#!/bin/bash

# Copyright 2018 Jarvan Wang Apache 2.0.

# Copyright 2016 Xiaohui Zhang Apache 2.0.

 
 

# This srcipt operates on lattice directories, such as exp/tri4a_lats

# the output is a new lat dir which has lattices from all the input lattice dirs

 
 

# Begin configuration section.

cmd=run.pl

extra_files=

num_jobs=4

# End configuration section.

echo "$0 $@" # Print the command line for logging

 
 

if [ -f path.sh ]; then . ./path.sh; fi

. parse_options.sh || exit 1;

 
 

if [[ $# -lt 3 ]]; then

echo "Usage: $0 [options] <data> <dest-lat-dir> <src-lat-dir1> <src-lat-dir2> ..."

echo "e.g.: $0 --num-jobs 32 data/train exp/tri3_lat_combined exp/tri3_lat_1 exp_tri3_lat_2"

echo "Options:"

echo " --extra-files <file1 file2...> # specify addtional files in 'src-lat-dir1' to copy"

echo " --num-jobs <nj> # number of jobs used to split the data directory."

echo " Note, files that don't appear in the first source dir will not be added even if they appear in later ones."

echo " Other than lattices, only files from the first src lat dir are copied."

exit 1;

fi

 
 

data=$1;

shift;

dest=$1;

shift;

first_src=$1;

 
 

mkdir -p $dest;

rm $dest/{lat.*.gz,num_jobs} 2>/dev/null

 
 

cp $first_src/phones.txt $dest 2>/dev/null

 
 

export LC_ALL=C

 
 

for dir in $*; do

if [ ! -f $dir/lat.1.gz ]; then

echo "$0: check if lattices (lat.*.gz) are present in $dir."

exit 1;

fi

done

 
 

for dir in $*; do

for f in tree; do

diff $first_src/$f $dir/$f 1>/dev/null 2>&1

if [ $? -ne 0 ]; then

echo "$0: Cannot combine lattice directories with different $f files."

fi

done

done

 
 

for f in final.mdl tree cmvn_opts num_jobs $extra_files; do

if [ ! -f $first_src/$f ]; then

echo "combine_lat_dir.sh: no such file $first_src/$f"

exit 1;

fi

cp $first_src/$f $dest/

done

 
 

src_id=0

temp_dir=$dest/temp

[ -d $temp_dir ] && rm -r $temp_dir;

mkdir -p $temp_dir

echo "$0: dumping lattices in each source directory as single archive and index."

for dir in $*; do

src_id=$((src_id + 1))

cur_num_jobs=$(cat $dir/num_jobs) || exit 1;

lats=$(for n in $(seq $cur_num_jobs); do echo -n "$dir/lat.$n.gz "; done)

$cmd $dir/log/copy_lattices.log \

lattice-copy "ark:gunzip -c $lats|" \

ark,scp:$temp_dir/lat.$src_id.ark,$temp_dir/lat.$src_id.scp || exit 1;

done

sort -m $temp_dir/lat.*.scp > $temp_dir/lat.scp || exit 1;

 
 

echo "$0: splitting data to get reference utt2spk for individual lat.JOB.gz files."

utils/split_data.sh $data $num_jobs || exit 1;

 
 

echo "$0: splitting the lattices to appropriate chunks according to the reference utt2spk files."

utils/filter_scps.pl JOB=1:$num_jobs \

$data/split$num_jobs/JOB/utt2spk $temp_dir/lat.scp $temp_dir/lat.JOB.scp

 
 

for i in `seq 1 $num_jobs`; do

lattice-copy scp:$temp_dir/lat.${i}.scp "ark:|gzip -c >$dest/lat.$i.gz" || exit 1;

done

 
 

echo $num_jobs > $dest/num_jobs || exit 1

 
 

echo "$0: checking the lattice files generated have at least 90% of the utterances."

for i in `seq 1 $num_jobs`; do

num_lines=`cat $temp_dir/lat.$i.scp | wc -l` || exit 1;

num_lines_tot=`cat $data/split$num_jobs/$i/utt2spk | wc -l` || exit 1;

python -c "import sys;

percent = 100.0 * float($num_lines) / $num_lines_tot

if percent < 90 :

print ('$dest/lat.$i.gz {0}% utterances missing.'.format(percent))" || exit 1;

done

rm -r $temp_dir 2>/dev/null

 
 

echo "Combined lattices and stored in $dest"

exit 0

combine_lat_dirs.sh的更多相关文章

  1. 数据库备份并分离日志表(按月)sh 脚本

    #!/bin/sh year=`date +%Y` month=`date +%m` day=`date +%d` hour=`date +%H` dir="/data/dbbackup/f ...

