Difficulties POSTing to RDP Hierarchy Browse Page

[Chris Lasher]

Hello,
I'm trying to write a tool to scrape through some of the Ribosomal
Database Project II's (http://rdp.cme.msu.edu/) pages, specifically,
through the Hierarchy Browser. (http://rdp.cme.msu.edu/hierarchy/)
The Hierarchy Browser is accessed first through a page with a form.
There are four fields with several options to be chosen from (Strain,
Source, Size, and Taxonomy) and then a submit button labeled "Browse".
The HTML of the form is as follows (note, I am also including the
Javascript code, as it is called by the submit button):

--------excerpted HTML----------------
<script language="Javas cript">

function resetHiddenVar( ){
var f_form = document.forms['hierarchyForm'];
f_form.action= "HierarchyContr ollerServlet/start";
return ;
}

</script>

<form name="hierarchy Form" method="POST"
action="Hierarc hyControllerSer vlet/start/">
<input type='hidden' name='printPara ms' value='no' />

<h1>Hierarchy Browser - Start</h1><div class="cart" style="float:
right">[&nbsp;<a href="hb_help.j sp">help</a>&nbsp;]</div>

<p>&nbsp;</p>
<div id="options">

<table summary="option s area" cellpadding="0" cellspacing="0"
border="0"><tr> <td align="left" valign="middle" >
<table border="0" cellspacing="0" cellpadding="0" summary="Option s"
align="left" class="borderup ">
<tr>
<th align="right" valign="middle" class="bottom greenbg"
nowrap="nowrap" >Strain:</th>
<td class="bottom formtext" nowrap="nowrap" ><input id="type"
name="strain" type="radio" value="type">
<label for="type">Type </label></td>
<td class="bottom formtext" nowrap="nowrap" ><input id="nontype"
name="strain" type="radio" value="nontype" >
<label for="nontype">N on Type</label>&nbsp;</td>
<td class="bottom formtext" nowrap="nowrap" ><input name="strain"
type="radio" id="strainboth " value="both" checked>
<label for="strainboth ">Both</label>&nbsp;</td>

</tr>
<tr>
<th align="right" valign="middle" class="bottom greenbg">Source :</th>
<td class="bottom formtext" nowrap="nowrap" ><input id="environment al"
name="source" type="radio" value="environ" >
<label for="environmen tal">Uncultured &nbsp;</label></td>
<td class="bottom formtext" nowrap="nowrap" ><input id="isolates"
name="source" type="radio" value="isolates ">
<label for="isolates"> Isolates</label></td>
<td class="bottom formtext" nowrap="nowrap" ><input name="source"
type="radio" id="sourceboth " value="both" checked >
<label for="sourceboth ">Both</label></td>
</tr>

<tr>
<th align="right" valign="middle" class="bottom greenbg">Size:</th>
<td class="bottom formtext" nowrap="nowrap" ><input
id="greaterthan 1200" name="size" type="radio" value="gt1200" checked>
<label for="greatertha n1200"><u>&gt;</u>1200</label></td>
<td class="bottom formtext" nowrap="nowrap" ><input id="lessthan120 0"
name="size" type="radio" value="lt1200">
<label for="lessthan12 00">&lt;1200</label></td>
<td class="bottom formtext" nowrap="nowrap" ><input id="sizeboth"
name="size" type="radio" value="both">
<label for="sizeboth"> Both</label></td>
</tr>
<tr>

<th align="right" valign="middle" class="bottom
greenbg">Taxono my:</th>
<td class="bottom formtext" nowrap="nowrap" ><input id="bergeys"
name="taxonomy" type="radio" value="rdpHome" checked>
<label for="bergeys">B ergey's</label></td>
<td colspan="2" class="bottom formtext" nowrap="nowrap" ><input
id="ncbi" name="taxonomy" type="radio" value="ncbiHome ">
<label for="ncbi">NCBI </label></td>
</tr>
</table>
</td>
<td align="left" valign="middle" >&nbsp;&nbsp;&n bsp;
<input name="browse" type="submit" id="browse"
onclick="resetH iddenVar(); return true;" value="Browse">

