[LON-CAPA-admin] auto-enrollment again

[Stuart Peter Raeburn]

Peter, 

The piece that needs to be customized for your institution is a 
fetch_enrollment() routine that will put XML files containing enrollment 
data in /home/httpd/perl/tmp/ 

with file names such as: 

gwu_134512831e340c5gwul1_ss06phy142201_classlist.xml 

where gwu is the domain of the course, 134512831e340c5gwul1 is the course 
name, and ss06phy142201 is an institutional course code for a section which 
maps to this particular LON-CAPA course.  There could be several of these 
files for a single LON-CAPA course, if the course has multiple sections.  
Files stored in /home/httpd/perl/tmp/ are deleted each time after the 
automated enrollment process has occurred. 

If you do not have access to a datastore which you can query for enrollment 
data, but instead will be using CSV files, then your implementation would 
most likely involve uploading the CSV files to some (permanent) directory on 
the file system on a regular basis. 

A fetch_enrollment() routine would then be crafted to convert the contents 
of the CSV files to XML and to write the output to appropriately named files 
in /home/httpd/perl/tmp/. fetch_enrollment() is called whenever 
auto-enrollment needs to occur, either for a single course as prompted via 
Course Coordinator action via ENRL -> Automated Enrollment Manager -> 
"Update Roster Now", or for a number of courses when 
/home/httpd/perl/Autoenroll.pl is called by the last entry in 
/etc/cron.d/loncapa 

A decision needs to be made about the format of the course code. In general  
you will need to adopt a format that provides for a unique code for each 
course section in your system (e.g., a course code of ss06phy142), with 
section numbers such as 001, 002 etc. used for the different sections. 

If you wish to use the dynamic select boxes to gather information from 
students submitting help requests via the "Ask Helpdesk" form on the log-in 
page you will want to provide a means of extracting information about the 
identity of the course from the course code.  There are two general 
approaches - either you maintain a look-up table that does this for you, or 
you adopt a format that can be broken into its constituent parts (year, 
semester or term, department, course number) via a perl regular expression. 

LON-CAPA places a few constraints on the format of the course code.  The 
following : , are not allowed in the course code as they are used as 
separators for storage of multiple sections or crosslistings.  This could 
change in the future by use of escapes, but for now the best strategy might 
be to restrict characters used in your course codes to letters, numbers and 
underscores. 

Once you have established the format you will use for the course code, as DC 
yu can use MCRS to associate course codes with courses, and ENRL -> 
Automated Enrollment Manager to associate section numbers and/or 
crosslistings with courses (and identify the desired GWU section -> LON-CAPA 
section mapping). 

Some more informattion about setting up auto-enrollment when enrollment data 
is stored in a directory, and retrieved from there rather than via a query 
to a data store can be found at:  
http://mail.lon-capa.org/pipermail/lon-capa-admin/2004-October/000721.html 

I can provide assistance if you decide to move forward with this, and could 
put together a CSV-to-XML conversion routine in perl.  To generalize this, 
it would probably be useful to provide an interface so that a DC could 
define a default mapping of columns to attributes.  This might be one of a 
number of categories of domain-wide data that we might start to store on the 
primary library server for a domain (a concept introduced in LON-CAPA 2.1 in 
the context of the MAIL function accessible to Domain Coordinators). 

This same applies if you decided to simply use the "Upload Classlist" 
function currently in ENRL (rather than implement auto-enrollment) - it 
would be useful if we could provide a means for a DC to define a default 
arrangement columns in uploaded CSV data to avoid the need for the user to 
select them each time they used upload claslist to a new course. Of course 
by going this route instead of auto-enrollment you do miss out out on the 
benefits of auto-dropping of students who are officially no longer 
registered for a class. 

If you prefer, feel free to contact me directly (offlist) if you have 
specific questions relating to implementation at GWU that need to be 
addressed. 

Stuart 

Stuart Raeburn
Div. Science & Math Education
Michigan State University
East Lansing MI 48824 


Peter Kovac writes: 

>    Hello again.  Per Guy's suggestion, I took a look through the comments 
> in /home/httpd/lib/perl/localenroll-std.pm  My perl is a bit rusty (been a 
> few years) and, to be honest, tends to be more hacks than pretty code so I 
> apologize if I seem dense. 
> 
>    First, my eventual goal.  I have been completely unsuccessful at 
> getting access to a proper enrollment datastore.  However, our faculty can 
> easily produce classlists in tabular (CVS/Excel) format.  I'd like to hack 
> together something that will convert uploaded classlist tables into proper 
> XML and insert them into the auto-enrollment system.  I will, of course, 
> pass on said ugly hack to whomever wants it. 
> 
>    I gather that fetch_enrollment is designed to generate XML files that 
> are then imported as class lists by other parts of the system.  My reading 
> of the perl seems to indicate that the actual work is done by code 
> somewhere other than /home/httpd/lib/perl/localenroll-std.pm or that I'm 
> supposed to write the interface between the enrollment and our data source 
> (vaguely recall something like that said at a conference).  
> 
>    So, any suggestions for other places to look to get a better handle on 
> the enrollment framework?  Other advice? 
> 
>    Thanks,
>    Peter
> _______________________________________________
> LON-CAPA-admin mailing list
> LON-CAPA-admin at mail.lon-capa.org
> http://mail.lon-capa.org/mailman/listinfo/lon-capa-admin 
>