{22} Trac tickets (2647 matches)

Both in 1.5.1b:

​http://thedatahub.org/api/search/dataset?groups=lodcloud&title=

and 1.5.2a (current master):

​http://test.ckan.net/api/search/dataset?groups=lodcloud&title=

Although the error message in 1.5.2a is more verbose:

"Bad request - Bad search option: HTTP code=400, reason=org.apache.lucene.queryParser.ParseException?: Cannot parse 'groups:lodcloud title:': Encountered \"<EOF>\" at line 1, column 22. Was expecting one of: \"(\" ... \"*\" ... <QUOTED> ... <TERM> ... <PREFIXTERM> ... <WILDTERM> ... \"[\" ... \"{\" ... <NUMBER> ..."

Some parameter validation before sending it to Solr should do the trick

• deploy celery
• deploy QA and archiver tasks
• write up a blog post announcing QA on thedatahub

• Update CKAN on DGU UAT to 1.5.1
• Deploy Celery
• Deploy QA extension

thedatahub.org/about contains info that is very general to the CKAN and really quite technical. The text should be changed to be both specific to thedatahub.org and provide the context in a non-technical way. It should be easy to customise the About page to be appropriate for say new-york.ckan.net - a bit of info about who runs it, plus the general stuff about CKAN powering it and it was written by OKF to further open data.

Currently there is a problem because the docs guide us to set the sqlalchemy url to use 'localhost' i.e. loopback, whereas paster commands don't specify '-h localhost' so use unix sockets (you need to do 'sudo -u postgres'). These should be the same.

Also do we need to tell people to add a line to their postgres authentication config /etc/postgresql/8.4/main/pg_hba.conf to help things? Florian suggests:

local   std         std                          md5

/etc/postgresql/8.4/main/pg_hba.conf

The dict returned by package_dictize does not include metadata_modified and metadata_created. These are really useful properties, so it's worth having them on the standard package dict representation, which is used in several places, like at the template level.

The download task in ckanext-archiver performs a HEAD request on the resource URL and checks if the "Content-Type" and "Content-Length" headers differ from the values stored to see if the resource needs to be updated [1].

The "Content-Length" header, although widely used, is not mandatory and some servers don't provide it, e.g.:

$ curl -I http://portfolio.theglobalfund.org/en/IATI/Activities?countryCode=AFG
HTTP/1.1 200 OK
Cache-Control: private
Transfer-Encoding: chunked
Content-Type: text/xml
Vary: Accept-Encoding
Server: Microsoft-IIS/7.5
Set-Cookie: ASP.NET_SessionId=3qhqekddgmre0kmk5cynq0sy; path=/; HttpOnly
X-AspNetMvc-Version: 3.0
content-disposition: attachment; filename=AFG_IATI_12012012.xml
X-AspNet-Version: 4.0.30319
X-Powered-By: ASP.NET
Date: Thu, 12 Jan 2012 12:36:43 GMT

Also worth noting that ​requests, the python library that uses ckanext-archiver, sets an "Accept-Encoding: gzip" header by default, which depending on the configuration of the remote web server, may prevent the "Content-Length" server from being sent, e.g.:

$ curl -H "Accept-Encoding: gzip" -I http://iatistandard.org/published-temp/adb-activities.xml
HTTP/1.1 200 OK
Date: Thu, 12 Jan 2012 12:12:46 GMT
Server: Apache
Last-Modified: Mon, 28 Nov 2011 15:55:35 GMT
Accept-Ranges: bytes
Vary: Accept-Encoding
Content-Encoding: gzip
Content-Type: application/xml

curl -I http://iatistandard.org/published-temp/adb-activities.xml
HTTP/1.1 200 OK
Date: Thu, 12 Jan 2012 11:56:23 GMT
Server: Apache
Last-Modified: Mon, 28 Nov 2011 15:55:35 GMT
Accept-Ranges: bytes
Content-Length: 2686720
Vary: Accept-Encoding
Content-Type: application/xml

All this can lead to some resources never getting updated, and of course the size property of the resource not being set.

As we need to download the resource anyway, it would be better to check if the real length of the data has been modified (and store it).