</td></tr></table></p>
</div>
<!-- end options -->
</form>
----------end excerpted HTML--------------
The options I would like to simulate are browsing by strain=type,
source=both, size = gt1200, and taxonomy = bergeys. I see that the
form method is POST, and I read through the urllib documentation, and
saw that the syntax for POSTing is urllib.urlopen( url, data). Since
the submit button calls HierarchyContro llerServlet/start (see the
Javascript), I figure that the url I should be contacting is
http://rdp.cme.msu.edu/hierarchy/Hie...rServlet/start
Thus, I came up with the following test code:

--------Python test code---------------
#!/usr/bin/python

import urllib

options = [("strain", "type"), ("source", "both"),
("size", "gt1200"), ("taxonomy", "bergeys"),
("browse", "Browse")]

params = urllib.urlencod e(options)

rdpbrowsepage = urllib.urlopen(
"http://rdp.cme.msu.edu/hierarchy/HierarchyContro llerServlet/start",
params)

pagehtml = rdpbrowsepage.r ead()

print pagehtml
---------end Python test code----------
However, the page that is returned is an error page that says the
request could not be completed. The correct page should show various
bacterial taxonomies, which are clickable to reveal greater detail of
that particular taxon.

I'm a bit stumped, and admittedly, I am in over my head on the subject
matter of networking and web-clients. Perhaps I should be using the
httplib module for connecting to the RDP instead, but I am unsure what
methods I need to use to do this. This is complicated by the fact that
these are JSP generated pages and I'm unsure what exactly the server
requires before giving up the desired page. For instance, there's a
jsessionid that's given and I'm unsure if this is required to access
pages, and if it is, how to place it in POST requests.

If anyone has suggestions, I would greatly appreciate them. If any
more information is needed that I haven't provided, please let me know
and I'll be happy to give what I am able. Thanks very, very much in
advance.

Chris

[Alan Kennedy]

[Chris Lasher]

I'm trying to write a tool to scrape through some of the Ribosomal
Database Project II's (http://rdp.cme.msu.edu/) pages, specifically,
through the Hierarchy Browser. (http://rdp.cme.msu.edu/hierarchy/)
I'm sure that urllib is the right tool to use. However, there may be one
or two problems with the way you're using it.
--------excerpted HTML----------------
<!-- snip -->
<form name="hierarchy Form" method="POST"
action="Hierarc hyControllerSer vlet/start/">
<input type='hidden' name='printPara ms' value='no' />
This is an omission from the params you are passing to the
HierarchyServle t. Although the "printParam s" field is not visible to you
in a browser, the browser still submits a name/value pair in its form
submission. So you should also in your code, as shwon below.
<input id="bergeys" name="taxonomy" type="radio" value="rdpHome" checked>
Also, you are using the wrong value for the taxonomy field. You are
setting a value of "bergeys", which is the ID of the field, not its
value. The correct value is "rdpHome".
--------Python test code---------------
#!/usr/bin/python

import urllib

options = [("strain", "type"), ("source", "both"),
("size", "gt1200"), ("taxonomy", "bergeys"),
("browse", "Browse")]
Try this

options = [ ("printParam s", "no"), ("strain", "type"),
("source", "both"), ("size", "gt1200"),
("taxonomy", "rdpHome"), ("browse", "Browse"),]

params = urllib.urlencod e(options)

rdpbrowsepage = urllib.urlopen(
"http://rdp.cme.msu.edu/hierarchy/HierarchyContro llerServlet/start",
params)

pagehtml = rdpbrowsepage.r ead()

print pagehtml
---------end Python test code----------

HTH,

--
alan kennedy
------------------------------------------------------
email alan: http://xhaus.com/contact/alan