Powershell - F5 iRules -- Extracting iRules - powershell

Question

I received a config file of a F5 loadbalancer and was asked to parse it with PowerShell so that it creates a .txt file for every iRule it finds. I'm very new to parsing and I can't seem to wrap my head around it.
I managed to extract the name of every rule and create a separate .txt file, but I am unable to wring the content of the rule to it. Since not all rules are identical, I can't seem to use Regex.
Extract from config file:
ltm rule /Common/irule_name1 {
SOME CONTENT
}
ltm rule /Common/irule_name2 {
SOME OTHER CONTENT
}
What I have for now
$infile = "F5\config_F5"
$ruleslist = Get-Content $infile
foreach($cursor in $ruleslist)
{
if($cursor -like "*ltm rule /*") #new object started
{
#reset all variables to be sure
$content=""
#get rulenames
$rulenameString = $cursor.SubString(17)
$rulename = $rulenameString.Substring(0, $rulenameString.Length -2)
$outfile = $rulename + ".irule"
Write-Host $outfile
Write-Host "END Rule"
#$content | Out-File -FilePath "F5/irules/" + $outfile
}
}
How can I make my powershell script read out what's between the brackets of each rule? (In this case "SOME CONTENT" & "SOME OTHER CONTENT")

Answer 1

Generally parsing involves converting a specific input ("string") into an "object" which PowerShell can understand (such as HTML, JSON, XML, etc.) and traverse by "dotting" through each object.
If you are unable to convert it into any known formats (I am unfamiliar with F5 config files...), and need to only find out the content between braces, you can use the below code.
Please note, this code should only be used if you are unable to find any other alternative, because this should only work when the source file used is code-correct which might not give you the expected output otherwise.
# You can Get-Content FileName as well.
$string = #'
ltm rule /Common/irule_name1 {
SOME CONTENT
}
ltm rule /Common/irule_name2 {
SOME OTHER CONTENT
}
'#
function fcn-get-content {
Param (
[ Parameter( Mandatory = $true ) ]
$START,
[ Parameter( Mandatory = $true ) ]
$END,
[ Parameter( Mandatory = $true ) ]
$STRING
)
$found_content = $string[ ( $START + 1 ) .. ( $END - 1 ) ]
$complete_content = $found_content -join ""
return $complete_content
}
for( $i = 0; $i -lt $string.Length; $i++ ) {
# Find opening brace
if( $string[ $i ] -eq '{' ) {
$start = $i
}
# Find ending brace
elseif( $string[ $i ] -eq '}' ) {
$end = $i
fcn-get-content -START $start -END $end -STRING $string
}
}
For getting everything encompassed within braces (even nested braces):
$string | Select-String '[^{\}]+(?=})' -AllMatches | % { $_.Matches } | % { $_.Value }

Answer 2

To parse data with flexible structure, one can use a state machine. That is, read data line by line and save the state in which you are. Is it a start of a rule? Actual rule? End of rule? By knowing the current state, one can perform actions to the data. Like so,
# Sample data
$data = #()
$data += "ltm rule /Common/irule_name1 {"
$data += "SOME CONTENT"
$data += "}"
$data += "ltm rule /Common/irule_withLongName2 {"
$data += "SOME OTHER CONTENT"
$data += "SOME OTHER CONTENT2"
$data += "}"
$data += ""
$data += "ltm rule /Common/irule_name3 {"
$data += "SOME DIFFERENT CONTENT"
$data += "{"
$data += "WELL,"
$data += "THIS ESCALATED QUICKLY"
$data += "}"
$data += "}"
# enum is used for state tracking
enum rulestate {
start
stop
content
}
# hashtable for results
$ht = #{}
# counter for nested rules
$nestedItems = 0
# Loop through data
foreach($l in $data){
# skip empty lines
if([string]::isNullOrEmpty($l)){ continue }
# Pick the right state and keep count of nested constructs
if($l -match "^ltm rule (/.+)\{") {
# Start new rule
$state = [rulestate]::start
} else {
# Process rule contents
if($l -match "^\s*\{") {
# nested construct found
$state = [rulestate]::content
++$nestedItems
} elseif ($l -match "^\s*\}") {
# closing bracket. Is it
# a) closing nested
if($nestedItems -gt 0) {
$state = [rulestate]::content
--$nestedItems
} else {
# b) closing rule
$state = [rulestate]::stop
}
} else {
# ordinary rule data
$state = [rulestate]::content
}
}
# Handle rule contents based on state
switch($state){
start {
$currentRule = $matches[1].trim()
$ruledata = #()
break
}
content {
$ruledata += $l
break
}
stop {
$ht.add($currentRule, $ruledata)
break
}
default { write-host "oops! $state" }
}
write-host "$state => $l"
}
$ht
Output rules
SOME CONTENT
SOME OTHER CONTENT
SOME OTHER CONTENT2
SOME DIFFERENT CONTENT
{
WELL,
THIS ESCALATED QUICKLY
}