gradint-build/samples/utils/splitter.py

-#!/usr/bin/env python
-
-# Program to support splitting a long sound file into
-# several little ones.
-
-# Needs 'sox' - if Windows, download from
-# sox.sourceforge.net
-# (e.g. http://prdownloads.sourceforge.net/sox/sox12172.zip
-# - note gives a "select a mirror" dialogue) and put sox.exe
-# in the same directory or on the path
-
-# -----------------------
-
-# lowpri: 2nd sort key by length ? (only matters if adding a lot of new words & phrases at same time)
-
-import time,os,sndhdr,sys
-try: import winsound
-except: winsound=None
-macsound = (sys.platform.find("mac")>=0 or sys.platform.find("darwin")>=0)
-if macsound: sys.stderr.write("Warning: You need to have qtplay (from gradint or wherever) in your PATH for this to work\n")
-
-def rawcut(allData,fromSecs,toSecs,rate=22050,bits=16,channels=1):
-    return allData[secbyte(fromSecs,rate,channels,bits):secbyte(toSecs,rate,channels,bits)]
-def secbyte(sec,rate,channels,bits):
-    # Convert a time in seconds to a byte offset in the raw
-    # data
-    # Note: Result MUST be a multiple of bytesPerSample
-    # 'sec' is not necessarily an integer
-    sampleNo = int(0.5+sec*rate) # nearest integer sample no
-    bytesPerSample = channels*int(bits/8)
-    return sampleNo * bytesPerSample
-
-def readTimings(langs):
-    if macsound: time.sleep(1) # OS X hack due to qtplay delay (1sec on an Intel 2GHz Core Duo running OSX 10.5)
-    sys.stdout.write("Starting clock\n")
-    # Now using time.time() rather than time.clock()
-    # due to clock units confusion
-    # Just have to hope the system is accurate enough
-    offset = time.time()
-    ret = [] ; ip=''
-    start = offset
-    while not ip=='q':
-        ip = raw_input(langs[len(ret)%len(langs)]+": ")
-        t = time.time()
-        if ip=="c" and ret: ret[-1]=(ret[-1][0],t-offset)
-        elif not ip: ret.append((start-offset,t-offset))
-        start = t
-    sys.stdout.write("Finishing at %f seconds\n" % (t-offset,))
-    return ret
-
-def instructions():
-    sys.stdout.write("Press Return between samples\n")
-    sys.stdout.write("Enter 'c' to change the time of the last Return to this one\n")
-    sys.stdout.write("Enter 'x' to omit this bit (e.g. silence)\n")
-    sys.stdout.write("Enter 'q' when done (AFTER stopping last sample)\n")
-    sys.stdout.write("PRESS RETURN TO START\n")
-    raw_input()
-
-def getParams():
-    wavFile=raw_input("Enter filename of main recording: ")
-    header = sndhdr.what(wavFile)
-    if not header:
-        sys.stdout.write("Problem opening that file\n")
-        return None
-    (wtype,rate,channels,wframes,bits) = header
-    sys.stdout.write("WAV file is %d-bit\n" % (bits,))
-    if bits==8: soxBits="-b -u" # unsigned
-    elif bits==16: soxBits="-w -s" # signed
-    elif bits==32: soxBits="-l -s" # signed
-    else:
-        sys.stdout.write("Unsupported bits per sample '%s'\n" % (bits,))
-        return None
-    soxParams = "-t raw %s -r %d -c %d" % (soxBits,rate,channels)
-    rawFile = wavFile + ".raw"
-    convertToRaw(soxParams,wavFile,rawFile)
-    lang1=lang2=None
-    while not lang1: lang1=raw_input("Enter first language on recording (e.g. zh): ")
-    interleaved=input("Are two languages interleaved? (1/0): ") # (horrible hack)
-    if interleaved:
-        while not lang2: lang2=raw_input("Enter second language on recording (e.g. en): ")
-    else:
-        lang2=lang1
-        sys.stdout.write("OK - should run this program again for other language's recording\n")
-    return soxParams,wavFile,rawFile,lang1,lang2,rate,bits,channels
-
-def convertToWav(soxParams,rawFile,wavFile):
-    os.system("sox %s \"%s\" \"%s\"" % (soxParams,rawFile,wavFile))
-def convertToRaw(soxParams,wavFile,rawFile):
-    os.system("sox \"%s\" %s \"%s\"" % (wavFile,soxParams,rawFile))
-
-def main():
-    tuple=None
-    while not tuple: tuple=getParams()
-    soxParams,wavFile,rawFile,lang1,lang2,rate,bits,channels = tuple
-    mainLoop(soxParams,wavFile,rawFile,lang1,lang2,rate,bits,channels)
-    os.unlink(rawFile)
-
-# Set lang1 & lang2 equal if not interleaving
-def mainLoop(soxParams,wavFile,rawFile,lang1="zh",lang2="en",rate=22050,bits=16,channels=1):
-    allData=open(rawFile,"rb").read()
-    open(wavFile,"rb").read() # to cache before starting clock and 'play' (especailly because just loaded the separate raw data) (could also play from raw data if got sox)
-    instructions()
-    # Start sound asynchronously - hope for the best that
-    # the first clock reading is near enough to the actual
-    # start of the sound
-    if winsound: winsound.PlaySound(wavFile,winsound.SND_FILENAME | winsound.SND_ASYNC)
-    elif macsound: os.spawnlp(os.P_NOWAIT,"qtplay","qtplay",wavFile)
-    # else: os.spawnlp(os.P_NOWAIT,"play","play",wavFile)
-    # Problem: What if 'play' o/p's at slightly less than the correct rate - will think the cuts are further on in the file than they really are.  (e.g. 16000Hz on a z61p Cygwin, "time play" shows it takes slightly longer than sox thinks the file is)
-    # Better convert to 44100 just to make sure.
-    else: os.system("sox \"%s\" -r 44100 -t wav - | play -t wav - &" % wavFile)
-    # Read timings, cut up, and write out the samples
-    samples = [ rawcut(allData,s,f,rate,bits,channels) for s,f in readTimings([lang1,lang2]) ]
-    formatString = "%0"+str(len(str(int(len(samples)/(2-(lang2==lang1))-1))))+"d_%s"
-    # (pad with 0s as necessary so it's in order)
-    # (len(samples)-1 gives highest number, so len(str(l..))
-    # gives number of digits in it)
-    for i in range(len(samples)):
-        if i%2: lang=lang2
-        else: lang=lang1
-        if lang1==lang2: c=i
-        else: c=int(i/2)
-        fname = formatString % (c,lang)
-        f=open(fname, "wb")
-        f.write(samples[i])
-        f.close()
-        convertToWav(soxParams,fname,fname+".wav")
-        os.unlink(fname)
-        sys.stdout.write("Written %s.wav\n" % (fname,))
-
-if __name__=="__main__":
-    main()

