import NITF files

[hvelarde]

Motivation

In order to support migration from deferent legacy systems we need a general mechanism to import data from various sources. Instead of writing a different version for each system we want to use NITF and probably NewsML which are news industry standards. This way the work needed is only at the legacy system.

Assumptions

All XML files must be valid. We are going to deal only with fields defined in collective.nitf. In case of any error or inconsistence we are just going to skip the object.

Proposal & Implementation

The implementation should be something like this:

• the XML files must be located somewhere in the file system and they must be accessible from the Plone instance
• read and process the files
• create one NITF object for every XML file
• create one Image, Link or File object inside the NITF for every media included in the XML file
• reindex the new objects

Deliverables

At the end we must have 2 transmogrifier pipeline sections: one source that will read the XML files and will convert them in a list of dictionaries including all the data we need, and another that will deal with the media included in the XML file.

The sections must be fully configurable and they must discard any data that is inconsistent.

Risks

None identified.