Site Scrape Import into Database

We'd like to automate the import of DVD data into our database from partner sites. This data includes title, stars, box art photo link, dvd features, runtime, date released etc. We'd like to scan DVD UPC's and have that data passed to a vendor site via a URL to find the product. This is supported by the partner site (eg. [login to view URL]).*contains adult content* If that product is found, we'd like the the path (with our affiliate code) to be written to our database as well as the dvd information found there. ----So site #1 data to be scraped based on UPC scan: Title Product Photo Affiliate link ([login to view URL];upc=BAR CODE #) Price/format Format (ie DVD) Release Date Cast/Stars Studio Name Runtime To get other dvd info that is not found on site #1 we'd like the title to be used to search site #2 and scrape the information and saved to our database. (ie. [login to view URL];lid=listing) ----So site #2 data to be scraped based on title search from site #2: Product ID DVD Features Price/format Format (DVD, PayPerView, VideoOnDemand...) DVD Features Director We will also need a regular feed setup to grab a zip file and import the data into our database from site #3. See project clarification board for format specifications. The matching Video on Demand title will be auto-populated when searching/scraping sites #1 & #2. If the UPC scanned does not give a result from site #1, we'd like the interface to provide entering the title that will search site #2 and the data from site #2. ---So if no UPC match, scrape of site #2 based on manual entry of a title includes: Title Product ID DVD Features Price/format Format (DVD, PPV, VOD, DIVX...) DVD Features Director Cast/Stars Runtime Studio Name Release Date If no match is found on either site by UPC scanning or manual title entry, we'd like an option to either enter the data manually or skip and continue scanning. This solution can be an online or client-based application. There must be a configuration setting for affiliate codes/paths that are written to the db, paths for photos being written, etc. Offline app. (Intel/PC) must be fault proof, meaning when end of scanning is completed and an upload of the data is performed, it can't choke on one record or entry being mal-formed. (we are not committed to any specific solution-so just an example).