run_all_postannotation.rb

require 'Datavyu_API.rb'
require 'rbconfig'
# include Config


# puts $pj.getDatabaseFileName()
# puts Dir.pwd
# puts $pj.getProjectName()
case RbConfig::CONFIG['host_os']
	when /mswin|windows/i
		$valid_id_newline = File.readlines('Z:\Seedlings\usedID.txt')
	else
		$valid_id_newline = File.readlines("/Volumes/pn-opus/Seedlings/usedID.txt")
end
$valid_id = $valid_id_newline.map{ |x| x.strip} # may have to add empty string


def checkCodes(cell)
	if not $valid_utt_type.include?(cell.utterance_type.to_s)
		puts "Cell# "+cell.ordinal.to_s+":  \""+cell.utterance_type.to_s+"\" is not a valid utterance_type code\n"
	end
	if not $valid_obj_pres.include?(cell.object_present.to_s)
		puts "Cell# "+cell.ordinal.to_s+":  \""+cell.object_present.to_s+"\" is not a valid object_present code\n"
	end
	if not $valid_id.include?(cell.id.to_s)
		puts "Cell# "+cell.ordinal.to_s+":  \""+cell.id.to_s+"\" is not a valid id code (this id is not in the list of used IDs)\n"
	end
end


def is_uppercase(some_string)
	return some_string == some_string.upcase
end


def check_mwu(cell)
	if cell.object_present.to_s != "NA"
		puts "Cell# " + cell.ordinal.to_s + ":  object_present must be NA in \"%com: mwu\" cell"
	end
	if cell.speaker.to_s != "NA"
		puts "Cell# " + cell.ordinal.to_s + ":  speaker must be NA in \"%com: mwu\" cell"
	end
end

$valid_utt_type = Array["q", "d", "i", "u", "r", "s", "n", "NA"]
$valid_obj_pres = Array["y", "n", "u", "NA"]



# $valid_id_newline = File.readlines('/Volumes/pn-opus/Seedlings/usedID.txt') #TODO put the right path


begin
	columns = getColumnList()

	# Check Codes
	#
	# checkValidCodes() makes sure that all the codes found in each cell of
	# the selected column are valid codes for that specific element. The
	# function is called with the value of the "column" variable defined above.
	# You need to change this by hand every time you run it on a different column.

	# for column in columns
	# 	checkValidCodes(column, "",
	# 				"utterance_type", ["q", "d", "i", "u", "r", "s", "n", "NA"],
	# 				"object_present", ["y", "n", "u", "NA"])
	# end

	puts

	for column in columns
		col = getColumn(column)
	# Make sure that all the speaker codes are exactly 3 letters long
	# and check the case
		for cell in col.cells

			if cell.object.to_s.start_with?("%com: mwu")
				check_mwu(cell)
				next
			end

			checkCodes(cell)

			if cell.speaker.to_s.length != 3 && cell.speaker.to_s != "NA"
				puts "check codes: (3 letter code required): [Column]: "+column+"  [Variable]: speaker\t[Cell]# : "+\
					cell.ordinal.to_s + "  [Current Value]: " + cell.speaker
			end

			# Speaker code has to be all uppercase
			if !is_uppercase(cell.speaker.to_s) && !cell.object.to_s.start_with?("%com:") && !cell.object.to_s.start_with?("%pho:")
				puts "check codes: speaker code must be uppercase: [Column]: " + column + " [Cell]# : "+\
				cell.ordinal.to_s
			end

			# object_present needs to be single character and lowercase
			if (cell.object_present.to_s.length != 1) && !cell.object.to_s.start_with?("%com:") && !cell.object.to_s.start_with?("%pho:")
				puts "check codes: object_present needs to be a single character: [Column]: " + column + " [Cell]# : "+\
				cell.ordinal.to_s
			end

			if is_uppercase(cell.object_present.to_s) && !cell.object.to_s.start_with?("%com:") && !cell.object.to_s.start_with?("%pho:")
				puts "check codes: object_present needs to be lower case: [Column]: " + column + " [Cell]# : "+\
				cell.ordinal.to_s
			end