[1] ​https://github.com/okfn/ckanext-archiver/blob/0a189262dca4ab5b286fb6a02b4ab8a201f639f3/ckanext/archiver/tasks.py#L72

Instead of alphabetically as we do currently, alternatively by most viewed datasets

for ​http://thedatahub.org/dataset

In this commit ​https://github.com/okfn/ckan/commit/1772a5c John Glover set map.explicit=True in ckan/config/routing.py.

The reason this was done was to avoid links collecting parameters. e.g. if you were on page /dataset/{id}/resource/{resource_id} then by default all the links on that page generated by url_for (Routes) would include the id and resource_id parameters as well. To avoid this, you had to go through all the links and add id=None and resource_id=None to the url_for parameters.

When map.explicit was changed to True, the value of the controller, action, id and any parameters were no longer automatically carried over into the generated links for the page. So previously links within the same controller didn't need to specify the controller (for example), but now they did. So when we did this we also had to fix up links that weren't explicit:

• ​https://github.com/okfn/ckan/commit/88e88380f61965f58fb105095e76a78673ce3e2c
• #1501 Pagination links
• ​https://github.com/okfn/ckan/commit/52d71102d9b07693ea240d4a66e8adb43e515414

John made the config change on 5/11/2011 which was merged to master ​https://github.com/okfn/ckan/commit/5a01e67 21/11/2011. The related fixes mentioned were in within the same week. This all went into release 1.5.1. The requirement

Provide configuration for reverse proxying that will correctly handle the mapping of a URL to a sub-folder (using X-SCRIPT-NAME)

• Analysis of the best solution [1d]

• Implement reverse proxying in tandem with #1653 [2d]

• Document and store the configuration files. [1d]

This requires adding a logic function for emitting an activity stream event, and then editing the logic functions for creating or updating users and making them call the new emit event function. This same emit event function can later be used to emit activity stream events for other types of object as well.

(and add test cases for it)

There's a discrepency in repoze.who versions between the source and package installs:

• repoze.who - package 1.0.18 vs source 1.0.19
• repoze.who-friendlyform - package 1.0b3 vs source 1.0.8

We get a test failure [1] with the 1.0b3 version (from the ubuntu 10.04 python-repoze.who-plugins package). But we've not noticed any problems on s057 instances (br, no, ie etc) which have the package versions of repoze.who.

The reason the package install uses the earlier packaged versions rather than the ones we'd like is that repoze uses all sorts of horrendous import hacks, making it too difficult to put into our 'ckan-conflict' source package.

James suggests we 'do something horrible like dynamically patch repoze on CKAN import'.

[1] ​http://buildbot.okfn.org/builders/builder-ckan/builds/1371/steps/shell/logs/stdio ERROR: ckan.tests.functional.test_user.TestUserController?.test_user_create_unicode

Last week I imported ca.ckan.net datasets into thedatahub.org, but the tags seem to have lost their dashes, underscores and dots.

I want to browse tags, but filtered for a particular group. Currently our tag API doesn't allow for filtering by group.

This is important for improving groups as communities within a site #1521. It would be easy to do this by adding an option to filter by a group. BUT are there any other use cases that would warrant a more complete faceted tag search?

--

BTW I can currently draw a tag cloud for a group - I can get the top tags used in a group like this:

curl ​http://thedatahub.org/api/action/package_search -d '{"q":"groups:country-ca", "facet.field":"tags", "rows":"0"}'

but it only contains the top 20 tags.

On thedatahub.org the favicon doesn't display. i.e. the CKAN logo should appear in the browser's tab.

Original ticket #48

The new group refactor allows for a hierarchy of groups where each group has a type - to be able to implement a tree of groups.

It will need a flag within the group to denote that it is a top level group, to circumvent the need to determine whether the group is a child of a parent node (for a specific use case - a publisher representing the department that only contains publishers).

Suggest 'is_top_level'

• Move into core (keeping as an extension) and default enable
• Do we need tests for controller? (No: already there)
• Documentation (?) - e.g. of basic config options
• Remove lib/stats.py which is still in core
  • dread moved it to extension in ​https://github.com/okfn/ckanext-stats/commit/293b2544737d8ab7610884bd5e999ee37f1ab813 but did not remove from core
• Deprecation notice on ckanext-stats repo and update extension registry

Est: 0.5d.

