIXSavU{esLPHNN=gs|C*vBKu4*l-^tDe2K3{}Zi7>J_B{QQ9U-K{bZA
zUW{HmGc7fB2|bYhWqyE7>oK*L-h6t5&2q^xY_e>{v-fPU5`VqF8h^e&3rF@R<?sb6
z2=nN#MsU-0>ZEJcCeXEndxjgOVbzagBM=^0i=W&q&>f{KIg47W1oDnE(qG9XA~|Te
zngI*3@Pr$Of(EES`_+Y5{OLT?)a@;hpCSmpbGT}J7%Oh1<CaxMY)2y=x&CJhzno2^
z!WiG!X@s8O7er3@`anr!uO5~cR|y-0O+tX!K+&v+7T861dJo;VU6gO27crfHO)cg~
zffhDm?4~Z=>ZqIF<nlSbRb6YV89ay|hdngeS?N_9HU&0q=<eCpH88Nty$`|CSkmlt
zJLu0^<FGlbcDv7Rb-QbARvV@H?%L|wI-g^ZSCDsH1yfy3ExXk<qZy=TF|thuCNz~+
z;L?L$IDmW1jp4u9c+D^nE(;>OR-JMld=$7Sbh&&OEA#U4c#N?0`e84Jq;GJ#KOJVL
zlO*+H#?<!?)>S7d@zg0#>apUfZ_<>`e>LUf7}3dFkn*74qw+>+wTR~;({&G;;nox7
z8oSMAA3*!KiI1WXTytysnp>yeXPpwNj<nO^{3!-+1@Xh1CnH`F)L0=J)O;M-bF)4o
zh2P3KDG>*!Ev%aeIB%Nt)8kZeMy{T4CjNwxCz$_WCS09V-`%%mOT$FV6||yU?ERxz
z`k7kl)k$YXC?5kY!~vUkn4+eAr#1fr59JpwHj9|>&}iZH=b_}A{Og6T?zw9x-6na`
zWiWL3C6HZXEGB{%!lhm`+?d6_t8!ZFqQTN29M^#s=?pzq9FW62%}<U4i{wpOes6Om
zRtoueU_<WMu}o+OD3cAR!B6hGB9<&rkGU=I0YsWpp$31uK548z4KBj*QV=71OQCJ7
z%>Zwxju(T7qB(OcvjVQdv=Itu9-(b4UIJ83&%n>y3h1kYKU(%j_UNHH^4U!Ap{K$S
z**hDqJ?NUDbQSheQ~oY_IoeUq_g!nmkP81{oG~6p#Jx9tk(dgKA|F}d4s7ISt;l;+
u*qZn#*(XnC?0yx9fbGt83N9Qtl#72Tt;ArN1xv)df4>I(&h1k=RR0stdf&(Z

