Custom Query (2152 matches)

Reported by dread, 2 years ago.

You can't delete things from the SOLR search index if the ckan.site_id and ckan.site_url are blank.

Should assert that one of these are set up.

Reported by shevski, 21 months ago.

On harvest page: ​http://s031.okserver.org:2375/harvest the checkbox appears above the text and not clear what it refers to/

Beter to relocate to the left (as on publicdata.eu) see screenshots attached.

Reported by rgrp, 5 years ago.

1. Given a url download from that url to disk
2. Given a package use download url to download to disk

Once on disk should uncompress (if necessary).

Details

• Should download either to tmp directory (defaulted in config) or a specified path on disk
• In download would like to support as many types of target urls as possible:
  1. Single files (usually compressed filesets but could be individual files)
  2. Index pages listing files to download
  3. Revision control repositories
  4. ...

This will obviously be a significant amount of work. Suggest:

1. Start off just supporting the first option.
2. Have something like a plugin system to handle new target types
3. Reuse existing work (e.g. easy_install and pip handle standard vcs such as svn, hg, git etc)

Cost: 1d