Questions:

• Why do this?
  • tiny extension with few dependencies - and really nice to have out of the box
  • trial for doing this on larger scale
• Do we keep as extension (even if in core)?
  • Ans: Yes, keep as extension because:
    • Already set up that way
    • Cleaner
    • Easier to disable / enable
• How do we integrate with main them (e.g. have stats link)
  • Ans: not sure (this is part of more general issue of how we update theme for varying changes elsewhere). Best answer is to have some widgetization in theme.
• Hide ratings section (at least until we reinstate ratings #1598)
  • Ans: no, let's not bother (and having ratings there encourages us to do #1598 and/or find out whether people are interested in ratings)

test_authz.py appears to test in great detail some very specific additional authz (related to total site lock-down it seems -- introduced I think for hri project).

I think there are simpler ways to get total site lockdown (use external auth!) and this test is slow and delicate (e.g. depends on specific words in templates). Suggest removing. If we don't remove we should at least refactor tests for access to certain pages to use a proper method of testing (e.g. agreed html comments in each page) rather than being depending on the presence of absence of specific wording.

Theme the DGU form.

Doesn't need to be an exact replica of DGU, but just enough to show it's possible.

Time estimate: 2d

• Fix logs (apache, ckan, harvest): rotate, set suitable levels
• Fix harvesting jobs: supervisord for gather consumer, cron job
• Fix backups

Also it may be worth setting up a test instance ( on s023 ?)

Migrate tables for activity streams

Set-up publicdata.eu to harvest datasets at rs.ckan.net (Serbian community CKAN).

Make sure accented characters are normalized when indexed and when searched for.

Document deployment on tenforce test server.

An API call that returns a user's public activity stream in JSON format.

The SOLR url, user and password defined in the CKAN config file are ignored and the default SOLR url is used.

This causes:

• "0 datasets" displayed on the home page
• Dataset searches result in 0 results and a small message "There was an error while searching". (Nothing about it in the logs)

To reproduce

This bug is only visible if your SOLR instance is not at the default place. To quickly reproduce this problem, setup your machine as a SOLR multicore instance and run: "paster db clean && paster create-test-data && paster serve development.ini". It quits with the error: "solr.core.SolrException?: HTTP code=400, reason=Missing solr core name in path"

Code affected

• Does not affect any released versions (1.5.1 is clean).
• Introduced to master: ​https://github.com/okfn/ckan/commit/627382e6ff575146da221efc50321a41152d8b0f by Ross

We want to make email announcements to users of thedatahub.org.

The most simple way to do this is to get a dump of the email addresses and subscribe/invite them to a mailman list.

Different ways to do it:

• 'paster dump-users-csv' creates a CSV file. Apache serves up the directory using password protection. Is this secure enough?
• CKAN admin pages provides a button to produce the CSV and download it. A config option may be necessary to enable this (disabled by default).

Something more advanced in the future could be:

• CKAN tells mailman about new subscribers using the mailman web interface. Could be brittle.
• Mailman doesn't have an Internet API, but it does have a python interface. So CKAN could tell mailman about new subscribers using a python proxy. There may be one out there, or maybe easy to do. But annoying to have to maintain another process running on machine.
• CKAN sends out the emails itself. A sysadmin is given a form. Users have an opt in/out option in their account. But receiving emails is a sensitive thing so it would need to be right first time, and there are several risks to this approach.

[Child of #1526]

Can't change a dataset from containing time-series resources to one containing individual resources or visa-versa.

Search results page of DGU.

Breakdown:

[X] ensure faceting by group (publisher) works [X] allow faceting by UKLP dataset type [X] allow faceting by INSPIRE

This ticket is about finalizing, then deploying (to ​http://thedatahub.org/), and finally announcing some of recent major new features such as:

• Deploy wordpresser extension for thedatahub (so we can start blogging) - #1533 - WONTFIX
• (Deploy new Resource and Dataset pages (invite comments))
• Get thedatahub geo-ified i.e. ckanext-spatial - #1601 (?)
  • Any additional dev that needs to be done? (e.g. geo in search ...?)
• (Deploy / re-enable ckanext-googleanalytics on thedatahub - #1451 (more stuff #1566))
  • Rename the extension to ckanext-analytics?
• Custom forms (and use of 'dataset types') - Ticket (??)
• (Get QA deployed and running - #1582)

It would be best if this was done by someone *unfamiliar* with the extensions so feedback could be provided on quality of install docs etc etc.