gradint-build/samples/utils/strip0.py

-#!/usr/bin/env python
-
-# Program to strip any silence from the beginning/end of a
-# sound file (must be real 0-bytes not background noise)
-
-# (This is useful as a "splitter" post-processor when
-# getting samples from CD-ROMs e.g. "Colloquial Chinese" -
-# don't use audacity here because some versions of audacity
-# distort 8-bit audio files)
-
-# Needs 'sox' + splitter
-
-from splitter import *
-
-for wavFile in sys.argv[1:]:
-    # Figure out sox parameters
-    header = sndhdr.what(wavFile)
-    if not header: raise IOError("Problem opening %s" % (wavFile,))
-    (wtype,rate,channels,wframes,bits) = header
-    if bits==8: soxBits="-b -u" # unsigned
-    elif bits==16: soxBits="-w -s" # signed
-    elif bits==32: soxBits="-l -s" # signed
-    else: raise Exception("Unsupported bits per sample")
-    soxParams = "-t raw %s -r %d -c %d" % (soxBits,rate,channels)
-    rawFile = wavFile + ".raw"
-    # Now ready to convert to raw, and read it in
-    convertToRaw(soxParams,wavFile,rawFile)
-    o=open(rawFile,"rb")
-    allData=o.read()
-    o.close()
-    # Now figure out how many samples we can take out
-    bytesPerSample = channels*int(bits/8)
-    if bytesPerSample==1: silenceVal=chr(128)
-    else: silenceVal=chr(0)
-    startIdx = 0
-    while startIdx < len(allData):
-        if not allData[startIdx]==silenceVal: break
-        startIdx = startIdx + 1
-    startIdx = int(startIdx/bytesPerSample) * bytesPerSample
-    endIdx = len(allData)
-    while endIdx:
-        if not allData[endIdx-1]==silenceVal: break
-        endIdx = endIdx - 1
-    endIdx = endIdx - len(allData) # put it into -ve notatn
-    endIdx = int(endIdx/bytesPerSample) * bytesPerSample
-    endIdx = endIdx + len(allData) # avoid 0
-    sys.stderr.write("Debugger: Clipping %s to %d:%d\n" % (wavFile,startIdx,endIdx))
-    allData = allData[startIdx:endIdx]
-    # Write back the file, and convert it back to wav
-    o=open(rawFile,"wb")
-    o.write(allData)
-    o.close()
-    convertToWav(soxParams,rawFile,wavFile)
-    # Clean up
-    os.unlink(rawFile)

