[super] Resource archiving

[rgrp]

We want to cache/archive data associated to a resource so it is available if the resource url disappears (and in order to support other processing we may wish to do e.g. webstorer ...)

Etherpad: ​http://ckan.okfnpad.org/queue (most relevant parts inlined here)

Preliminaries

• Add task_status table to store qa/archiever/webstore information that does not need to be versioned. - #1363 (and #1371 - related logic functions)

Configuration setup for daemons

• Standard ini file
• Sections are named after daemon / extension. E.g. [my-daemon]
• Arbitrary values but anticipate at least 2 stanard values:
  • ckan_url e.g. ​http://thedatahub.org/
  • ckan_apikey e.g. xxxxxx
  • celery_config

celeryd config:

All providers of tasks will add an item to the following entry point:

[ckan.tasks]
name = ckanext.{name}.tasks:....

celeryconfig.py

from pkg_resources import iter_entry_points
for entry in iter_entry_points:
     celeryimports.appen(....)

CELERY_IMPORTS = celeryimports

Work Items

1. Resource change notifications in core - Make an IResourceChange and IResourceUrlChange. [1d] [0.75d] - #1383
2. Generate archiving request on resource url change [0.25d][0.25d] - #1399
3. Archiver daemon #891
   1. implement link-check function and task (point 2 from Archiver.update above) [1d] [0.5d]
   2. Rewrite archiver to use external storage. (decide how!)[3d][~2d]
4. Write to resource and task status table.[1d][0.75d]
5. [Required?] Make archived data available in WUI - #892
6. Documentation - #1400

Archiver process

Generate archiving request on resource url change: #

Archiver.update: see #891

Link checker: same as Archiver.update up to 2.1