Eurostat datasets mainly.

This CSV resource used to preview but now the format appears to be unsupported: "We are unable to preview this type of resource: x-osdata-csv"

​http://thedatahub.org/dataset/lbhf-spending-2010/resource/9661abbd-2816-4d58-8b20-3cb0eb770c69

This is used as an example by the OpenSpending? team all the time.

From Mark on the ckan-dev list:

In the table of broken links at <​http://thedatahub.org/qa/dataset/broken_resource_links>, the links as printed should be abridged to a fixed length (but link to the full URL). Otherwise the 'reason' column vanishes off the screen. Also a table would probably be better since as things stand the 'reasons' are not lined up

We already have first pass of Data Preview that was done as part of #1357. This is about v2 of this. Two main components:

• A new version of the separate Data Explorer library (now called Recline Data Explorer) - #1446 Est: 10-15d
  • Why:
    • Cleaner, backbone based solution.
    • Provides editing support if wanted.
    • Collaboration with Max Ogden
• Integrate this into CKAN (into resource view page) - see #1630 Est: 1d
• Improvements to Data Viewer / Previewer system in CKAN itself - Est: 2-3d
  • (Update Data Preview-er (View-er) to support images ...)
  • (WONTFIX here - (?) Update Data Preview-er to support geodata - #1151)

Richard Cyganiak: I often see a problem where half the system thinks that I'm logged in, while the other half insists I'm not. Logout then login again may fix it.

David Read: I think I've seen it several times that the "My Account" link is given (suggesting you're logged in) but when you click it it asks you to log in, which means you weren't.

We need an API for changing the authorization (permissions) on a model object.

This was neglected in #1253 because of various changes to authz were looming. But six months later it still hasn't been done.

Suggest this is just moving _add_user_object_role from ckan/lib/base.py to the logic layer, but no doubt it will be more complicated.

Estimate: 3 days

Test failures in master due to tom and rufus not running tests. 0.3d.

Right now is quite difficult to get the most current version of a package via the harvest objects, and the queries involved are prone to errors. A new "current" field will be added which will identify unequivocally the most recent harvested object. Unfortunately this will involve some major changes:

• Model update (See #1726)
• Update queries for getting harvest sources' datasets
• Update ckanext-csw queries for getting the objects
• Update inspire and ckan harvesters
• Document new field

Super #1669

Authz for normal users in publisher profile as per sysadmin

Super #1669

Editor authz for publisher groups need to be implemented and tested.

The HTML and CSS for the activity streams templates needs to be tidied up. Currently the CSS is inlined in the templates themselves, and there is much unnecessary repetition of CSS.

Perhaps put this one in hold until there has been some design discussion about how the activity streams and the user, dataset and group pages with activity streams should look?

For example, after deleting a dataset old activity stream events will still hyperlink to that dataset.

The current harvesting implementation can only have one instance per server.

We could either:

• Accept this
• Change the RabbitMQ code to support multiple instances
• Switch the entire harvesting to use Celery
• Wait for James's redis based live feedback based system

To allow navigation between sections to still work, also potentially rephrase to avoid confusion that you have to save every time to add a tag or resource and can't say add description, tag and resource in one go

Super ticket: #1506

e.g. ​http://thedatahub.org/user The users with Gravatar have their nice user icons, but the majority of users (without Gravatars) have 'broken link' symbols. Same problem wherever users are shown, such as the dataset history pages.

Broken on 1.5.2b - currently on thedatahub.org.

Super ticket #1744

If you edit a dataset, simply add a tag and click on "Save" then you get the message "You have unsaved changed. Make sure to click 'Save Changes' below before leaving this page." inserted whilst the submission takes place. (This is because the click moves focus away from the tag field). This (reasonably common) edge-case could usefully be squashed.

Also note typo: "unsaved changed"

Part of #1698

I have a package with name: "Štatistické prehľady agendy cestovných pasov" and when I try to search for "Štatistické" there are no results.

CKAN setup:

• psql DB with UTF-8 tables

• psql search backend

• ckan version: 1.3.1b

Example:

​http://verejnedata.sk/package/agenda-cestovnch-pasov

try to search for the first word in the package name.