shotgunsoftware / python-api Goto Github PK

View Code? Open in Web Editor NEW

304.0 304.0 198.0 69.17 MB

A Python-based API for accessing Flow Production Tracking.

Home Page: https://developer.shotgridsoftware.com/python-api

License: Other

Python 99.93% Batchfile 0.07%

python-api's People

Contributors

Stargazers

Watchers

Forkers

schworer nvizible amorton mbeal beat1823 bdeluca guniorobot hughmacdonald brandonvfx daytongirl82 samuderapase miketuffy secure96 dahiro jlepis hzhang2 westernx aag oopoopoop jperring boldstelvis robertkraig cmihaly jlnitro purestvfx ajng npglambert aj-dneg maxgyver pscadding krets host3d dushyantk trizoy chloechan tweakoz satishgoda manu3dmatchmovers dentas hybridhalo atomicfiction claudiuvadean pixomondo czg200809 richardssam bbqstudy dwarflabs derrickmeyer mikepkes titoudev johnahartman cgcloud lavenda arork flyingai eestrada moonbot ridkuma pboucher dan-roboticarm mor-vfx nuternativ alexpnyy bravo1010 renevolution iyu-zz zaltu cuthb3rt rodeofx mikeweissmv eoghancunneen silarr squeezestudioanimation mike1158 iverb cundesi utsdab singhrz-zz cmwaura michael-delaporte babygiantstudio loonghao cpichard dzjack jfbalanc nkpatro gst craftyapes laikastudios ykdvd kushquad chrisheckey olessyaz satishgodapearl ogranada rane-hs col-one accidentallythecable nixworks apetrynet

python-api's Issues

Should Uploading to S3 automatically retry when receiving a `HTTP Error 503: Slow Down`?

Hey,

Just hit this myself when trying to create a Shotgun PublishedFile and include a thumbnail image with it.

It appears that the sg-media-usor-01 got a bit overloaded with traffic and S3 told us (python api) to slow down.

Because I was naively using the create method to also handle my "image" field, I now get an error after the entity has been created, however I don't know what id the created entity has.

My workaround has been to copy out the upload logic from the create method and wrap it in a retry loop with some sleeping, but it is not ideal.

Is this something that I could propose and get merged in if I wrote a proper implementation for it?

Thanks
-Alex

IronPython unicode issue: ('unknown', u'\xf1', 4, 5, '')

IronPython 2.7.4 has an issue on value.decode from unicode in many unit tests:

File "C:\dev\wspy\sandbox_shared\src\shotgun\python_api3\tests\test_api.py", line 24, in setUp
self.sg.update('Note', self.note['id'], {'content':u'La Pe\xf1a'})
.....
File "C:\dev\wspy\sandbox_shared\src\shotgun\python_api3\shotgun_api3\shotgun.py", line 1823, in _outbound_visitor
return value.decode("utf-8")
UnicodeDecodeError: ('unknown', u'\xf1', 4, 5, '')

see here for list of unit tests:
https://dl.dropboxusercontent.com/u/4995745/201411/shotgun-IronPython-api-unittests_in_test_api_20141118.html

followers() causes an internal server error if the entity id does not exist

If you give Shotgun.followers() an id that doesn't exist it causes the server to have a 500 error. This happens in Shotgun v6.0.3 (build 04eae45) with python-api master (e92bece).

In [3]: sg.followers({'type': 'Shot', 'id': 1167})
---------------------------------------------------------------------------
ProtocolError                             Traceback (most recent call last)
<ipython-input-3-b5726fc21ace> in <module>()
----> 1 sg.followers({'type': 'Shot', 'id': 1167})

/home/ashworth/.virtualenvs/tmp-3463e352d8613412/lib/python2.7/site-packages/shotgun_api3/shotgun.py in followers(self, entity)
   1054         )
   1055 
-> 1056         return self._call_rpc('followers', params)
   1057 
   1058     def schema_entity_read(self, project_entity=None):

/home/ashworth/.virtualenvs/tmp-3463e352d8613412/lib/python2.7/site-packages/shotgun_api3/shotgun.py in _call_rpc(self, method, params, include_auth_params, first)
   1683         LOG.debug("Completed rpc call to %s" % (method))
   1684         try:
-> 1685             self._parse_http_status(http_status)
   1686         except ProtocolError, e:
   1687             e.headers = resp_headers

/home/ashworth/.virtualenvs/tmp-3463e352d8613412/lib/python2.7/site-packages/shotgun_api3/shotgun.py in _parse_http_status(self, status)
   1853                                 error_code,
   1854                                 errmsg,
-> 1855                                 headers)
   1856 
   1857         return

ProtocolError: <ProtocolError for brandon.shotgunstudio.com: 500 Internal Server Error>

Update function returns datetimes as string

Tested on python api v3.2.4, but haven't seen anything in 3.2.6 that would fix this.
When running the update function on a datetime field the return value is a string, rather than the datetime.datetime expected.
This string is also not compatible with plugging back in to further updates.

