Custom API built to extract data from website 2 levels deep

General information for the business: Fashion Store XML/JSON Feed of Trending Products
Description of requirements/functionality: Need custom API script/software built to extract data from website - http://Polyvore.com and auto-generate an xml/json file.

The API must automatically pull newest "trending categories" and "trending products" (with data) from each of these categories from the website on a recurring basis.
- Must auto-generate an xml and/or json file on a recurring schedule such as: Hourly and/or when polyvore site updates
(I will be using a wordpress plugin to automatically import the xml file into my wordpress website with cronjobs)

Import Data needs to include:
- links to all the dynamically changing trending categories (with images)
- lists of products (with links, images & specific data fields as listed below) from within each trending category
(100-250 per category) and
- lists of related products (with links, images & data) as shown on each product page (50-100 per product)

More details....

1. "Trending categories" will dynamically change from this page (see link below) with their title, image and link:
Note: there are about 250 or so trending categories at any given time and update frequently

Sample list of trending categories
:
http://render.import.io/?url=http%3A%2F%2Fwww.polyvore.com%2Ftrending_now%2Fsplash.trend%3Ftrend%3DTrending%2BNow&inf=9

I need all trending categories to be automatically scraped each time page updates with new trending categories.
It must also automatically scrape a second level deep to include the linked products listed in each trend category and their specific product page details.

SPECIFIC DATA TO INCLUDE PER PRODUCT (as shown on product page):
-------------------------------------------------------------------
1. Product Title
2. Price
3. Sale Price
4. Product Description
5. Product Image
6. Product buy link
7. Product store name (.com)
8. brand (if listed)
9. # of likes (on polyvore)
10. tags (if available to include)
11. category (if available to include)
12. department (if available to include)
13. color (if available to include)
14. related key word searches (shows up on product pages as "also try searching for" links)
15. Top 50-100 related products (as shown on product page)

- All these data fields will need to be scraped from the Product pages of the related products also.
Any questions?? Let me know...
Thanks!

Specific technologies required: api, xml, python, c++, php, mysql
OS requirements: Windows, Linux, Other
Extra notes: The script/software MUST BE AUTOMATED to update this data in real-time API - for the trending categories as well as all the products and product data generated in each category. MUST BE SOFTWARE OR SCRIPT. NOT A "SCRAPING SERVICE". I also need programmer to be available for occasional updates to software (if and whenever Polyvore might change site).