delta 10778
zcmeHNYj9h|71q6xP2xC?9S29T^RQ`5<k*V!v?P;)^|WlsdRp?MF3z>yvMedm)iZIV
zG!!UF4b}pNJQ$C7b^3z|5zs<Q#|6sJFifX2{0dBlVLBb?^heqW1=>P-R+jQ`5^#dU
zqhvjw&)KuP_wL=Z_uF%Km6>H`{?2@Wi%ut3u&No1at4F34Dk8cvu8mt2te(#rUtTV
z=aZxCiKA~5JZ#pW6qzDcqmRh_<RJp3mZi$}gSj7lx~w6+5wsNmxWw5?kW9cfKm%Y4
zpaO6!U@ZXGX#=?)@EPd!McXQnbpYHCHpp85YXG?ZW{}N*O#p7uwh|-@u${pOFd9Kw
z4#1ULK-K_S0abt>0)7Jc0RWe;1K9ysUCd(}PGZPQ08&6F;5NYR09<wlNEx6Dz(K@{
z+U9Q1a7h|u51<#I1?T{J04}pkn`&zYiWzLc$v$mrCqmOE0g;$Ctz*Zb7(a3nWD2kw
zum>;+z;>K6pl3;cRp>Dk@q<IGkiU#W()WS6|7Kh`e}CK-JX{!Ntd|nLE_TyR#d_FH
zhN50_lMwPHH{o{JO(hrO{6aUG7>sF4&H4>c4SNPxDJATf&(Klkx~88MOPVgIf!)v!
z*a0w3n;L3Fpy2Vw&AUKyV3>%0C~3p)##S>(H2{0KuV@p1v;hW-d9<@(Euzad5PQ*d
z**eiUnEikWKmZT~>;mA+g+WFDG$0DVk5+KYG}X|!4-B{xLChB{1IYxG74wKy(TJX_
z=qlTc4bM{hiEJfJRH0~PnoKd!zbf~Z`QQ?BD7k9)3JqxY;7N>l^|krx`axnjz8MAz
zm9fyh4I4x_gL`*jZ?P=|Hr(S<Zxf*7-o^lNe7&~cfQD=C1&OESU4X@RJGX|+?wDo1
z2#wGaJGz+oO8F0FSwyapoLRGi=*oAs%@N0|Xo8qR(bWgCok}KtM^VPwmZ!U(A@a9u
zIz=4!WeDOfBx}9{&5g?uSsA{<SgdqH+rn7DC1fAnmhas1I)TEc>QLPi+p=w4`OLc+
zqKOZcpdV-PL*mEzDd29vJ%D=w_rtVkL7?DDr8Ki3{|MwC1%P)B6S-~V@p{ivj(g1|
z<BmR)y^dZbdHFN)N+N5d$s_~CjBk|@n8WaRA=-2>JElahkz7QN^)vgieQJhX_GeHI
z%utnt7QMD^$4sb_fJ*Cj<b>tN8%Yw8wto^M=p$R2_z{W@9LU{H=0=zoJEHR-fxX2?
zr_NO6yVT&ZTNDKD45BIHaw}XP6K&yUB+S@%^Pk#pB69&Urzgws0P?3zn?-oI4*(7V
zeh!!dJOIEIra@)_4;J$efqWS7dA>>cG?6nib0IP}$ei!QiPMe9%-#+&M%KLRAs?}6
zrSGZM(;xq1>aH&TA=J>)yuk>asB3~{qMX6_4^Z{~%Ja@dvxm2!OiLygCbRNa&yOhU
zwdEGLqX~c?1Z7gp{3rDCWjQ{1xzWzkZkUt{#_grb+z6T5L+0)ya~UR*-(8olJNSIr
zD$K#qr2=&5^M`c!g-27<o6ve>MSQfWmJ3S|{=-YrykwwyBzk-fEz8|WqPdwd^#05y
zB<p9PsiW-2=N={y+1j-d&-Gepw48AQ?QCg8?I%shaHJl+V_~7`V{GCGTK!n3#NIg)
zO46QSK+wSv@JIaKv5|<Ec5*_LeVlSTBO`%eEJ%gCBV%B21{@<1XC&ec2GG>8*7VDS
zvXcvcy9G*xN+y)bI1;r|%n|ZMB94sD7jQ&kC{gl7Qjv-~xyKUpM`>u=*h4!fB8zOK
z=o~QwoXW5_<{TNY=!(7N;BKM!GV4+H*j6;!(sb_e(wQ3c+grEVPhS7NA?joUdS|k^
zD&`E+qaheX=V-)9<Do%SC&Pu&-ZNRv-m_a$NOmDd(BYDDghG*kBjr*Qhc9(eLI>X|
z5_3Bye}93STF|k>Y~rm8EcNJ^sH$Sr6%E+&7)A<rdHTsJ;<DZ)D;TJ5>*!KAxuWRt
z?V;pt*E(CiLfi5HH2!rtn$B!VZ|4@$9<NPqFvq#tZg)@yNpG=wR8}`=a-5Z!6tW>K
