function convert-pdfdict2hash {
	param ([System.IO.FileStream]$pdstream, [long]$pdoffset)
	#converts << >> delimited pdf dictionary to hash. Called from convert-pdf2txt and show-pdfcontent
	$PDResult=@{}
	[byte[]]$pdbuffer=@(0xff,0xff)
	[bool]$addtohash=$false
	$line=''
	for ($pdfPos=$pdoffset+1;$pdfPos -lt $pdstream.length;$pdfpos=$pdfpos+2) {
		$pdstream.position=$pdfpos
		$iChars=$pdstream.read($pdbuffer,0,$pdbuffer.length) > $nul
		$line=$line+([char]$pdbuffer[0])+([char]$pdbuffer[1])
		$line=$line.replace([char]10+[char]13,[char]10)
		$line=$line.replace([char]13+[char]10,[char]10)
		$line=$line.replace([char]13,[char]10)

		if ($pdbuffer[0] -eq 10 -or $pdbuffer[1] -eq 10 -or $pdbuffer[0] -eq 13 -or $pdbuffer[1] -eq 13) {
			$newline=''
			if ($pdbuffer[0] -eq 10 -or $pdbuffer[0] -eq 13) {$newline=([char]$pdbuffer[1])}
			$line=$line.split([char]10)[0].trim()
			if ($line -eq '>>') {
				$pdfPos = $pdstream.length
			}
			elseif ($line -eq '<<') {
				$addtohash=$true
			}
			elseif ($addtohash) {
				$pdKey=$line.split(' ')[0].replace('/','')
				$valueStart=$line.indexof($pdKey)+$pdkey.length
				$pdValue=($line.substring($valueStart,$line.length-$valueStart)).replace('<<','').replace('[','').replace(']','').trim()

				#add code to handle data types dictionary and array
				#dictionary within dictionary
				#if ($pdValue.contains('<<') {
				#	$pdValue=convert-pdfdict2hash $pdstream $pdOffset
				#}
				#array
				#elseif ($pdvalue.contains('[') {
				#
				#}

				if (!$pdResult.containskey($pdKey)) {
					$pdResult.add($pdKey,$pdValue)
				}
				else {
					$pdResult[$pdKey]=$pdResult[$pdKey]+'%'+$pdValue
				}
			}
			$line=$newline
		}
	}	
	$PDResult
}

function show-pdfcontent {
	#requires function convert-pdfdict2hash. Called from convert-pdf2txt
	param ([System.IO.FileStream]$scfilestream, [long]$iscoffset)

	$dContentInfo=@{}
	$iContentLength=0
	$sContentFilter=''
	$line=''
	[byte[]]$scbuffer=@(0xff,0xff)
	$dContentInfo=convert-pdfdict2hash $scfilestream $iscoffset
	$iContentLength=$dContentInfo['length']
	if ($dContentInfo.containskey('filter')) {
		$sContentFilter=$dContentInfo['filter']
		switch ($sContentFilter.tolower()) {
			{"/flatedecode"} {
				#Get Stream offset
				for ($iFindStream=$iscoffset+1;$iFindStream -lt $scfilestream.length;$iFindStream=$iFindStream+2) {
					$scfilestream.position=$iFindStream
					$iChars=$scfilestream.read($scbuffer,0,$scbuffer.length) > $nul
					$line=$line+([char]$scbuffer[0])+([char]$scbuffer[1])
					if ($scbuffer[0] -eq 10 -or $scbuffer[1] -eq 10) {
						$newline=''
						if ($scbuffer[0] -eq 10) {$newline=([char]$scbuffer[1])}
						$line=$line.split([char]10)[0].trim()
						#write-host $line
						if ($line.tolower().contains('stream')) {
							#Header Can be two or six bytes??? See RFC 1950 & 1951 and http://blogs.msdn.com/bclteam/archive/2007/05/16/system-io-compression-capabilities-kim-hamilton.aspx.
							$iStreamOffset=$iFindStream + 4
							if ($newline) {$iStreamOffset = $iStreamOffset - 1} #$iFindStream needs to be decremented by 1 if stream fell between two bytes we're reading.
							$iFindStream=$scfileStream.length
						}
						$line=$newline
					}
				}
				#Deflate zlib compressed stream
				#"Stream offset: $($iStreamOffset)"
				$scfilestream.position=$iStreamOffset
				$decompressed=new-object System.IO.Compression.DeflateStream -argumentlist $scfilestream,Decompress
				$sdecomp=''
				for ($iSC=1;$iSC -lt $iContentLength;$iSC++) {
						$decompressed.read($scbuffer,0,$scbuffer.length) > $nul
						$sdecomp=$sdecomp+[char]$scbuffer[0]+[char]$scbuffer[1]
						#$scbuffer|%{write-host -nonewline ([char]$_)}
				}
				$finalresult=''
				foreach ($sLine in $sDecomp.split([char]10)) {
					#$sline
					if ($sline -match "^\((.*)\) (T.|\'|\"")$") {
						#$finalresult=$matches[1]
						$matches[1]
					}
				}
				