  2. Linux下如何查找.sh后缀的文件

    find / -name *.sh或locate *.shfind 与locate的区别:locate 配合数据库查看文件位置 find 实际搜寻硬盘查询文件名称

  3. sh 测试网段在线主机

    yum install nmap nmap -sP  192.168.21.1/24 查看网段在线主机 grep -vwf file1 file2 文件内容比较 #!/bin/bash # day=` ...

  4. Maven打包生成可运行bat/sh脚本文件

    利用Maven的appassembler-maven-plugin插件,就可以实现自动打包可运行的脚本,还可以跨平台.   <plugin>      <groupId>org ...

  5. 记一个简单的保护if 的sh脚本

    真是坑爹,就下面的sh,竟然也写了很久! if [ `pwd` != '/usr/xx/bin/tomcat' ] then echo "rstall is not allowed in c ...

  6. 主机巡检脚本:OSWatcher.sh

    主机巡检脚本:OSWatcher.sh 2016-09-26更新,目前该脚本只支持Linux操作系统,后续有需求可以继续完善. 注意: 经测试,普通用户执行脚本可以顺利执行前9项检查: 第10项,普通 ...

  7. Oracle巡检脚本:ORAWatcher.sh

    Oracle巡检脚本:ORAWatcher.sh #!/usr/bin/ksh echo "" echo "ORAWatcher Version:1.0.1" ...

  8. 解决:/bin/sh: 1: /home/**/custom_app.sh: Permission denied错误

    出现如下错误,一般是执行权限不够. /bin/sh: : /home/custom_app.sh: Permission denied 解决方法是:cd 到此文件目录,对提示的文件赋予可执行权限或读写 ...

  9. SH Script Grammar

    http://linux.about.com/library/cmd/blcmdl1_sh.htm http://pubs.opengroup.org/onlinepubs/9699919799/ut ...

随机推荐

  1. jquery字符串操作

    目的:把自己常用到的jquery操作字符串总结一下 w3cSchool关于js字符串的整理:有需要的时候可以查 http://www.w3school.com.cn/js/jsref_obj_stri ...

  2. Rsync + sersync 实时同步备份

    一      Rsync + Sersync  实时同步介绍 1.Rsync 服务搭建介绍 云机上搭建Rsync server,在本地搭建Rsync Clinet. 2. Sersync 服务搭建介绍 ...

  3. Java的基本类型和包装类型

    测试的JDK版本:java version "1.7.0_79".Java的9中基本类型:void.char.boolean.byte.short.int.long.float.d ...

  4. 二、Redis安装

    一.下载Redis: 下载地址:https://github.com/MSOpenTech/redis/releases. 由于redis并不支持window系统,而window版本的redis的是由 ...

  5. SOme USeful NOtes for MYself.

    SOme USeful NOtes for MYself. B站神奇的频道(YouTube里同名):关于微积分/线代/梯度下降/DL等数学知识的理解,对理解DL很有帮助 https://space.b ...

  6. Bootstrap 模态框(Modal)插件id冲突

    <!DOCTYPE html><html><head>    <meta charset="utf-8">     <titl ...

  7. Nuget 多平台多目标快速自动打包

    构建现代的 .Net 应用离不开 Nuget 的支持,而快速打包 Nuget 成了提高生产率的有效方法. 1. 前置条件 为了实现 Nuget 的快速打包,我们需要先解决一些前置依赖,无论是 .Net ...

  8. vim常用快捷键整理

    搜索快捷键 /  关键字n 向下匹配N 向上匹配 移动光标快捷键 gg 命令将光标移动到文档开头,等同于 1GG 命令将光标移动到文档末尾0 或功能键[Home] 这是数字『 0 』:移动到这一行的最 ...

  9. Java技术栈思维导图

    Java技术栈思维导图 Java IO流体系 设计模式

  10. 牛客网 272B Xor Path(树上操作)

    题目链接:Xor Path 题意:每个顶点的点权为Ai,任意两点路径上点权异或和为Path(i,j),求所有Path(i,j)和. 题解:考虑每个顶点被用到的次数,分以下三种情况: 1.本身和其他顶点 ...