zU6OXG2P?8cy~WJ!k%c2ddxr@d%mI^?2elOmUpyumGFh>)*RJvm+v75sGSUZGu~KGt
z#CCP#1{$ZuYEH{-N{=yQ>q}5Py;*G(JI97~7QdVF)B2Fbrbt@M)`X6$mwANh-jHS>
zl&}xWta6pHyIrQ$C?htxUgYc!D+0kGtwI@8daVkrR_>#R<<?=kmml;E^@fGqsI#?I
zC>|HsQcAAJ6qX7N2B$+J5eeD{$A#h)->y*gkMs31_qb%jq%`$;JZhg-)ZRBHjj4ik
zk1`w$sv_!OOqH_5hR2N5fUe(el13AXkWCj7`di1jt$vkJE^3eW7$?MLn?f4BB4wuw
zlno?;Q$o=fn&4Xd9ompt>*Wr36{e&hZlxt^MJz71S@el9GZpF<#)8Ut)ZXn*`UVx^
zv6Mz4bQ)=%CK(yiw0C<3g8B%ZlJN{di`p|BO_+A6-I9naDpW^lnWME|qP3`oCd4su
zcQ1D+9ndNK?jao|v-_-4z0j7BDjk%>AWw-6zIZZj(MNg)hXnBCI&psj20X@h8U&sJ
zjdhqOP1q&0moKt9VuLA1NI#(A%9TDzuUBUX8wXu+ry(kxFuTnj!=P1V42yfba@UZP
zvdg6!O0GA@Od6pp-O=iI_o>YB?nJLq><+7fyQC41PijrI+PXs%JzimK&>nZB{C20&
zozlso!$z;AJ1*&v_PSfWab906F5(7t-S%kEBjzO~9-byH=k-P;R+m~Ah+FLXfZXF4
zvgrmQNnTG}D;n$*b!(`&G#cgVoEAm8&*2!B3Hbh$IOQ{#5`A`gQp>X#tpY>4*%zR^
zeAl>7pOUnTbYgLT($w1ylej~y)4K#F=K!A5GL>K5KW>SdCKQT!H=h5`Vh|*QF?-I?
zH1`Y88v)7_^X!RSE!8;(EZ60TdL&v}I_g=jVA$?-j8cBDJ22r4(4&}+;XO4W-Vq6|
zMs72fRJ8dyHkv)spu>@KDP=%u1Sqd@3CUhzNiO4D)#8HwkP+e)q`HcBMre1KnwX%%
z@Eigt5D<m`<hO%+My@CK<WgkrcNNDEzWEiyk@O-+?0Rt+W~cPmOiv5hDg8Cm(?WJi
zf6er?ke!&Ao^AxBHv-a2KR0d!q_7tMzXhb}>q+K|;PUiR#+nPsd=XrpUdmW=A(=0N
z%hO94Yc3@7MTDB_rHM5!`pPYCPh4A!nZ`S`^QmbL?<m9O&NauFR~1RlZ&RQ*Guyu#
z@#HuCEqjAHmhdz6J4hYrtEG+~7vF0vMvA)f+Bf-H7QF-c1-pE8Ywf|n#hZWd@rb{C
zVViZbdCirJYA-4Jc6?^upoJyey}EKlam6j1^tJ8bEZNJ^ewEwv*X89%U*_eQ-}&W!
zFLwPL*AR+bEuE@S6Ily~eU~{pLF*6I5SqgFT{=^JUPCV*Qmk5ej)tq})}mF1A4G7%
zAzb7M$7&BfJ{pURj#GYz7jcg4fX1s4_lUG0i~EE%=k5SX8Gy*w9R4LyxAC04dJdm!
z;-vht6OxS$lsy=Y&|drKM9|^%pFckVqP7&C5sp==E^l6Sk-4;w*@(}AONk8%?l6us
z@sC<?n&DLvo6(UMs?k1nZ|+e7?K;fGhc`=m7eq}q;H%zzQ779dCtpl;j1^A<!)WiR
z3RLmJQ{|yxWQ4M3A3aOZDD!*<?c2BU`+l5qBLw|^hoHP0oC$I2rKif@1ZU>&c_T@n
z<G!^(R=FZ?W_w>KG3cA@HXu|MQfjxLJ0D?H5B>@N!Vi{)zpOv;3hzT5GHO}qy-8Mn
z?qnB%uBvucldIWZ{^05CR`TQfr~bXZ#^3x51Aa$`Ue~;Yy0q+}(t$fH7<)?VoB~er
zPrb;NKcjy8l~NznGNF>MDxBnh;i3aO1=ff&9Jw`5{OxR6?tNzN6Y`ty1bXGD5H)_t
z%u7zcK_K%VeQ55?1~et8&(D6ca~YauyYj9x@H@<N%53wR{F?v%l_2gz``#0w)1N9)
Y=eFu!v2G$hAo1^r9!AUW|B@R21GL|`<p2Nx