Example below;

import datetime
from tank_vendor.shotgun_api3.lib.sgtimezone import UTC
from tank.util import shotgun

utc_now = datetime.datetime.now(UTC())
sg = shotgun.create_sg_connection()

# update date field with utc date, returns as string
sg.update('PublishedFile', 433118, {'sg_accessed_at': utc_now})
>>> {'sg_accessed_at': '2020-11-27 09:52:29 UTC', 'type': 'PublishedFile', 'id': 433118}

# for comparison, find_one returns as datetime with tzinfo
sg.find_one('PublishedFile', [['id', 'is', 433118]], ['sg_accessed_at'])
>>> {'sg_accessed_at': datetime.datetime(2020, 11, 27, 9, 52, 29, tzinfo=<tank_vendor.shotgun_api3.lib.sgtimezone.LocalTimezone object at 0x7fdff227b8d0>), 'type': 'PublishedFile', 'id': 433118}

# plugging that string back into an update throws an error
sg.update('PublishedFile', 433118, {'sg_accessed_at': '2020-11-27 09:52:29 UTC'})
>>> Traceback (most recent call last):
  File "<input>", line 1, in <module>
  File "/TCP/home/spencere/toolkit_dev/install/core/python/tank_vendor/shotgun_api3/shotgun.py", line 1435, in update
    record = self._call_rpc("update", params)
  File "/TCP/home/spencere/toolkit_dev/install/core/python/tank_vendor/shotgun_api3/shotgun.py", line 3314, in _call_rpc
    self._response_errors(response)
  File "/TCP/home/spencere/toolkit_dev/install/core/python/tank_vendor/shotgun_api3/shotgun.py", line 3619, in _response_errors
    raise Fault(sg_response.get("message", "Unknown Error"))
Fault: Invalid date time format: 2020-11-27 09:52:29 UTC. Correct format is 2011-01-21T13:26:09Z (UTC) or 2011-01-21T13:26:09-07:00 (UTC Offset)

# note, the recommended format also throws an error
sg.update('PublishedFile', 433118, {'sg_accessed_at': '2020-11-27T09:52:29Z (UTC)'})
>>> Traceback (most recent call last):
  File "<input>", line 1, in <module>
  File "/TCP/home/spencere/toolkit_dev/install/core/python/tank_vendor/shotgun_api3/shotgun.py", line 1435, in update
    record = self._call_rpc("update", params)
  File "/TCP/home/spencere/toolkit_dev/install/core/python/tank_vendor/shotgun_api3/shotgun.py", line 3314, in _call_rpc
    self._response_errors(response)
  File "/TCP/home/spencere/toolkit_dev/install/core/python/tank_vendor/shotgun_api3/shotgun.py", line 3619, in _response_errors
    raise Fault(sg_response.get("message", "Unknown Error"))
Fault: Invalid date time format: 2020-11-27T09:52:29Z (UTC). Correct format is 2011-01-21T13:26:09Z (UTC) or 2011-01-21T13:26:09-07:00 (UTC Offset)

Add package to pypi

Hey, would it be possible to get this package into pypi?

We currently download all our other python dependencies from pypi, but with this package, we have had to fork it and upload it to our own internal pypi server. Thanks

Add Python 3 support

http://www.vfxplatform.com/

29th April 2017 - CY2018 Draft published with significant upgrades to gcc and glibc. This also comes with notice of our intention to move to Python 3 for CY2019 in advance of Python 2 support ending in 2020. We are currently soliciting feedback on this Draft so please either send to [email protected] or share on vfx-platform-discuss. We expect to publish CY2018 Final around SIGGRAPH in July.

Mockgun does not support 'any' filter_operator

When using an 'any' filter_operator and mocking, I get an :

Traceback (most recent call last):
  File "/home/marie/rdoenv/python_libs/pure/rodeo/tests/integration_tests/test_ilm_ocio_profile.py", line 121, in test_shotFilter
    for config in configs.shotConfigMap.iterkeys():
  File "/home/marie/rdoenv/python_libs/pure/rodeo/color/ilm_ocio_profile.py", line 122, in shotConfigMap
    for sgShot in self.sgShots:
  File "/home/marie/rdoenv/python_libs/pure/rodeo/color/ilm_ocio_profile.py", line 63, in sgShots
    self._sgShots = self.sg.find('Shot', queryFilters, ['code', 'sg_cdl_path', 'sg_client_name'])
  File "/home/marie/rdoenv/python_libs/pure/shotgun_api3/lib/mockgun.py", line 339, in find
    elif f[1] == "in" and not isinstance(f[2], list):
KeyError: 1

The same code is working when not mocking.

