This patch makes it possible to use the api to follow a media stream.

Reacting to real-time updates about a tag requires the use of the "min_tag_id" parameter which is included in the pagination object returned by api methods that paginate. Unfortunately, python-instagram only exposes the "next_url" member of that object, which is not needed if you only want the "latest since last time I polled". Mike described the use of "min_tag_id" here.

This patch fixes that by adding a "pagination_format" parameter to the api methods which defaults to "next_url" so it won't break existing code. Setting the parameter to "dict" will return the full pagination dictionary instead of just its "next_url" member. Setting "as_generator" to "True" will cause the "pagination_format" parameter to be ignored since the generator does not return pagination information anyway.

I hope you find this patch useful and unobtrusive enough to include it. Thank you! Henning

Add more recent Python versions including development branches and nightly build.

The motivation came from reading brettcannon's article : https://snarky.ca/how-to-use-your-project-travis-to-help-test-python-itself/ . Trying to activate the newest Python versions on CI jobs is in most case a win-win situation: if everything works fine, there is nothing to worry about. If an issue is spotted, it is good to know about it to fix it on your side or to open a bug on CPython ( https://bugs.python.org/ ).

Also, if a failures is spotted, you can use the allow_failures option in your matrix build (more information about this in the link above).

Also, nightly and 3.7-dev may be a bit too recent/unstable so I can remove then if need be.

More information about how this PR happened : https://github.com/SylvainDe/CIthon .

More information about the various Python versions available on Travis : https://docs.travis-ci.com/user/languages/python/ .

More information about the Python 3.6 version schedule : https://www.python.org/dev/peps/pep-0494/ . More information about the Python 3.7 version schedule : https://www.python.org/dev/peps/pep-0537/ .

Addresses #127: Add an exclude_format parameter to bind_method to allow us to build a path without the format.

This is necessary because unlike the other endpoints, media/shortcode/{shortcode}.json 404s and should be requested as just `media/shortcode/{shortcode}.

Previous behavior

api = InstagramAPI(access_token=access_tken)
api.media_shortcode('os1NQjxtvF')

raises InstagramClientError: (404) Unable to parse response, not valid JSON.

Corrected behavior

api = InstagramAPI(access_token=access_tken)
api.media_shortcode('os1NQjxtvF')

returns Media: 933107621857404944_231015329 as expected.

As discussed here: https://github.com/Instagram/python-instagram/issues/214

The code i'm trying to merge solve my problem. Not sure if it's the right solution, but it looks correct.

I found this a bit unclear myself. I can't find it mentioned in the provided Python wrapper docs or the official API docs themselves. To me this feels like an important param to understand and clarify since most API results are paginated, and most people will be interested in returning multiple pages of data via the native objects (as opposed to fetching the URL with requests or something).

I was getting True, True, True instead of the values for error code, type and message. I assume we want the important information instead of testing for the keys existence.

In response to the questions related to pagination on "Stack Overflow", etc I have implemented get_users_follows with pagination in the sample web application.

The User ID is @instagram, i.e. 25025320, but can be changed if required?

Output from the test harness is quoted below:

python-instagram cmlh$ python tests.py 
.........................
----------------------------------------------------------------------
Ran 25 tests in 0.324s

OK
python-instagram cmlh$

I have registered [email protected] for the Contributor License Agreement (CLA)

I want to implement infinity scroll of user's photos on my page. I don't want to send to user link to next page and don't want to store this link in session. I don't want to cut next id from the link (it is not durable).

Fixes #27

• user_relationship - determine whether the relationship status between the current user and the target user
• more verbose simplejson import
• except ValueError on simplejson.loads() - This needs to be caught. An example can be

ingr = InstagramAPI(**creds)
ingr.user_relationship(user_id='mikehelmick')

mikehelmick is not a user id so Instagram returns

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
   "http://www.w3.org/TR/html4/loose.dtd">

<html lang="en">
<head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
    <title>404</title>
</head>
<body>
    Couldn't find this page. Sorry.
</body>
</html>

Which would cause L134 of bind.py to fail because

content_obj['meta']['error_type'], content_obj['meta']['error_message'])

meta does not exist cause the response wasn't parsed.

This patch includes:

• User information for likes rather than just a count.
• Caption for each Media.
• API method for extracting recent media for geography_id
• Bug fix for Location.point

It's seems Instagram just return number of comments and do not return the detail of comments in this method that i checked! So it makes problem because there's is no key that named 'data' to append it on 'new_media.comments'

I corrected some grammar mistakes in the code,clean up some whitespace,editing documents,create a specification file and edit some comments.Thank you and I trying my best to do these.

…ent_id

Problem: It was not receiving access_token while exchanging code. Error: You must provide a client_id

Solution: File: oauth2.py Function: exchange_for_access_token +headers = {'Content-type': 'application/x-www-form-urlencoded'} +response, content = http_object.request(url,headers=headers,method="POST", body=data) -response, content = http_object.request(url,method="POST", body=data)