diff --git a/acoustic_model/check_novoapi.py b/acoustic_model/check_novoapi.py
index 40defe6..9a3057a 100644
--- a/acoustic_model/check_novoapi.py
+++ b/acoustic_model/check_novoapi.py
@@ -28,11 +28,13 @@ stimmen_transcription_ = pd.ExcelFile(default.stimmen_transcription_xlsx)
 phonelist_novo70_      = pd.ExcelFile(default.phonelist_novo70_xlsx)
 df = pd.read_excel(phonelist_novo70_, 'list')
 
-translation_key = dict()
-for ipa, novo70 in zip(df['IPA_simple'], df['novo70_simple']):
-    if not pd.isnull(ipa):
-        print('{0}:{1}'.format(ipa, novo70))
-        translation_key[ipa] = novo70
 
-#df = pd.read_excel(stimmen_transcription, 'check')
+## novo phoneset
+#translation_key = dict()
+## *_simple includes columns which has only one phone in.
+#for ipa, novo70 in zip(df['IPA_simple'], df['novo70_simple']):
+#    if not pd.isnull(ipa):
+#        print('{0}:{1}'.format(ipa, novo70))
+#        translation_key[ipa] = novo70
+#phonelist_novo70 = np.unique(list(df['novo70_simple']))
 
diff --git a/acoustic_model/defaultfiles.py b/acoustic_model/defaultfiles.py
index 9f4d4fa..ae510fc 100644
--- a/acoustic_model/defaultfiles.py
+++ b/acoustic_model/defaultfiles.py
@@ -3,7 +3,7 @@ import os
 #default_hvite_config = os.path.join(os.path.dirname(os.path.realpath(__file__)), 'data', 'htk', 'config.HVite')
 
 cygwin_dir = r'C:\cygwin64\home\Aki\acoustic_model'
-kaldi_dir = r'C:\OneDrive\WSL\kaldi-trunk\egs\fame\s5'
+
 #config_hcopy = os.path.join(cygwin_dir, 'config', 'config.HCopy')
 #config_train = os.path.join(cygwin_dir, 'config', 'config.train')
 config_hvite = os.path.join(cygwin_dir, 'config', 'config.HVite')
@@ -30,11 +30,16 @@ repo_dir = r'C:\Users\Aki\source\repos'
 ipa_xsampa_converter_dir    = os.path.join(repo_dir, 'ipa-xsama-converter')
 forced_alignment_module_dir = os.path.join(repo_dir, 'forced_alignment')
 
-fame_dir        = r'C:\OneDrive\WSL\kaldi-trunk\egs\fame\s5\corpus'
+WSL_dir   = r'C:\OneDrive\WSL'
+fame_dir        = os.path.join(WSL_dir, 'kaldi-trunk', 'egs', 'fame')
+fame_s5_dir     = os.path.join(fame_dir, 's5')
+fame_corpus_dir = os.path.join(fame_dir, 'corpus')
 
 experiments_dir = r'c:\OneDrive\Research\rug\experiments'
 stimmen_transcription_xlsx = os.path.join(experiments_dir, 'stimmen', 'data', 'Frisian Variants Picture Task Stimmen.xlsx')
 stimmen_data_dir           = os.path.join(experiments_dir, 'stimmen', 'data')
 phonelist_friesian_txt     = os.path.join(experiments_dir, 'friesian', 'acoustic_model', 'config', 'phonelist_friesian.txt')
-phonelist_novo70_xlsx      = os.path.join(experiments_dir, 'Nederlandse phonesets_aki.xlsx')
+
+novo_api_dir = os.path.join(WSL_dir, 'python-novo-api')
+cmu69_phoneset = os.path.join(novo_api_dir, 'novoapi', 'asr', 'phoneset', 'en', 'cmu69.phoneset')
 
diff --git a/acoustic_model/htk_vs_kaldi.py b/acoustic_model/htk_vs_kaldi.py
index 3ec6a2c..f9985f1 100644
--- a/acoustic_model/htk_vs_kaldi.py
+++ b/acoustic_model/htk_vs_kaldi.py
@@ -28,12 +28,12 @@ htk_dict_dir       = os.path.join(default.experiments_dir, 'stimmen', 'dic_short
 fa_dir             = os.path.join(default.experiments_dir, 'stimmen', 'FA_44k')
 result_dir         = os.path.join(default.experiments_dir, 'stimmen', 'result')
 
-kaldi_data_dir = os.path.join(default.kaldi_dir, 'data', 'alignme') 
-kaldi_dict_dir = os.path.join(default.kaldi_dir, 'data', 'local', 'dict')
+kaldi_data_dir = os.path.join(default.fame_s5_dir, 'data', 'alignme') 
+kaldi_dict_dir = os.path.join(default.fame_s5_dir, 'data', 'local', 'dict')
 lexicon_txt    = os.path.join(kaldi_dict_dir, 'lexicon.txt')
 
-#lex_asr	 = os.path.join(default.fame_dir, 'lexicon', 'lex.asr')
-#lex_asr_htk = os.path.join(default.fame_dir, 'lexicon', 'lex.asr_htk')
+#lex_asr	 = os.path.join(default.fame_corpus_dir, 'lexicon', 'lex.asr')
+#lex_asr_htk = os.path.join(default.fame_corpus_dir, 'lexicon', 'lex.asr_htk')
 
 
 # procedure
@@ -274,8 +274,8 @@ if make_kaldi_lexicon_txt:
 
 ## ======================= load kaldi forced alignment result =======================
 if load_forced_alignment_kaldi:
-    phones_txt = os.path.join(default.kaldi_dir, 'data', 'lang', 'phones.txt')
-    merged_alignment_txt = os.path.join(default.kaldi_dir, 'exp', 'tri1_alignme', 'merged_alignment.txt')
+    phones_txt = os.path.join(default.fame_s5_dir, 'data', 'lang', 'phones.txt')
+    merged_alignment_txt = os.path.join(default.fame_s5_dir, 'exp', 'tri1_alignme', 'merged_alignment.txt')
     
     #filenames	   = np.load(stimmen_data_dir + '\\filenames.npy')
     #words		   = np.load(stimmen_data_dir + '\\words.npy')