			# utterance_type needs to be single character and lowercase
			if (cell.utterance_type.to_s.length != 1) && !cell.object.to_s.start_with?("%com:") && !cell.object.to_s.start_with?("%pho:")
				puts "check codes: utterance_type needs to be a single character: [Column]: " + column + " [Cell]# : "+\
				cell.ordinal.to_s
			end

			if is_uppercase(cell.utterance_type.to_s) && !cell.object.to_s.start_with?("%com:") && !cell.object.to_s.start_with?("%pho:")
				puts "check codes: utterance_type needs to be lower case: [Column]: " + column + " [Cell]# : "+\
				cell.ordinal.to_s
			end

			split_object = cell.object.to_s.split("+")
			if split_object.length > 1
				for word in split_object[1..-1]
					if is_uppercase(word[0].chr)
						puts "check codes: only the first word in a multi-word object can be uppercase: [Cell]# : "+ cell.ordinal.to_s
					end
				end
			end


			cell.argvals.each_with_index { |code, i|
				# pho can be empty
				if code == "" && cell.speaker.to_s != "CHI" && cell.arglist[i].to_s == "pho"
					next
				end
				# codes can't be empty
				if code == ""
					puts "check_codes (Found empty code): [Column]: " + column+\
						"       [Variable]: " + cell.arglist[i].to_s + "    [Cell#]: " + cell.ordinal.to_s
				end

				# "NA" needs to be all uppercase
				if code == "na" || code =="nA" || code == "Na"
					puts "check_codes: NA needs to be uppercase: [Column]: " + column+\
						"       [Variable]: " + cell.arglist[i].to_s + "    [Cell#]: " + cell.ordinal.to_s
				end

				# codes cannot contain space, unless it's inside comment, or its a pho code.
				if cell.arglist[i].to_s != "pho" && !code.start_with?("%com:") && !cell.object.to_s.start_with?("%pho:") && code.match(/\s/)
					puts "check_codes: code cannot contain space: [Column]: " + column+\
						"       [Variable]: " + cell.arglist[i].to_s + "    [Cell#]: " + cell.ordinal.to_s
				end
			}
		end
	end



	# Check Comments
	#
	# Make sure that comment cells don't have any content for
	# the non-comment codes. Everything except the comment itself
	# should be "NA". Also makes sure that onset/offset times for
	# comments are equal.

	for column in columns
		col = getColumn(column)
        for cell in col.cells
            if (cell.object.to_s.start_with?("%com:") &&
                    ((cell.utterance_type.to_s != "NA") ||
                     (cell.object_present.to_s != "NA") ||
                     (cell.speaker.to_s != "NA")))
								if cell.object.to_s.start_with?("%com: mwu")
									next
								end
                puts "comments ERROR: one of the values is not \"NA\": [Column] "	 + column + "[Cell#]: " + cell.ordinal.to_s

                if (cell.onset != cell.offset) && !(cell.object.to_s.include?("personal information"))
                  diff = cell.offset - cell.onset
                  if diff.abs <= 1
                    next
                  end
                    puts "comments ERROR: onset and offset are not equal: [Column] " + column + " [Cell#]: " + cell.ordinal.to_s
                end

            elsif (cell.object.to_s.start_with?("%com:") &&
                   (cell.onset != cell.offset) && !(cell.object.to_s.include?("personal information")))
                puts "comments ERROR: onset and offset are not equal: [Column] " + column + " [Cell#]: " + cell.ordinal.to_s
            end
        end
	end



	# Check Intervals
	#
	# this makes sure all onsets are < offsets
	#	scan_for_bad_cells(col)
	for column in columns
		col = getColumn(column)
		for cell in col.cells

			if cell.onset > cell.offset
				puts "intervals ERROR: onset is greater than offset: [Column] " + column + " [Cell#]: " + cell.ordinal.to_s

			end
			if !cell.object.to_s.start_with?("%com") && !cell.object.to_s.start_with?("%pho:") && (cell.onset == cell.offset)
				puts "intervals ERROR: onset and offset are equal in non-comment cell: [Column] " + column + " [Cell#]: " + cell.ordinal.to_s
			end
		end
	end
	puts "\n\n\n"