			}
		} #switch
	}


}



function convert-pdf2txt {
	#Requires function: convert-pdfdict2hash
	param ([string]$pdfsource, [string]$txtdest)

	if ($pdfsource.length -lt 4) {$pdfsource=$pdfsource+'.pdf'}
	if (!$pdfsource.endswith('.pdf')) {$pdfsource=$pdfsource+'.pdf'}

	if (!$txtdest) {$txtdest=$pdfsource.substring(0,$pdfsource.length - 4)+'.txt'}
	if (!(test-path $pdfsource)) {
		"Couldn't find file $($pdfsource)"
		break
	}

	$pdfsource=([System.IO.FileInfo]"$($pdfsource)").fullname
	trap { 
		if ($_.Exception.GetType().Name -eq "FileNotFoundException") {
			"File not found exception"	
			"Try copying the file to a local drive."
			break
		}
		else {
			"trapped " + $_.Exception.GetType().Name + " in script "+ $error[0].invocationinfo.scriptname + " at line " + $error[0].invocationinfo.scriptlinenumber + " offset "+ $error[0].invocationinfo.OffsetInLine +":"
			$error[0]
		}
	}

	$pdfstream=[IO.File]::OpenRead($pdfsource)
	[byte[]]$pdfbuffer=@(0xff,0xff)

	#Get xref into hash
	for ($pdfPos=$pdfstream.length-2;$pdfPos -gt 1;$pdfpos=$pdfpos-2) {
		$pdfstream.position=$pdfPos
		$iChars=$pdfstream.read($pdfbuffer,0,$pdfbuffer.length) > $nul
		$tail=([char]$pdfbuffer[0])+([char]$pdfbuffer[1])+$tail
		if ($tail.tolower().contains("startxref")) {$pdfpos=0}
		
	}
	
	$tail=$tail.replace([char]10+[char]13,[char]10)
	$tail=$tail.replace([char]13+[char]10,[char]10)
	$tail=$tail.replace([char]13,[char]10)

	foreach ($ln in $tail.split([char]10)) {
		if ($ln -match "^[0-9].*$") {
			$xrefoffset=[int]$matches[0]
		}
	}

	if (!$xrefoffset) {"Couldn't get xref offset. Quitting";break}
	#$xrefoffset=[int]$tail.split([char]10)[2]
	
	$xref=@()
	for ($pdfPos=$xrefoffset;$pdfPos -lt $pdfstream.length;$pdfpos=$pdfpos+2) {
		$pdfstream.position=$pdfpos
		$iChars=$pdfstream.read($pdfbuffer,0,$pdfbuffer.length) > $nul
		$line=$line+([char]$pdfbuffer[0])+([char]$pdfbuffer[1])
		$line=$line.replace([char]10+[char]13,[char]10)
		$line=$line.replace([char]13+[char]10,[char]10)
		$line=$line.replace([char]13,[char]10)
		if ($line.contains([char]10)) {
			$xref+=$line.replace([char]10,'').trim()
			$line=''
		}

	}	

	$dxref=@{}
	for ($i=[int]$xref[1].split(' ')[0];$i -lt [int]$xref[1].split(' ')[1];$i++) {
		$dxref.add([int]$i,[long]$xref[$i+2].split(' ')[0])
	}

	#Get root obj # from trailer
	for ($pdfPos=$pdfstream.length-2;$pdfPos -gt 1;$pdfpos=$pdfpos-2) {
		$pdfstream.position=$pdfPos
		$iChars=$pdfstream.read($pdfbuffer,0,$pdfbuffer.length)
		$tail=([char]$pdfbuffer[0])+([char]$pdfbuffer[1])+$tail
		if ($tail.tolower().contains("railer")) {$pdfpos=0}
	}

