crate-archive / crate-site Goto Github PK

From crate.io website, crate.io is planning to provide api to access information about packages, it would be great if the api can provide dependencies. For certain packages requirements.txt is needed(web)and others mention in setup.py, if crate.io can provide both as json list, it would be great.

Currently, the releases tab on package detail pages, sorts releases by version number. While this is a sensible default, chronological sort order would also be useful.

Making the tab's contents into an actual table might help as well, especially if any sort of JS or AJAX driven "click column to sort-by" behavior is easy to obtain OOB. (Though a more static-ish subnav or whatnot that allows toggling between the two sort orders, would doubtless suffice too, if simpler to implement.)

P.S. crate.io is awesome, @dstufft is awesome. 🍰

I think language names shouldn't be translated. Imagine if site currently switched to russian suddenly and someone from brazil wanna switch to portuguese and can't find it coz they don't know russian word "португальский" which means portuguese. Languages should be defined in settings and listed in native languages. I think so.
Example https://github.com/ImaginationForPeople/imaginationforpeople/blob/master/settings.py#L46
P.S languages in english or each language is written in its language will work best.

It might useful to provide a simple announcement system to allow maintainers to announce new directions, backward incompatibilities...

It'd be fantastic if I could click on an author's name on a package's info page and get a page with a list of all the packages that person has authored.

Note, I don't care about the pypi user. I care about the author string.

The detail views (like https://crate.io/packages/django-stripe/) present a helpful pip install ___ command, which is great, but doesn't work if the package isn't hosted or installable. Building out a means of detecting this & conditionally showing the command (perhaps with other information) would be awesome.

Need some Informational Pages

• About The Project
• Tech Stack
• Others?

Currently "Crate" is a big monolithic Django Site. It would be better if instead all the parts that build "Crate the app" were made more generic, and split out into their own repos/projects.

Crate.io's repo would then just become the parts that are specific to Crate.io (settings, urlconf, etc), while all the logic lives inside the generic app.

When using openid django-social-auth set's your password to ! (unusuable). We should allow people to set their password to something and bypass using OpenID etc.

Parse dependencies from requirements.txt and/or setup.py, list them and directly link them to the corresponding crate.io pages.

provide useful analytics to easily learn about package trends

It would be nice to allow people to submit wether a package worked or didn't work for them on a particular Python Version.

Possibly also support for versions of popular frameworks?

Potential issues include needing to wait until we have enough votes (How many votes?) before displaying Works/Doesn't Work Information for a version.

was at a new machine and was looking for the info contained in here:

https://github.com/crate-archive/crate-site/blob/master/crateweb/templates/helpdocs/setting-up-pip.html

Which is not available from the home page via anything that I could see.

Context: Ubuntu 11.10, postgresql.

Here's an example of the stacktrace I can see when running the sync:

Traceback (most recent call last):
File "/home/django/crate/local/lib/python2.7/site-packages/celery/execute/trace.py", line 47, in trace
return cls(states.SUCCESS, retval=fun(_args, *_kwargs))
File "/home/django/crate/local/lib/python2.7/site-packages/celery/app/task/init.py", line 247, in call
return self.run(_args, *_kwargs)
File "/home/django/crate/local/lib/python2.7/site-packages/celery/app/init.py", line 175, in run
return fun(_args, *_kwargs)
File "/home/django/crate/crate-site/crate_project/apps/pypi/tasks.py", line 336, in process_release_data
release.save()
File "/home/django/crate/crate-site/crate_project/apps/packages/models.py", line 177, in save
return super(Release, self).save(_args, *_kwargs)
File "/home/django/crate/src/django/django/db/models/base.py", line 464, in save
self.save_base(using=using, force_insert=force_insert, force_update=force_update)
File "/home/django/crate/src/django/django/db/models/base.py", line 566, in save_base
created=(not record_exists), raw=raw, using=using)
File "/home/django/crate/src/django/django/dispatch/dispatcher.py", line 172, in send
response = receiver(signal=self, sender=sender, **named)
File "/home/django/crate/crate-site/crate_project/apps/packages/models.py", line 367, in version_ordering
Release.objects.filter(pk=release.pk).update(order=i)
File "/home/django/crate/src/django/django/db/models/query.py", line 533, in update
rows = query.get_compiler(self.db).execute_sql(None)
File "/home/django/crate/src/django/django/db/models/sql/compiler.py", line 986, in execute_sql
cursor = super(SQLUpdateCompiler, self).execute_sql(result_type)
File "/home/django/crate/src/django/django/db/models/sql/compiler.py", line 818, in execute_sql
cursor.execute(sql, params)
File "/home/django/crate/src/django/django/db/backends/util.py", line 40, in execute
return self.cursor.execute(sql, params)
File "/home/django/crate/src/django/django/db/backends/postgresql_psycopg2/base.py", line 52, in execute
return self.cursor.execute(query, args)
DatabaseError: deadlock detected
DETAIL: Process 30521 waits for ShareLock on transaction 1220140; blocked by process 30523.
Process 30523 waits for ShareLock on transaction 1220141; blocked by process 30521.
HINT: See server log for query details.

On package pages (crate.io/packages/<name>/), I can't find links to home pages of packages.

