coffeered / python-flickr-api Goto Github PK

flickr.stats.getPhotoDomains should be flickr.stats.getPhotosetDomains on line 
4776

There's a try/except block on line 3110 of objects.py, in the search method, 
that tries to access both a key and an attribute. The except clause should 
target both, like this:

       except (KeyError, AttributeError) : pass

That's bit me a couple of times. Thanks.

What steps will reproduce the problem?

This code:

  user = flickr_api.Person.getByUrl('http://www.flickr.com/photos/earlyadopter/')
  photos,resp_infos = user.getPublicPhotos()

returns this error:

  ValueError: too many values to unpack

I do have lots of images, and it would be nice to return at least some 
predefined number of them, or have in user.getPublicPhotos() pagination 
mechanism.

What version of the product are you using? On what operating system?

object.py file has "Date   : 05/08/2011" in line 21.

Here's my code: http://dpaste.org/GOjCe/

Here's the stacktrace: http://dpaste.org/NtQKN/

I discovered very quickly that the error was happening because I was specifying 
photoset_id instead of id. Changing it to "id" fixed the problem and gave me 
expected results.

I think there should be error checking somewhere in this recursion loop to 
check that the attribute is set before trying to retrieve it.

Your meta programming approach is really cool and helpful, but it seems to be 
very prone to recursion like this. I think error catching is pretty important 
when the code has a lot of magic.

on line 3726 you have a typo

says 
        photoset_ids = arsg["photoset_ids"]

should be
        photoset_ids = args["photoset_ids"]

Hey,
I have two concerns:
The first about the streaming limit, when I'm streaming photos and comments, it 
doesn't stream all of them like if there is a limit, Can I make it unlimited?

The second question, is there a way to stream users without specifying their 
email or username, just stream random users?

Thank you very much 
Khaled

Line 2800 of objects.py includes the typo "recentlUpdated" which should be  
"recentlyUpdated". :)

What steps will reproduce the problem?
1. Try to import the library on Python 2.5
2. It fails.
3. Try to fix a few thing (like importing with_statement from __future__, 
replacing json by simplejson)
4. It still fails.

What is the expected output? What do you see instead?
If the documentation doesn't state that it requires 2.6+, I would have expected 
it to work on 2.5, since 2.5 is still supported by the major distributions.

What version of the product are you using? On what operating system?
Python 2.5 on Ubuntu 8.04 LTS

Please provide any additional information below.
1. The library does work on Python 2.6
2. The zip file from the download area does not include the instruction to 
create "flickr_keys.py", which is necessary for this library

It would be great if your library had some way to easily hook into the DJango 
cache library (or some other form of caching).

The other flickrapi has this function. While your library kicks its ass in most 
regards, it would be great have support for this: 
http://stuvel.eu/media/flickrapi-docs/documentation/#caching-of-flickr-api-calls

Is that something you'd be open to having added?

It took me a while to trace this down, but currently FlickrObject.__getattr__ 
checks if the object has been "loaded" yet or not. If it hasn't it forces a 
getInfo() API kill.

When I call Photoset.getPhotos(extras=['url_m'])

Then loop through all photos and try to access url_m (which is already has 
access to!!), it forces a getInfo call on every single photo. This REALLY slows 
down my app.

I'm sure there is some structural reason why having that getInfo() call made 
things easier, but it's an extra HTTP request that the library user isn't 
expecting.

Can't offer a solution because I don't know why it's necessary to have it there 
in the first place.

But the real method is "set_auth_handler". Docs need to be updated.

In multiple places, you call args.pop("user_id").id

If someone passes in a user_id, it won't have a .id attribute. Either you meant 
to have them pass in a user obvject or need to use the "user_id" directly.

Insteances of this issue

* Photo.Person.addPerson
* Photo.search

There might be others. The code is re-written a bunch of different ways across 
the system. Should probably have a helper function that checks if the user 
passed in a user and adds user_id to the args. That way it's only in the code 
once.

What steps will reproduce the problem?
1. Install Flickr_api
2. Get the keys from Flickr and set them up.
3. follow the wiki and reach upto retrieving a user

What is the expected output? What do you see instead?


