Write some software to match new products against a customer "wants" database & generate personalised emails for matched wants

I sell antique maps and prints online. I want to develop an application that, whenever I list new stock, will generate a personalised email to past customers highlighting any products relevant to them. I have a database containing customer email addresses and keywords or phrases of interest to each customer. This should be matched against a file containing details of new inventory. So for example, if a customer has “Paris” as a keyword, then when I list a map or print of Paris, the customer will receive an email highlighting that product. There are 2 CSV files to upload to the app: 1) a keyword file and 2) an inventory file. Examples are attached. • The inventory file is for an atlas of French departement maps. It contains a number of fields for each inventory item. The relevant ones are: o Column B) product category o D) Product Title o F) product description o G) Price in GBP o H) “Image 1” – whether or not the image has successfully uploaded to the website. Do not match if not o J) Stock quantity o K) Second product category o L) Image 1 address – the URL of the image to use in the email for the thumbnail o N) webaddress – the URL of the product page to use in the email for the link to the product • The attached dummy keyword file contains 1 keyword for each email address, but the app should be able to cope with up to 10 keywords per record. The file contains 3 additional fields – min price, max price & map or print. The match results should ignore any items below the min price or above the max. Some customers are just interested in maps, some in prints, and some in both. The match results should reflect the customer’s requirements. Each keyphrase should be matched against 1) product category 2) product title 3) first part of the product description 4) second product category. Matches should be ignored if 1) the image has not been uploaded to the website 2) the item is out of stock, 3) the price is above the max or below the min specified for that customer 4) the customer is interested in maps only, and the item is a print or vice versa The matching algorithm should be able to locate and match a multi-word keyword (eg Newcastle-upon-Tyne, St Ives) if it exists as a standalone phrase within the specified fields of the inventory file. It should be able to account for differences in capitalisation (ie to match “Paris” against “PARIS”), it should be able to deal with differences in the use of foreign characters (to match “Finistère” against “Finistere”), and in punctuation (so that “St-Tropez” matches against St Tropez”, “St James’s” matches “St Jamess”). It should not produce a match when the keyword or phrase is embedded within a longer word, except in the case of a plural (so “Paris” should not match against “Parish”, or “Eure” against “Inferieure”).