gradint-build/src/Makefile

-all:
-	echo "Please run 'make' in the directory above, not here."
-	exit 1

gradint-build/thindown.py

-# This file is part of the source code of
-# gradint v0.9945 (c) 2002-2009 Silas S. Brown. GPL v3+.
-#    This program is free software; you can redistribute it and/or modify
-#    it under the terms of the GNU General Public License as published by
-#    the Free Software Foundation; either version 3 of the License, or
-#    (at your option) any later version.
-#    This program is distributed in the hope that it will be useful,
-#    but WITHOUT ANY WARRANTY; without even the implied warranty of
-#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-#    GNU General Public License for more details.
-
-# program to "thin down" the gradint .py for low memory environments
-# by taking out some of the code that's unused on that platform
-
-import sys
-
-if "s60" in sys.argv: # S60 version
-  version = "S60"
-  to_omit = [
-"if use_unicode_filenames:", # WinCE
-"if paranoid_file_management:",
-'if not extsep==".":',
-"if macsound:","elif macsound:",
-"if winsound:",
-"if unix:",
-"if winsound or mingw32:",
-"elif unix and not macsound:",
-"if gotSox and unix:",
-"class SoundCollector(object):",
-"def lame_endian_parameters():",
-"def soundfile_to_data(file,soxParams):",
-# TODO SH sound collector will have problem with indentation due to """..""" strings
-"def decode_mp3(file):",
-"class Mp3FileCache(object):",
-"if outputFile:",
-"class InputSourceManager(object):",
-"def wavToMp3(directory):",
-"def makeMp3Zips(baseDir,outDir,zipNo,direc=None):",
-"class RecorderControls:",
-"def doRecWords():",
-"if app:","elif app:",
-"def addStatus(widget,status,mouseOnly=0):",
-"def removeStatus(widget):",
-"def addButton(parent,text,command,packing=None,status=None):",
-"def addLabel(row,label):",
-"def CXVMenu(e):",
-"def selectAll(e):",
-"def selectAllButNumber(e):",
-"def addTextBox(row,wide=0):",
-"def addLabelledBox(row,wide=0):",
-"def addRow(parent,wide=0):",
-"def addRightRow(widerow):",
-"def make_output_row(parent):",
-"def select_userNumber(N,updateGUI=1):",
-"def updateUserRow(fromMainMenu=0):",
-"def renameUser(i,radioButton,parent,cancel=0):",
-"def deleteUser(i):",
-"def setupScrollbar(parent,rowNo):",
-"def focusButton(button):",
-"class ExtraButton(object):",
-"def make_extra_buttons_waiting_list():",
-"def startTk():",
-"def guiVocabList(parsedVocab):",
-"def synchronizeListbox(listbox,masterList):",
-"if useTK:",
-"def openDirectory(dir):",
-"if winCEsound:",
-"def check_for_slacking():",
-"def gui_outputTo_end():",
-"def gui_outputTo_start():",
-]
-else: assert 0, "Unrecognised version on command line"
-
-revertToIndent = -1
-lCount = -1
-for l in sys.stdin.xreadlines():
-  lCount += 1
-  if lCount==2: print "\n# NOTE: this version has been automatically TRIMMED for "+version+" (some non-"+version+" code taken out)\n"
-  l=l.rstrip()
-  assert not "\t" in l, "can't cope with tabs"
-  indentLevel=-1
-  for i in range(len(l)):
-    if not l[i]==" ":
-      indentLevel = i ; break
-  if indentLevel<0 or indentLevel==len(l) or (revertToIndent>=0 and indentLevel>revertToIndent): continue
-  revertToIndent = -1
-  if (l+"#")[:l.find("#")].strip() in to_omit:
-    print (l+"#")[:l.find("#")]+" pass # trimmed"
-    revertToIndent = indentLevel
-  else: print l

hanzi-prompts/begin_zh-yue.txt

+開頭

hanzi-prompts/begin_zh.txt

+开端

hanzi-prompts/end_zh-yue.txt

+今日個堂上完啦

hanzi-prompts/end_zh.txt

+今天的课结束了

hanzi-prompts/longpause_zh-yue.txt

+而家我哋要等一陣，然後翻溫。喺第一課我哋仲未學習好多嘅詞語，所以停頓會比較長，但係喺未來嘅課程，我哋唔會有咁長嘅停頓

hanzi-prompts/longpause_zh.txt

+现在我们要等, 然后复习。 在第一课我们还没有学习很多词语所以停顿比较长。 但是在未来的课我们没有这样长的停顿。

hanzi-prompts/meaningis_zh-yue.txt

+意思係

hanzi-prompts/meaningis_zh.txt

+意思是

hanzi-prompts/nowPleaseSay_zh-yue.txt

+而家請講

hanzi-prompts/nowPleaseSay_zh.txt

+现在请说

hanzi-prompts/pleaseSay_zh-yue.txt

+請講

hanzi-prompts/pleaseSay_zh.txt

+请说