Code sample :

queryFilters = [['project.Project.code', 'is', self.projectCode],
                ['sg_status_list', 'not_in', ['bid', 'awd', 'omt', 'void', 'hdn']]
                ]

if self.shotFilters:
    # Keep only the shots that are corresponding to the filters
    shotQueryFilters = []
    for shotFilter in self.shotFilters:
        shotQueryFilters.append(['code', 'starts_with', shotFilter])
    queryFilters.append({
        "filter_operator": "any",
        "filters": shotQueryFilters
    })

self._sgShots = self.sg.find('Shot', queryFilters, ['code', 'sg_cdl_path', 'sg_client_name'])

Thanks !

Unlinking Asset in Shot's Related Assets tab

I am new to this api and am wondering if there is a way to unlink an asset that was added to the Related Assets tab on a Shot. I tried

from shotgun_api3.shotgun import Shotgun
Shotgun.update("Shot", 2531, {"assets":[{"type":"Asset", "id":3425}]},
multi_entity_update_modes={"assets":"remove"})

but this didn't seem to work. Any suggestions? Thank you :)

SSO Support

A facility like oAuth is a necessity for a completely secure ecosystem. For those serving their Action Menu Items on a Web server, oAuth would offer a way to seamlessly authenticate the session with shotgun.

Currently, after clicking an Action Menu Item, users are prompted to login separately on the Web server. There are two ways to approach this:

basic auth
auth with user's shotgun credentials

Option 1 Requires users to login with a separate username and password from their shotgun account. It uses a static API key for subsequent API requests.
Option 2 Requires the storage of the user's plaintext shotgun credentials in a cookie or on the server. It uses these credentials to authenticate the subsequent API requests

There needs to be a way to checkout a temporary session token for API requests. This way users can leverage the shotgun API with their own identity and with the security of a single login facility with the guaranteed security already built into shotgun's authentication.

Connection Management/Timeouts

What's the recommended way to handle connection issues, i.e. slow or broken connections?

Is it safe just to use signals to cancel out of stalled/slow calls to the API?

Installation docs in README file

pip install git+git://github.com/shotgunsoftware/[email protected]

should be replaced by

pip install git+https://github.com/shotgunsoftware/[email protected]

Shotgun(connect=False) no longer working as expected

python-api/shotgun_api3/shotgun.py

Lines 590 to 594 in 41c5369

    
           if connect: 
        
               self.server_caps 
        
           # Check for api_max_entities_per_page in the server info and change the record per page value if it is supplied. 
        
           self.config.records_per_page = self.server_info.get('api_max_entities_per_page') or self.config.records_per_page

Previously, e.g. in v3.0.22, instantiating a Shotgun object did NOT contact the server. Determining the value of records_per_page should be deferred when connect=False.

Python 3 support

The API doesn't seem to be compatible with Python 3

Unable to specify correct filter to query and return all Project, Episode, and Shot fields together

Using the following variables:

project_name = 'SuperStickman'
episode_name = 'Stic-101'

I want to extract ALL the Project, Episode and Shot fields together.

I tried these two different ways, but they don't display Project, Episode and Shot fields.

sg_shots = sg.find("Shot", [['project.Project.name', 'is', project_name],['episode.Episode.name', 'is', episode_name]], ['code', 'sg_sequence.Sequence.sg_status_list'])

sg_shots = sg.find("Shot", [['project.Project.name', 'is', project_name]], ['code', 'sg_sequence.Sequence.sg_status_list'])

How do I return a query result that returns ALL the Project, Episode and Shot fields together?

Thank you in advance.

Retrieve fields inside arrays of entities with sg.find_one()

I've been looking to find an answer to this for quite a while but couldn't find anything in the docs…

I try to retrieve a version together with its notes, but I need to have more of the notes than just the standard fields that are retrieved.
Here's the relevant part:

def getVersion(id):
    sg = Shotgun("supplied","supplied","supplied")

    response = sg.find_one("Version", [["id", "is", int(id)]], ['id', 'notes', 'notes.Note.content'])
    
    return response

I thought notes.Note.content would be the right way to do it but since it is a collection, it does not seem to work. The response does not give me the notes content fields but just this:

{
    "notes": [
        {
            "type": "Note",
            "id": 25797,
            "name": "Example User on 008_010_v103 and 008_010 - this is a test"
        }
    ],
    "type": "Version",
    "id": 4068
}

Is there any way to get specified fields of the notes list or do I need to perform another query to get these?

Using find() and ordering by playlists.PlaylistVersionConnection.sg_sort_order returns CRUD ERROR

The following code:

shotgun.find(
    "Version",
    order=[{"field_name": "playlists.PlaylistVersionConnection.sg_sort_order", "direction": "asc"}],
)

Results in the following error:

shotgun_api3.shotgun.Fault: API read() CRUD ERROR #10: Read failed for entity type [Version]

This is probably an issue on the server side rather than the python api client, but I would expect to be able to sort by the Playlists > Sort Order field though the API in the same way you can through the UI.