	# Personal Info
	#
	# This part runs through the datavyu file
	# and pulls out all the personal info comments.
	# It then writes out a file containing the
	# timestamps of those regions. That output file
	# should then be fed into the program that actually
	# does the silencing.
	#
	#
	# These comments should take the form:
	#
	#   %com: personal information [audio]: credit card number
	#
	# or
	#
	#   %com: personal information [video]: butt
	#
	#
	#  You need to set the output path of the .csv file that
	#  will be generated. This is the first line after this comment


	# output_path = $pj.getProjectDirectory() + File::SEPARATOR + $pj.getProjectName() + "_personal_info.csv"
	# output_path.gsub!('\\', '/')
	#
	# split_path = output_path.split(File::SEPARATOR)
	#
	# case CONFIG['host_os']
	# when /mswin|windows/i
	# 	no_persinfo_file_path = File.join(split_path[0], "Seedlings/Scripts_and_Apps/no_personal_info.txt")
	# else
	# 	no_persinfo_file_path = "/Volumes/pn-opus/Seedlings/Scripts_and_Apps/no_personal_info.txt"
	# end
	#puts output_path
	#output = File.expand_path(output_path)

	# no_pi = IO.readlines(no_persinfo_file_path)
	# already_in_nopi = false
	# for element in no_pi
	# 	if element.include? $pj.getProjectName()
	# 		already_in_nopi = true
	# 	end
	# end

	for column in columns

        col = getColumn(column)
				if col.cells.length == 0
					puts "\nPlease delete the empty extra column....\n"
					next
				end
        # arrays containing millisecond onset/offsets for personal information
        audio_regions = Array.new
        video_regions = Array.new

        entry = nil
				for cell in col.cells

            entry = cell.object.to_s
            if (entry.start_with?("%com: personal info"))
                if (entry.include?("[audio]"))
                    audio_regions.push([cell.onset, cell.offset])
								end
                if (entry.include?("[video]"))
                    video_regions.push([cell.onset, cell.offset])
								end
                if (!(entry.include?("[audio]") || entry.include?("[video]")))
                    puts "Malformed personal information comment:  cell#: " + cell.ordinal.to_s
                end
            end
				end

				# used_to_be_nopi = false
				# if already_in_nopi && (!audio_regions.empty? || !video_regions.empty?)
				# 	used_to_be_nopi = true
				# 	# no_pi.delete($pj.getProjectName())
				# 	# no_pi.delete($pj.getProjectName()+"\n")

				# 	File.open(no_persinfo_file_path, 'w') do |f|
				# 		f.truncate(0)
				# 		# f.puts(no_pi)
				# 	end
				# 	puts $pj.getProjectName() + " used to have no personal info. It has been removed from the no_personal_info list\n\n"
				# end

				# if there are no personal info regions, add the name of the file
				# to the no_personal_info.txt manifest in /seedlings/Scripts_and_Apps/
				# if (audio_regions.empty? && video_regions.empty? && !already_in_nopi)
				# 	open(no_persinfo_file_path, "a") do |f|
				# 		f.puts $pj.getProjectName()
				# 	end
				# 	puts "There were no personal info regions"
				# 	exit
				# end
				# if already_in_nopi && !used_to_be_nopi
				# 	exit
				# end
				# output_file = File.open(output_path, "w")

        # for region in audio_regions
        #         output_file.puts("audio,#{region[0]},#{region[1]}")
        # end

        # for region in video_regions
        #     output_file.puts("video,#{region[0]},#{region[1]}")
        # end

        # output_file.close()

        # puts "personal info timestamps written to: " + output_path + "\n\n"
	end



end