The plugin can be made for chrome (chrome extension) or firefox (firefox plugin) and would have the following functionalities:
(1) A list of url patterns would decide if the current page is from a targeted domain and activate the plugin button if so
(2) If the plugin is active (because it's domain is in the list described in (1)), the user will have one or several choices of buttons he can select to launch the scraper.
(3) The scraper will parse the page, producing a structured data in json format.
(4) According to what button was selected, the resulting data (originally in json format) will be saved in one of several formats (specified in the plugins' preferences), which could be:
* json (text)
* mongodb (not necessary, but would be nice)
The plugin should be made to work for a single specific website (I will communicate details to applicants), but should be organized so that I can extend it to others (simply by adding some domains and corresponding scraping code to the list).
It would be nice if the applicant knows how to do it all, but if he/she doesn't know how to parse, we can still discuss his/her application and consider giving the parsing task to someone else.