###########################################################################
# Removes two sentences from out of eight sentences.  Does this for the whole .group files
# Written by Kyuchul Yoon ( kyoon@ling.osu.edu )
# If you've got 50 files with each file containing 8 sentences, you choose two numbers.
# Then the script divides each file into 6 sentences and 2 sentences
############################################################################

# Specify files and folders
form Select files
	comment Choose one from each set.  These will be omitted from training sets
	choice SentenceNum1: 7
		button 1
		button 2
		button 3
		button 4
		button 5
		button 6
		button 7
		button 8
	choice SentenceNum2: 8
		button 1
		button 2
		button 3
		button 4
		button 5
		button 6
		button 7
		button 8
	word inputFileExt group
	word sentSeparator SFN
	integer colNumForSentSeparator 16
	word outputFileExt .cross-validate
endform

#outFile$ = "selected.txt"

Create Strings as file list... fileList *.'inputFileExt$'
Sort
numFiles = Get number of strings
pause 'numFiles' labeled textgrids identified.  Continue?

# Loop throught each file
for iFile to numFiles
	# Get the input filename
	select Strings fileList
	doneFile$ = Get string... iFile
	Read Strings from raw text file... 'doneFile$'
	Rename... myStrings
	numLines = Get number of strings

	# Initialize the index for the number of sentences
	iSentence = 1

	# Loop through each line
	for iLine to numLines
		#####  By examining the file numbers to omit, decide which output file to use
		if (sentenceNum1 <> iSentence) and (sentenceNum2 <> iSentence)
			outFileToUse$ = "selected.txt"
		else
			outFileToUse$ = "omitted.txt"
		endif
		currentLine$ = Get string... iLine
		tempString$ = currentLine$

		####### Get the column string that the user designated as separating sentences #######
		for iColumn to (colNumForSentSeparator-1)
			lenCurrentLine = length(tempString$)
			iTab = index(tempString$, tab$)
			tempString$ = right$(tempString$, (lenCurrentLine-iTab))
		endfor
		##################### End of "gettting to" the column string ################
	
		#### Now, actually extract the first three letters which could be "SFN"
		sentSep$ = left$(tempString$, 3)

		# If the line is indeed the last token (its POS being SFN), then increase the sentence number
		if (sentSep$ = "SFN")
		iSentence = iSentence + 1
		endif	

		# Write out the line according to the filename determined above
		fileappend 'outFileToUse$' 'currentLine$''newline$'
	endfor
	select Strings myStrings
	Remove
endfor
select Strings fileList
Remove

#### END OF SCRIPT ####