On PyPI package pages (pypi.python.org/pypi/<name>), the link is shown next to the label "Home Page:" in the package metadata section at the bottom of the page.

Just like it's the case with Pypi, but add trending packages feed

Allow pointing to a github (or just git?) and possible bitbucket (hg?) repository which will be looked at for requirements files (buildout too? How does buildout work?) and will give notification whenever one of the packages your project uses has been updated.

It would be nice to record what people are searching for, what results they got back, and what the ultimately ended up clicking on.

We could then use this data to determine how well we are doing in our search, and what could be done to improve it.

Independently of #74 add personal category. So, I can sort all my favorite packages by my own category, like: flask packages, django packages, work, beta, and other.

Try and implement good stats screens for package owners with some awesome stat information.

Crate currently has No tests. This needs to be rectified.

Document Crate (and eventually all the apps once those are split out)

Create personal dashboard (like github dashboard), where user can watch all activity of favorite packages (new version and other). And add RSS for personal dashboard. So, I just add packages to "My favorite" and can track all releases by RSS feeds.

subj

Create Trove Specific Pages that can be used to discover packages that are associated with a certain Trove.

The Keywords field is currently unused on Crate. It would be nice to tokenize this field through some method and use it's contents as tags.

Will be greate to see top lists: top 10, top 100 regarding to favorite action. So, I'll can see all popular packages, which people likes.

Allow people to register to receive web hooks. Any action of significance should result in a webhook being fired.

Thoughts:

• Unique ID per Event
• Receiver should confirm that they got it
• Retry a number of times until we give up.

Allow package authors to add Google Analytics to their package packages.

Currently there is a Free Pingdom Account that alerts when Crate.io goes down. This should either be upgraded and the alerts made more comprhensive, or Nagios (or another software) should be employed for the same purpose.

Need to all the strings and formats marked for i18n and create projects on Transifex to allow people to contribute localizations in their own language.

Need to Parse the AWS Download stats from S3 (Stored in a Bucket).

Once we have this data we also need to make it available at pypi.crate.io. This is the last item
needed to be a compliant mirror with PyPI.

I am able to put in my google open id (url to profile page) which forwards me to google to authorize crate.io. This seems to work successfully and the first time I'm able to choose a username and login. Once I log out and try to log back in, however, the exact same process occurs - I can login using openid but crate.io always asks me to choose a username so it seems it's not able to link my gmail openid to my local crate.io account...

People might find it useful to list exact versions of packages instead of just overall packages. This kind of feature also lends itself to a "raw" mode that gives the list in requirements.txt format.

Currently Download Count Sync happens every 10 minutes, it grabs the 150 pacakges that were synced the longest time ago and it gets a list of files for all of those packages and checks with PyPI.

Problems:

• Inefficient
• Not as Friendly to PyPI as it could be
• Can generate up to 2k Celery Tasks

Ideas:

• There is a stats interface that records daily stats (and monthly) for PyPI
  • These are integrated in multiple times a days
  • How to switch over, these are daily stats not totals?

I maintain a list of python packages that I consider part of my 'standard toolbox'; libraries I consider 'best of breed' for a given task.

It would be great if I could manage this list on crate.io. Something as simple a a toggleable star per package and a page which lists my favourite packages would be useful.
Given this basic functionality, further features could be trivial to implement:

• custom rss feed for favourite updates
• one-click download of all favourite packages in a .zip

Would anyone else be interested in this? Has anything similar been proposed? I'm willing to have a crack at implementing something if there is any interest.

There is no pagination on favorites, this can break.

Hi there, nice work.

I just noticed that after I implemented PEP 386, the latest release (0.1.post18) of fexpect, my package for scripting remote prompts in fabric seems to be ordered as older than some previous releases on crate.io:

https://crate.io/packages/fexpect/

Whereas pypi correctly seems to recognize it as the latest release:

http://pypi.python.org/pypi/fexpect

Would be great if this were consistent.

Allow users to Watch particular packages. From their watched packages they can get an Activity Stream of changes (Available in RSS format)

Like http://djangopackages.com/ will be greate to have category.
P.S. And many other cool features from http://djangopackages.com/

I would love to query crate.io for package downloads and then display them on: pythonpackages.com. E.g. http://pythonpackages.com/info/django could contain a link to: http://packages.crate.io/a/f/9/1/af9118c4e8a063deb0b8cda901fcff2b805e7cf496c93fd43507163f3cde156b/Django-1.3.1.tar.gz. It looks like I could construct the link if could get a hold of the sha256 hash. Any plans to expose this info via an API? Great job with the site!

Any chance you will support uploads? pythonpackages.com would like to support releases-to-multi-indexes and would love to test with crate.io,

It would be nice for crate to provide a Mirror of Packages not hosted on PyPI.

Issues:

• Legality, Not all packages have a license that allows redistribution
  • Can Trove Classifiers Be Used to Determine the License?
• No Change Log; Package Author could Change Release and We won't know.
• Ill Will; Package Authors Arn't hosting on PyPI for a reason (Why?) Would doing this cause friction?

This can already be done using OpenID, but we need to take it further and once a PyPI openid is saved on Crate make the UX and the machinery to say that user X on Crate is equivalent to user Y on PyPI