Automating site interaction

Question

I'm working with a CMS and need to import data to it using typical html forms. The data itself is in csv files with one page per row. Such is the CMS that importing directly to db isn't possible due to the complexity of the design. It's pretty important that i "fake" usual user interaction because the CMS does a lot of background work that's crucial for the import.

Basically, for each row in the csv file, I need to copy a csv column to a html textfield, or select a checkbox, or click a certain button. One major issue is mapping the data in the csv to actions in the CMS. So if one column contains the string 'foobar' is really means "set the firstName dropdown widget to 'foobar'".

Is there a tool to automate this? I´ve been looking at AutoHotKey, Selendium, Web-Harvester and many other tools but I'm not convinced they are the correct tools. The main problem is being able to interact with the html pages in a easy way.

Answer 1

There are a bunch of tools that do that. Visual Studio Team Test Edition will do this by recording your actions and allow you to modify the resulting C# programming. You can then read from your CSV and replay in a loop.

You can also do this relatively easily if your interface doesn't change much using HTML Agility Pack.

Also I've written regular C# ( HttpWebRequest and Regex ) programs to do this and it's not very difficult either.