From: Erika <geagneske83@...>
Date: 2008-12-10T17:15:01+09:00
Subject: Re: pop3 body email

--0-215819000-1228897304=:41638
Content-Type: text/plain; charset=us-ascii

Hi,

I tried using Hpricot in the following way:
.....
Connect with POP3 and get the email
.....
email = TMail::Mail.parse(m.pop)
body2=Hpricot(email.body)
elements = body2.search("/html/body/table//td").collect{|k| k.inner_html.split(',') unless k.inner_html =~ /</ }.flatten.compact
puts elements

....

Is there any better way to extract the info?

My html email looks like something like this:
<table>
<tr><td>Info1</td><td>Info2</td></tr>
</table>
<table>
<tr><td>Info1</td><td>Info2</td></tr>
</table>
<table>
<tr><td>Info1</td><td>Info2</td></tr>
</table>
<table>
<tr><td>Info1</td><td>Info2</td></tr>
<tr><td>Info1</td><td>Info2</td></tr>
<tr><td>Info1</td><td>Info2</td></tr>
</table>

So the general rule is that i have 3 tables and I need to check that for Info1 the correct Info2 is shown. Every Info1 / Info2 can vary.

Is there a better way to have in to arrays for example Info1 and Info2. Because my solution is ok, only I subtract all the info from the html code which I need to parse one more time.

Thanks,
Erika


________________________________
From: Robert Dober <robert.dober@gmail.com>
To: ruby-talk ML <ruby-talk@ruby-lang.org>
Sent: Tuesday, December 9, 2008 11:02:39 PM
Subject: Re: pop3 body email

On Tue, Dec 9, 2008 at 9:29 PM, Brian Candler <b.candler@pobox.com> wrote:
> Erika wrote:
>> Hi,
>>
>> I have to check the body of the email which is a html code, like the
>> following:
>> <tr><td>Text1: </td>
>>
>> <td>Text2</td>
>>
>> I have to check that for "Text1" what "Text2" is shown, which can vary
>> for different reasons.
>
> If you are already able to read the E-mail via POP3, and the body
> consists of a single text/html part, then this just becomes a question
> about parsing HTML. Hpricot is usually cited as the best library for
> doing that. I'd say it's not worth digging about with regexps when you
> can do the job properly.
Thx a lot there, but please do not underestimate regexen, especially
as they will become more powerful in 1.9.

Are you aware of dependencies of Hpricot notwithstanding that it is a
wonderful tool.
Being an old Unix guy however I feel that you do not need a full
fledged library + dependencies if a three liner
can do the job.
However if this was only an example and OP needs more parsing Hpricot
is a very sensible way to go.
Cheers
R


      
--0-215819000-1228897304=:41638--