How to use a web-embedded model in a computational workflow?

Question

There is a model embedded in a web browser (Caprio 1998) that I would like to use in an MCMC algorithm. What is the best way to do this?

I could implement the model in my favorite language but I would like to know if it is possible to either write a web-scraping / mechanize script to evaluates the model and dumps the output into a file, or somehow capture the existing model implementation as a desktop executable.

The interface is fairly straightforward:

And I envision a script that takes the parameters in the fields on the left side as inputs, and returns the results provided in the right column of fields as outputs:

 ./runmodel.sh 0.001 0.002 1 0.04 0.001 1.0 1.0 0.5 0 > model.out

Can I efficiently do this using the current implementation (in Java 1.0), or should I just start from scratch?

Caprio, MA (1998) The evolution of resistance: a simple deterministic model. http://web.archive.org/web/20011118202519/http://www.msstate.edu/Entomology/PGjava/ILSImodel.html

Answer 1

Extracting data from a Java applet on the web seems complicated (see this answer on SO and this forum post), but if, for example, the data was available as ordinary web forms or, in general, in a more accesible way, there are some solutions in Python for web scraping:

• Scrapy
• urllib to grab de data + lxml / BeautifulSoup to parse it (the latter option depending on its well-formedness)