IronPython issue: unknown encoding: idna

Looks like the idna encoding needs to be converted again to utf-8 for IronPython:

h = Http(".cache")
thumb_resp, content = h.request(version_with_thumbnail.get('image'), "GET")

test_humanuser_upload_thumbnail_for_version
Error
Traceback (most recent call last):
File "C:\dev\wspy\sandbox_shared\src\shotgun\python_api3\tests\test_api.py", line 1551, in test_humanuser_upload_thumbnail_for_version
thumb_resp, content = h.request(version_with_thumbnail.get('image'), "GET")
LookupError: unknown encoding: idna

see: https://dl.dropboxusercontent.com/u/4995745/201411/shotgun-IronPython-api-unittests_in_test_api_20141118.html

Sharing a playlist via API produces unexpected behavior

When I share a playlist (example below) via shotugun API, associated note submissions hang on submitting...

def sharePlaylist(sg, playlistID, clientID):
    filters = [ ['playlist', 'is', {'type': 'Playlist', 'id': playlistID}] ]
    fields  = ['playlist', 'user']
    share = sg.find_one('PlaylistShare', filters, fields)
    if not share:
        data = {'playlist': {'type': 'Playlist', 'id': playlistID}, 'user': {'id': clientID, 'type': 'ClientUser'}}
        fields  = ['playlist', 'user']
        return sg.create('PlaylistShare', data, fields)
    else:
        return share

@pboucher explained this is not supported in the API. Can we get support for this? Perhaps I'm simply missing something in my code above?

Accessing the fields data with the API

Hello,

Is there a way to retrieve the items of a List type Field with the API?

Thanks

ShotgunError exception too broad?

I generally base exceptions on StandardError because it's usually bad policy to catch Exception because it's too broad. PyCharm complains about it so I know it's not just my own issue.

Would it harm downstream exc handlers to base off of StandardError or something even lower in the exc mro?

Any handlers which are already catching Exception won't be affected. StopIteration and Warming excs don't seem to be necessary for the ShotgunError exc hier.

Bundle httplib2 should be adapted for the integration you've made of it

Hi,

noticed with this stack trace :

 File "../python-2.7/shotgun_api3/sg_26.py", line 5, in <module>
    from .lib.httplib2 import Http, ProxyInfo, socks, SSLHandshakeError
 File "../python-2.7/shotgun_api3/lib/httplib2/__init__.py", line 59, in <module>
    from httplib2 import socks
  File "../python-2.7/httplib2/__init__.py", line 915, in <module>
    class HTTPSConnectionWithTimeout(httplib.HTTPSConnection):
AttributeError: 'module' object has no attribute 'HTTPSConnection'

As you can see :

you have integrated httplib2 into shotgun_api3.lib sub-package.

you correctly use it with it : from .lib.httplib2 import Http, ProxyInfo, socks, SSLHandshakeError

but then you should also have adapted the import at line 59 of shotgun_api3/lib/httplib2/__init__.py to from shotgun_api3.lib.httplib2 import socks or the better from . import socks.

cause like that you actually use a part of the integrated httplib2 you made a copy and a part of the normal httplib2 which can comes in any environment where we need to use shotgun_api3.. which may or may not be compatible..

I don't know if there are no other such import to adapt in the different things you have integrated into/under shotgun_api3/lib/*

workaround to install python-api on Fedora Silverblue 37

On Fedora Silverblue 37 I tried:
sudo rpm-ostree install python3-pip
but it failed.

Here's a workaround that works for Silverblue 37:

git clone https://github.com/shotgunsoftware/python-api.git sgpythonapi
cd sgpythonapi/
sudo python3 setup.py install

I'm able to create sg and do sg.find()/sg.find_one() after that.

Unable to Run PyInstaller EXE with shotgun_api 3.3.2

As I'm in the midst of troubleshooting SSL error for select users, I saw there is update to httplib2 and decided to update shotgun_api3.

While it can run fine when executing through a Python interpreter, my studio pipeline requires the tool to be compiled as a standalone exe for artists usage. Here's the error message when trying to run the EXE after upgrading from 3.3.0 to 3.3.2.

Currently my workaround is to pip install pyparsing==2.4.7 and modify both site-packages/shotgun_api3/lib/httplib2/python2/auth.py and site-packages/shotgun_api3/lib/httplib2/python3/auth.py

# Existing code
# from ... import pyparsing as pp
# My fix after pip install pyparsing
import pyparsing as pp

After the changes above, I can build normally and run the EXE without issue.

My dev setup is the following:

Python 3.8.10
pyinstaller 4.5.1

I'm excluding the other libraries since they don't use pyparsing

P/S: This is more for other users, there is no need for me to modify the shotgun_api3 package for prior version from 3.3.0 and older when building an EXE using pyinstaller.