	$tail=$tail.replace([char]10+[char]13,[char]10)
	$tail=$tail.replace([char]13+[char]10,[char]10)
	$tail=$tail.replace([char]13,[char]10)
	foreach ($trailerLine in $tail.split([char]10)) {
		$trailerLine=$trailerLine.tolower().trim()
		if ($trailerline.startswith("/root ")) {$iRootObj=$trailerLine.split(' ')[1]}
	}

	#Get Root object offset from xref hash
	#"Root offset: $($dxref[[int]$iRootObj])"
	$rootOffset=$dxref[[int]$iRootObj]

	if ($rootOffset) {

		#"Pages: $($dRoot['Pages'])"

		$dRoot=convert-pdfdict2hash $pdfstream $rootOffset
		$iPagesObjectNum=$dRoot['Pages'].split(' ')[0]
		#"Pages Offset: $($dxref[[int]$iPagesObjectNum])"

		$iPagesOffset=$dxref[[int]$iPagesObjectNum]
	
		$dPages=convert-pdfdict2hash $pdfstream $iPagesOffset
		$kids=$dPages['Kids'].split(' ')

		#Every 3rd entry in $kids is object number of page
		for ($iKid=0;$iKid -lt $kids.length;$iKid=$iKid+3) {
			$iPageObjectNum=$kids[$ikid]
			$iPageOffset=$dxref[[int]$iPageObjectNum]
			$dPage=@{}
			$dPage=convert-pdfdict2hash $pdfstream $iPageOffset
			$contents=$dPage['Contents']
			if ($contents.contains(' ')) {
				$contents=$contents.split(' ')
				#every 3rd entry is object number of content object
				for ($icontentnum=0;$icontentnum -lt $contents.length;$icontentnum=$icontentnum+3) {
					$iContentObjectNum=$contents[$iContentnum]
					$icontentOffset=$dxref[[int]$iContentObjectNum]
					#"Content offset: $($iContentOffset)"
					show-pdfcontent $pdfstream $iContentOffset
				}
			}
		}
	
	}
	#Can't find root offset and page structure.  Read through file to find streams and send to show-pdfcontent function. 
	else {
		"Document may be encrypted!";break
		$thislineoffset=0
		for ($pdfPos=0;$pdfPos -lt $pdfstream.length;$pdfpos=$pdfpos+2) {
			$pdfstream.position=$pdfpos
			$iChars=$pdfstream.read($pdfbuffer,0,$pdfbuffer.length) > $nul
			$line=$line+([char]$pdfbuffer[0])+([char]$pdfbuffer[1])

			$line=$line.replace([char]10+[char]13,[char]10)
			$line=$line.replace([char]13+[char]10,[char]10)
			$line=$line.replace([char]13,[char]10)

			if ($pdfbuffer[0] -eq 10 -or $pdfbuffer[1] -eq 10 -or $pdfbuffer[0] -eq 13 -or $pdfbuffer[1] -eq 13) {
				$newline=''
				if ($pdfbuffer[0] -eq 10 -or $pdfbuffer[0] -eq 13) {$newline=([char]$pdfbuffer[1])}
				$line=$line.split([char]10)[0].trim()
				#write-host $line
				
				if ($line.tolower().contains(' obj')) {
					$lastobj=$thislineoffset
				}
				if ($line.tolower().contains('stream')) {
					show-pdfcontent $pdfstream $lastObj
				}
				$line=$newline
				$thislineoffset=$pdfpos+2
				if ($pdfbuffer[0] -eq 10 -or $pdfbuffer[0] -eq 13) {$thislineoffset--}
			}

		}	
	}
		#$pdfstream.position=$rootOffset
		#$iChars=$pdfstream.read($pdfbuffer,0,$pdfbuffer.length)
		#$tail=([char]$pdfbuffer[0])+([char]$pdfbuffer[1])
		#$tail
}

$usage="type-pdf.ps1 <doc.pdf>`nWill display unformatted plain text from pdf file if the file uses zlib compression."

if ($args.length -ne 1) {$usage;break}

if ($args[0].contains('?') -or $args[0].contains('*')) {$usage;break} 

if ($args.count -eq 1) {
	convert-pdf2txt $args[0]
}
else {
	convert-pdf2txt $args[0] $args[1]
}