Media Importer

From EChase
Jump to: navigation, search

So far i've been trying to get to grips with how bits of the webservice are actually gonna work. After i figured that out i started working on the Media Engine Interface itself. At the moment i have the ability to import single images. These images are stored in a temporary location on the webserver which can be specified by a settings file. They are also registered as temporary items in the database. Connections to the database come from a thread safe connection pool which i found and modified.

Images which are in the database are encapsulated as StoredItem objects. A StoredItem object has all the information required to analyse said image using whatever algorithm as well as use said image in any query.

The Media Engine importer's collection handeling will work primarily through the webservice and partially through external tools such as FTP, SCP and so on. The current plan is to allow the importer to deal with generic "ItemCollections" which define different ways of storing a bunch of media items. So the user could upload a zip (or jar) of items which could be extracted, or the user could upload an array of java images or the user could point to an existing directory on the webserver and so on.

Once these collections are scanned and added to the database, they can be treated like generic StoredItems and subsequently analysed using whatever algorithms required. Methods will be provided to analyse sets of StoredItems (entire collections for example) using the same algorithms.

The ability to deal with collections, analyse generic StoredItems and perform querys should be relativly easy compared to the webservice and database access stuff. Probably at the end of next week (August 5th) if not the week after that