ShotGrid Legacy Login post-Autodesk ID - Personal Access Token Clarification

According to this migration support doc re: personal access tokens, for continued functionality of my company's scripts, we need to bind a Personal Access Token from our Autodesk ID to our ShotGrid account.

It's unclear, however, what combinations of credentials we should use with the Shotgun API for authentication going forward, after binding this PAT to our ShotGrid account. Additionally, I've yet to find an unambiguous way to prove that binding PATs to our shotgrid accounts is all we need to do for continued functionality.

After binding a PAT to my ShotGrid account, I wrote the script at the end of this issue (following the demo code in the video here) to test all possible combinations of credentials - only the following combinations worked:

server=PROXY, login=SG_USER, password=SG_PASS, tfa=SG_2FA
server=SG_SERVER, login=SG_USER, password=SG_PASS, tfa=SG_2FA

This indicates that only shotgrid credentials - and not those associated with my autodesk ID - work for authentication, even after PAT binding. It is implied in the documentation - though not made sufficiently explicit - that this is the intended behavior, and that by binding a PAT to our ShotGrid account all scripts should be able to work with no changes whatsoever. But, again - I've found no way to definitively prove this. If I'm wrong, several hundred of our employees may find themselves unable to continue with work after Friday.

If I'm right - if all we need to do is bind PATs to each of our employees ShotGrid accounts - there's the question of implementation. Do all of our employees need to manually generate PATs for each of their ShotGrid accounts, or is there some way to do this automatically for everyone in our domain?

Here's the script I used to check things:

from shotgun_api3.shotgun import Shotgun
import json
import itertools as it

params = json.load(open('params.json'))

servers = [
    'PROXY',
    'SG_SERVER'
]
logins = [
    'AD_USER',
    'AD_EMAIL',
    'SG_USER',
]
passwords = [
    'AD_PASS',
    'SG_PASS',
]

twofa = [
    'SG_2FA',
    'AD_2FA'
]

if __name__ == "__main__":
    import sys

    filters = [['sg_status_list', 'is', 'act']]
    fields = ['id']
    tfa_cache = {
        'SG_2FA': 'NUMS',
        'AD_2FA': 'MORE_NUMS'
    }
    for server, login, password, tfa in it.product(servers, logins, passwords, twofa):
        server_value = params[server].encode('utf-8')
        login_value = params[login].encode('utf-8')
        pw_value = params[password].encode('utf-8')
        tfa_value = tfa_cache[tfa].encode('utf-8')
        print("trying:")
        print('server="{}", login="{}", password="{}", tfa="{}"'.format(server, login, password, tfa))
        try:
            sg = Shotgun(
                server_value,
                login=login_value,
                password=pw_value,
                auth_token=tfa_value
            )
            n_users = len(sg.find('HumanUser', filters, fields))
            print('WORKED!')
        except Exception as e:
            print("FAILED... Error: {}".format(e))

Relative imports and python 2.4

I can't use the newer versions of shotgun.py because of the relative imports. I'm using python 2.4, which I know doesn't support them, but Shotgun is supposed to support that version. I don't think python gets to check the sys.version_info and then decide which import to try as the author intended (line 51ish); I seem to be crashing immediately because 2.4 considers it invalid syntax.

mimetypes initialization fails on Windows because of TypeError

We just encounter the issue on Windows that Python 2.7 threw a TypeError on mimetypes initialization.

C:\Python27>python
Python 2.7.8 (default, Jun 30 2014, 16:03:49) [MSC v.1500 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> exit()

C:\Python27>python -c "import mimetypes; mimetypes.init()"
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\Python27\lib\mimetypes.py", line 351, in init
    db.read_windows_registry()
  File "C:\Python27\lib\mimetypes.py", line 254, in read_windows_registry
    with _winreg.OpenKey(hkcr, subkeyname) as subkey:
TypeError: must be string without null bytes or None, not str

The suggested workaround is to catch the TypeError when the Shotgun API first uses mimetypes

import mimetypes    # used for attachment upload
try:
    mimetypes.add_type('video/webm','.webm') # try adding to test for unicode error
except (UnicodeDecodeError, TypeError):
    # Ticket #25579 python bug on windows with unicode
    # Ticket #23371 mimetypes initialization fails on Windows because of TypeError (http://bugs.python.org/issue23371)
    # Use patched version of mimetypes
    from .lib import mimetypes as mimetypes

See also: mimetypes initialization fails on Windows because of TypeError (http://bugs.python.org/issue23371)

Can't build Pyinstaller executable with python 2.7 on shotgun_api3 3.2.2

Since Python 3 support was added, I can't build Pyinstaller executables with shotgun import in them. The build completes, but I get the following error:

File "dependecies\shotgun_api3\lib\six.py", line 82, in _import_module __import__(name) ImportError: No module named xmlrpclib

It was working with version <= 3.0.41

Installation does not work for Python 3.7

Hi,

Even though Python 3.7 is specified as a requirement, the install does not work with it:

> D:\python-api-3.2.3\python-api-3.2.3>python setup.py install --user
running install
running bdist_egg
running egg_info
creating shotgun_api3.egg-info
writing shotgun_api3.egg-info\PKG-INFO
writing dependency_links to shotgun_api3.egg-info\dependency_links.txt
writing top-level names to shotgun_api3.egg-info\top_level.txt
writing manifest file 'shotgun_api3.egg-info\SOURCES.txt'
reading manifest file 'shotgun_api3.egg-info\SOURCES.txt'
writing manifest file 'shotgun_api3.egg-info\SOURCES.txt'
installing library code to build\bdist.win-amd64\egg
running install_lib
running build_py
creating build
creating build\lib
creating build\lib\shotgun_api3
copying shotgun_api3\shotgun.py -> build\lib\shotgun_api3
copying shotgun_api3\__init__.py -> build\lib\shotgun_api3
creating build\lib\shotgun_api3\lib
copying shotgun_api3\lib\mimetypes.py -> build\lib\shotgun_api3\lib
copying shotgun_api3\lib\sgsix.py -> build\lib\shotgun_api3\lib
copying shotgun_api3\lib\sgtimezone.py -> build\lib\shotgun_api3\lib
copying shotgun_api3\lib\six.py -> build\lib\shotgun_api3\lib
copying shotgun_api3\lib\xmlrpclib.py -> build\lib\shotgun_api3\lib
copying shotgun_api3\lib\__init__.py -> build\lib\shotgun_api3\lib
creating build\lib\shotgun_api3\lib\httplib2
copying shotgun_api3\lib\httplib2\__init__.py -> build\lib\shotgun_api3\lib\httplib2
creating build\lib\shotgun_api3\lib\mockgun
copying shotgun_api3\lib\mockgun\errors.py -> build\lib\shotgun_api3\lib\mockgun
copying shotgun_api3\lib\mockgun\mockgun.py -> build\lib\shotgun_api3\lib\mockgun
copying shotgun_api3\lib\mockgun\schema.py -> build\lib\shotgun_api3\lib\mockgun
copying shotgun_api3\lib\mockgun\__init__.py -> build\lib\shotgun_api3\lib\mockgun
creating build\lib\shotgun_api3\lib\httplib2\python2
copying shotgun_api3\lib\httplib2\python2\certs.py -> build\lib\shotgun_api3\lib\httplib2\python2
copying shotgun_api3\lib\httplib2\python2\iri2uri.py -> build\lib\shotgun_api3\lib\httplib2\python2
copying shotgun_api3\lib\httplib2\python2\socks.py -> build\lib\shotgun_api3\lib\httplib2\python2
copying shotgun_api3\lib\httplib2\python2\__init__.py -> build\lib\shotgun_api3\lib\httplib2\python2
creating build\lib\shotgun_api3\lib\httplib2\python3
copying shotgun_api3\lib\httplib2\python3\certs.py -> build\lib\shotgun_api3\lib\httplib2\python3
copying shotgun_api3\lib\httplib2\python3\iri2uri.py -> build\lib\shotgun_api3\lib\httplib2\python3
copying shotgun_api3\lib\httplib2\python3\socks.py -> build\lib\shotgun_api3\lib\httplib2\python3
copying shotgun_api3\lib\httplib2\python3\__init__.py -> build\lib\shotgun_api3\lib\httplib2\python3
copying shotgun_api3\lib\httplib2\python2\cacerts.txt -> build\lib\shotgun_api3\lib\httplib2\python2
copying shotgun_api3\lib\httplib2\python3\cacerts.txt -> build\lib\shotgun_api3\lib\httplib2\python3
creating build\bdist.win-amd64
creating build\bdist.win-amd64\egg
creating build\bdist.win-amd64\egg\shotgun_api3
creating build\bdist.win-amd64\egg\shotgun_api3\lib
creating build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2
creating build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2
copying build\lib\shotgun_api3\lib\httplib2\python2\cacerts.txt -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2
copying build\lib\shotgun_api3\lib\httplib2\python2\certs.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2
copying build\lib\shotgun_api3\lib\httplib2\python2\iri2uri.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2
copying build\lib\shotgun_api3\lib\httplib2\python2\socks.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2
copying build\lib\shotgun_api3\lib\httplib2\python2\__init__.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2
creating build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3
copying build\lib\shotgun_api3\lib\httplib2\python3\cacerts.txt -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3
copying build\lib\shotgun_api3\lib\httplib2\python3\certs.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3
copying build\lib\shotgun_api3\lib\httplib2\python3\iri2uri.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3
copying build\lib\shotgun_api3\lib\httplib2\python3\socks.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3
copying build\lib\shotgun_api3\lib\httplib2\python3\__init__.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3
copying build\lib\shotgun_api3\lib\httplib2\__init__.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2
copying build\lib\shotgun_api3\lib\mimetypes.py -> build\bdist.win-amd64\egg\shotgun_api3\lib
creating build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun
copying build\lib\shotgun_api3\lib\mockgun\errors.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun
copying build\lib\shotgun_api3\lib\mockgun\mockgun.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun
copying build\lib\shotgun_api3\lib\mockgun\schema.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun
copying build\lib\shotgun_api3\lib\mockgun\__init__.py -> build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun
copying build\lib\shotgun_api3\lib\sgsix.py -> build\bdist.win-amd64\egg\shotgun_api3\lib
copying build\lib\shotgun_api3\lib\sgtimezone.py -> build\bdist.win-amd64\egg\shotgun_api3\lib
copying build\lib\shotgun_api3\lib\six.py -> build\bdist.win-amd64\egg\shotgun_api3\lib
copying build\lib\shotgun_api3\lib\xmlrpclib.py -> build\bdist.win-amd64\egg\shotgun_api3\lib
copying build\lib\shotgun_api3\lib\__init__.py -> build\bdist.win-amd64\egg\shotgun_api3\lib
copying build\lib\shotgun_api3\shotgun.py -> build\bdist.win-amd64\egg\shotgun_api3
copying build\lib\shotgun_api3\__init__.py -> build\bdist.win-amd64\egg\shotgun_api3
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2\certs.py to certs.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2\iri2uri.py to iri2uri.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2\socks.py to socks.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python2\__init__.py to __init__.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3\certs.py to certs.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3\iri2uri.py to iri2uri.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3\socks.py to socks.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\python3\__init__.py to __init__.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\httplib2\__init__.py to __init__.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\mimetypes.py to mimetypes.cpython-37.pyc
  File "build\bdist.win-amd64\egg\shotgun_api3\lib\mimetypes.py", line 571
    print USAGE
              ^
SyntaxError: Missing parentheses in call to 'print'. Did you mean print(USAGE)?

byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun\errors.py to errors.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun\mockgun.py to mockgun.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun\schema.py to schema.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\mockgun\__init__.py to __init__.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\sgsix.py to sgsix.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\sgtimezone.py to sgtimezone.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\six.py to six.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\xmlrpclib.py to xmlrpclib.cpython-37.pyc
  File "build\bdist.win-amd64\egg\shotgun_api3\lib\xmlrpclib.py", line 191
    MAXINT =  2L**31-1
               ^
SyntaxError: invalid syntax

byte-compiling build\bdist.win-amd64\egg\shotgun_api3\lib\__init__.py to __init__.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\shotgun.py to shotgun.cpython-37.pyc
byte-compiling build\bdist.win-amd64\egg\shotgun_api3\__init__.py to __init__.cpython-37.pyc
creating build\bdist.win-amd64\egg\EGG-INFO
copying shotgun_api3.egg-info\PKG-INFO -> build\bdist.win-amd64\egg\EGG-INFO
copying shotgun_api3.egg-info\SOURCES.txt -> build\bdist.win-amd64\egg\EGG-INFO
copying shotgun_api3.egg-info\dependency_links.txt -> build\bdist.win-amd64\egg\EGG-INFO
copying shotgun_api3.egg-info\not-zip-safe -> build\bdist.win-amd64\egg\EGG-INFO
copying shotgun_api3.egg-info\top_level.txt -> build\bdist.win-amd64\egg\EGG-INFO
creating dist
creating 'dist\shotgun_api3-3.2.3-py3.7.egg' and adding 'build\bdist.win-amd64\egg' to it
removing 'build\bdist.win-amd64\egg' (and everything under it)
Processing shotgun_api3-3.2.3-py3.7.egg
creating c:\users\thomas\appdata\roaming\python\python37\site-packages\shotgun_api3-3.2.3-py3.7.egg
Extracting shotgun_api3-3.2.3-py3.7.egg to c:\users\thomas\appdata\roaming\python\python37\site-packages
  File "c:\users\thomas\appdata\roaming\python\python37\site-packages\shotgun_api3-3.2.3-py3.7.egg\shotgun_api3\lib\mimetypes.py", line 571
    print USAGE
              ^
SyntaxError: Missing parentheses in call to 'print'. Did you mean print(USAGE)?

  File "c:\users\thomas\appdata\roaming\python\python37\site-packages\shotgun_api3-3.2.3-py3.7.egg\shotgun_api3\lib\xmlrpclib.py", line 191
    MAXINT =  2L**31-1
               ^
SyntaxError: invalid syntax

Adding shotgun-api3 3.2.3 to easy-install.pth file

Installed c:\users\thomas\appdata\roaming\python\python37\site-packages\shotgun_api3-3.2.3-py3.7.egg
Processing dependencies for shotgun-api3==3.2.3
Finished processing dependencies for shotgun-api3==3.2.3

Any plan to update the vendored modules/packages?