What version of the product are you using? On what operating system?
Later version on website

Please provide any additional information below.

I am a newbie. So this might be due my lack of info on what else to do

When I call any API methods, I get this flickr API error:

100 : Invalid API Key (Key has invalid format)

Which is due to way API_KEY was being passed in as a default argument.

API_KEY had a value, but the local api_key wasn't receiving that value and just 
had a blank string as its value.

My guess is that you might be developing with a different version of python 
than I am. Maybe? I'm using 2.7. Are you using python 3?

Here is the commit I made to fix it for my setup: 
https://github.com/auzigog/python-flickr-api/commit/e99d3374040fa54b9ff1f5246c23
d48e012a6df1

I'd highly encourage you to move this project over to GitHub. It's a much more 
popular 

Honestly, I'd be willing to write code to help improve this project and fix 
bugs, but only if it's on GitHub. It's so much easier to collaborate, fork 
code, make my own commits, and pass them back to you.

Plus it gets way more exposure for your code. Most developers are using GitHub 
as their primary open source repo host these days.

Would you be willing to make the switch over?

http://code.google.com/p/python-flickr-api/source/browse/trunk/setup.py

The version number currently says "0.1" but it appears as though you've 
released a 1.0 version.

I find it helpful to store my version number in app/__init__.py and have 
setup.py import from there. Then I only have to change it in one place.

What steps will reproduce the problem?
1. pip install 
svn+http://python-flickr-api.googlecode.com/svn/trunk/#egg=flickr-api
2. import flickr_api
3. Run any of the function calls

What is the expected output? What do you see instead?

The function returns normally.

Instead, It errors saying flickr_keys module wasn't available. I checked the 
install directory and contrary to what INSTALL.txt says, there's no place for 
me to put my key.

The solution for this shouldn't be to add the flickr_keys module. It should be 
to require you to specify your key and secret when instantiating a new module.

Calls should look like this: 

import flickr_api
api = flickr_api.FlickrAPI(my_key, my_secret)
photos = api.Photos.search('cats')

That's much cleaner than making me put config details inside your application 
install.

(I'll even make a patch for this change if you move it over to github :D )

Looks like you've introduced a SyntaxError in the latest revision. Line 1151 in 
objects.py includes:

        return = test.login(token)

which should not have the assignment operator.

Calling this

{{{

        portfolio = Portfolio(settings.FLICKR_UESR_ID)
        photosets = portfolio.getPhotosets()
        context_data['photos'] = portfolio.getPhotos(photosets[1])
}}}

which calls this:

{{{

class Portfolio(object):

    def __init__(self, user_id):
        self.user = flickr_api.Person(id=user_id)

    def getPhotosets(self):
        photosets = self.user.getPhotosets()
        return photosets

    def getPhotos(self, photoset):
        extras = ['url_sq', 'url_s', 'url_t', 'url_m', 'url_o']
#        photosA = photoset.getPhotos(extras=extras)
#
#        for photo in photosA:
#            photo.__dict__.update({'loaded':True})

        args = dict()
        args['extras'] = ','.join(extras)
        photos = call_api(method="flickr.photosets.getPhotos", photoset_id=photoset.id, **args)
#        photos = [p.getPhotoFile(size_label='Square') for p in photos_all]
        return photos['photoset']['photo']
}}}

Which gives me this error: http://d.pr/WbFM

At the very least, there should be better error handling here to tell me what 
I'm doing wrong quickly. I'm going to debug it more to see if I have a more 
specific suggestion.

Thanks!

What steps will reproduce the problem?

There is following code in objectrs.py in line 1173:

  r = method_call.call_api(method = "flickr.urls.lookupPerson", url = url)

which returns:

  "flickr_api.base.FlickrAPIError: 112 : Method "flickr.urls.lookupPerson" not found"

Flickr's corresponding method is currently called lookupUser, not lookupPerson.

What version of the product are you using? On what operating system?

object.py has "Date   : 05/08/2011" in line 21.