[Bug 766394] Auto-generate share links for KB articles #2006
Conversation
| show_share_link = forms.BooleanField( | ||
| label=_lazy(u'Share Link:'), | ||
| initial=True, | ||
| required=False) |
There was a problem hiding this comment.
You can probably be smarter about this and automatically figure this out based on category, is_archived, etc.
|
Ready for review! |
| class Migration(DataMigration): | ||
|
|
||
| def forwards(self, orm): | ||
| "Write your forwards methods here." |
There was a problem hiding this comment.
You should probably edit, or at least remove this docstring ;)
dean
changed the title
|
@mythmon Fixed the things you pointed out. |
max_length is 24 characters, which should cover the url lengths as they grow for years to come.
They are not generated yet, but there is a base set up for putting them in the database if the box is checked.
Share links are generated via bit.lys API, and the configuration for accessing that API can modified in settings.py.
Added migrations, form element, and now share_links are auto-generated for each new KB article, but you can choose to hide or show them.
Also editted some styling and safeguards for KeyErrors.
We now only generate share links if the article is in a category that should have share links generated for itself.
This is the user we should use for data migrations when having to assign owners/creators to new objects.
This is a data migration for share links. If the template tag sharing a link is found in the most recent revision for a document, we then create a new revision based off the most recent one with content that excludes the share link.
It is now done with a celery task because it's an API call and we want it to be asynchronous. Additionally, I added more descriptive exceptions for any status code responses form bit.ly's api other than a 200.
We now quit the generate_short_url function if either a username or API key is not found. There was also an update to an error message.
There is now a cron job for generating share links where they are missing.
We need to make sure the bitly settings are None by default, so that if the settings are missing it is noticed by generate_short_link in wiki/tasks.py
We now spawn up workers to take care of mass generating share_links in chunked bits, similarly to rebuiding kb articles. Our query for valid docs to generate share links for is now much more efficient and should only grab articles that need the links.
Tests run and pass. Also added custom exceptions for generate_short_url bitly errors.
This includes some comment updates and checking for bitly exceptions instead of generic ones.
|
@dean you rebased, didn't you? That makes it hard to see what you changed. During reviews, you should probably avoid that. |
| is_template=False, | ||
| is_archived=False, | ||
| category__in=settings.IA_DEFAULT_CATEGORIES) | ||
| .exclude(slug='')) |
There was a problem hiding this comment.
How is ~Q(foo=bar) different then .exclude(foo=bar) ? Did you know you can have multiple .exclude or .filter calls?
There was a problem hiding this comment.
I tried the .exclude on both of those ! statements before, without much luck, but it seems to work now. It must have been another issue. Fixed!
Removed Q filters, and subclassed BitlyException for other bitly exceptions.
| category__in=settings.IA_DEFAULT_CATEGORIES) | ||
| .exclude(slug='', | ||
| current_revision=None, | ||
| html__startswith=REDIRECT_HTML)) |
There was a problem hiding this comment.
Does this work as expected? I thought we had a problem where it was only filtering things that had all these properties, instead of any of these properties. If you tested it, then no worries. I was expecting something like .exclude(slug='').exclude(current_revision=None).exclude(...)
|
I pulled this down to test the code:
This is almost ready. You just need to fix the cron job, and then it's ready to go. |
|
Fixed the cron job typo. |
|
Cool. r+ |
This addresses bug 766394, which is linked at the bottom of this comment.
https://bugzilla.mozilla.org/show_bug.cgi?id=766394