From: Luke Ivers <technodolt@...>
Date: 2007-02-01T02:33:00+09:00
Subject: Re: parsing indented plain text

On Thu, 1 Feb 2007 02:26:49 +0900
B Mills <bmill@merion.com> wrote:

> Hi.
> 
> I'm still kind of new to Ruby and scripting in general, and am trying to
> select specific entries from the plain text output of the Apple System
> Profiler which arranges data like this:
> 
> Network:
> 
>     Internal Modem:
> 
>       Type: PPP (PPPSerial)
>       Hardware: Modem
>       BSD Device Name: modem
>       IPv4:
>           Configuration Method: PPP
>       IPv6:
>           Configuration Method: Automatic
>       Proxies:
>           FTP Passive Mode: Yes
> 
>     Built-in Ethernet:
> 
>       Type: Ethernet
>       Hardware: Ethernet
>       BSD Device Name: en0
>       IPv4 Addresses: 10.3.9.249
> 
> So on and so forth. (There is an XML output option, but it uses the
> Apple plist format, which provides very little useful structure and a
> much larger filesize). I want to find a way to parse this based on the
> number of preceding white spaces at the beginning of the line (and then
> by colon), so I can get a specific property from a Hash:
> 
> profile['Network:'][ 'Built-in Ethernet:']['IPv4 Addresses:']
> 
> I started out by using grep to create an array of of each level of
> indentation, and nesting loops to select a range between two indexes of
> that level:
> 
> key1 = file.grep(/^\w/)
> key1.each do |k1|
>   range1 = file.index(key1[key1.index(k1)]
>   if k1 == key1.last
>     range2 = range1
>   else
>     range2 = file.index(key1[key1.index(k1) + 1]
>   end
>   stub = file[range1...range2]
>   key2 = stub.grep(^/\s\s\s\s\w/)
>   key2.each do |k2|
>   ....rinse and repeat...
>     key3.each do |k3|
> 
> ...and so on until I run out of indentation levels. I'm running into
> problems where the loop stops at the last element and won't dig down
> into anything below it. It also doesn't handle context, as there are
> duplicate values in the file. I also tend to have the problem of making
> things more complicated than they have to be, so I figure there is a
> more elegant and Ruby-ish way to do this. Any suggestions?
> 
> -- 
> Posted via http://www.ruby-forum.com/.
> 
This is already built for using YAML... it's structured as YAML already.
If this is saved as a file named network.yml, you can do this:
profile = {} 
File.open('network.yml') { |f| profile = YAML.load(f) }
You'll get the following
=> {"Network"=>{"Built-in Ethernet"=>{"BSD Device Name"=>"en0", "IPv4 Addresses"
=>"10.3.9.249", "Hardware"=>"Ethernet", "Type"=>"Ethernet"}, "Internal Modem"=>{
"Proxies"=>{"FTP Passive Mode"=>true}, "BSD Device Name"=>"modem", "IPv4"=>{"Con
figuration Method"=>"PPP"}, "IPv6"=>{"Configuration Method"=>"Automatic"}, "Hard
ware"=>"Modem", "Type"=>"PPP (PPPSerial)"}}}

Meaning you can do profile["Network"]["Build-in Ethernet"]...
You get the idea.
Gotta go eat lunch now.
-- 
Luke Ivers <technodolt@gmail.com>