<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<font face="Courier New" size="-1">I forgot to reply to the list as
well:<br>
</font>
<pre><font face="Courier New" size="-1">On 2011-09-20 15:25, Alvin Google wrote:
> On 2011-09-20 14:58, Martin Peach wrote:
>> On 2011-09-20 01:22, Alvin Google wrote:
>>> I have been interested in using streaming information in a pd patch. I
>>> have taken a look at the simplewebclient and httpget patches listed
>>> here
>>> in the thread, but I can not figure out how to target specific
>>> tables or
>>> div class information from a web page.
>>>
>>> I assume that you have to download the entire page.
>>> Can you parse for specific information : name of content and it's
>>> value?
>>>
>>
>> Yes you get the whole content-length.
>> You can either operate on a list or a table. I prefer to load
>> everything into a large table and parse that.
>> I think pdlua would be ideal for parsing text, but so far I made my
>> own abstractions from Pd objects to extract numbers from particular
>> locations in lists. The problem with converting things to symbols is
>> that the symbol table can get filled with mostly junk symbols.
>> You can go line by line by scanning for 10, the ASCII line feed
>> character.
>>
>> Martin
>>
> Thanks for your reply Martin.
> I am not sure how to go about this yet, but it sounds like one has to
> look through data line by line.
> Can you look for specific strings on a webpage like the contents of a
> specific table tag?
[mrpeach/tabfind] can be used to find a string in a table.
> Maybe there is another way to approach this?
>
There are probably many ways to do it. They are all a bit tedious with
Pd though.
I attach a copy of a patch I use. Maybe it will help.
Martin
</font></pre>
</body>
</html>