Cheers,

Thomas

Make shotgun api usable with pyinstaller

We're trying to use pyinstaller for easy internal deployment of a single executable written in python.

Compiling works:

pyinstaller sgflop.py -F --upx-dir=/usr/bin/ --hidden-import=shotgun_api3 --hidden-import=ssl --hidden-import=httplib2 --hidden-import=_ssl --hidden-import=certifi --hidden-import=os

Execution fails with this error:

Traceback (most recent call last):
  File "sgflop.py", line 25, in <module>
  File "shotgun_api3/shotgun.py", line 439, in __init__
  File "shotgun_api3/shotgun.py", line 458, in server_caps
  File "shotgun_api3/shotgun.py", line 485, in info
  File "shotgun_api3/shotgun.py", line 1640, in _call_rpc
  File "shotgun_api3/shotgun.py", line 1761, in _make_call
  File "shotgun_api3/shotgun.py", line 1779, in _http_request
  File "shotgun_api3/lib/httplib2/__init__.py", line 1448, in request
  File "shotgun_api3/lib/httplib2/__init__.py", line 1200, in _request
  File "shotgun_api3/lib/httplib2/__init__.py", line 1135, in _conn_request
  File "shotgun_api3/lib/httplib2/__init__.py", line 902, in connect
  File "shotgun_api3/lib/httplib2/__init__.py", line 84, in _ssl_wrap_socket
  File "ssl.py", line 381, in wrap_socket
  File "ssl.py", line 141, in __init__
ssl.SSLError: [Errno 185090050] _ssl.c:340: error:0B084002:x509 certificate routines:X509_load_cert_crl_file:system lib
Failed to execute script sgflop

I found a solution in this thread. Now it works, but I believe this should be elegantly implemented in the httplib2 lib inside shotgun.

Is this something you would include if I file a pull request? Any suggestion how to wrap the certficate in the code?

P.S. Did I mention I would prefer to use a go package for shotgun

Support async workflows

Hi all,

Wanted to raise this issue regarding shotgun_api3 and support for async workflows?

Happy to do the work and PR myself, but I'd like to get opinions on this first!

Thanks,
Josh

faster json decoding with ujson?

Hi! While profiling some of our internal code that we needed to speed up, I noticed that the Shotgun API seemed to have a bottleneck at the json decoding stage (in _json_loads_ascii) I've been running my tests on Python 2.7 and I'm 98% sure we've got the C speedups for simplejson in place, so I don't think that's the issue.

For these preliminary benchmarks, I'm using cprofile to find the cumulative time of making an SG.find() call 10x, which returns the number of entities specified in the lefthand column. The difference is negligible at low numbers, but ujson seems to be exponentially faster at decoding when larger amounts of data are returned (times are in seconds):

entities	simplejson	ujson
1	0.003	0.000
10	0.013	0.000
100	0.105	0.001
1000	0.955	0.012
10000	7.791	0.108

I'm wondering if there's a known reason to not utilize a faster library for decoding json returned by the Shotgun API? I did notice that ujson doesn't appear to have the object_hook arg currently being used in _json_loads_ascii; removing this wasn't having any negative impact in my tests, but I'd be interested in knowing when this object_hook is necessary (since we might consider moving our own fork to ujson.)

Param type incorrectly set to int instead of list in the docstring.

python-api/shotgun_api3/shotgun.py

Line 851 in b3034c1

    
                   :param int order: Optional list of fields to order the results by. List has the format::

Does it make sense to completely work with the rest api?

As there is the rest api, does it make sense to completely rely on that and ditch the python api, or are there still benefit's?

nose in maintenance mode

The nose testing framework is in maintenance mode according to its official docs:

Nose has been in maintenance mode for the past several years and will likely cease without a new person/team to take over maintainership. New projects should consider using Nose2, py.test, or just plain unittest/unittest2.

Are there plans to switch to a different library or framework that is actively being developed?

I was planning to build unit tests based on the tests in this repository.

Could we get Error Code in Shotgun's Exception

Could you add the http return/error code in the Shotgun's Exception ?
I would like to manage some exception based on 4xx and 5xx exceptions.

for instance, here I'll loose all error code :
https://github.com/shotgunsoftware/python-api/blob/master/shotgun_api3/shotgun.py#L2689

same issue here https://github.com/shotgunsoftware/python-api/blob/master/shotgun_api3/shotgun.py#L4121

but it seems _call_rpc is fine since you only do a raise at the end

Import statements in shotgun.py

Hi,

May I make a suggestion as to a slight change.

At the moment, in shotgun.py, you have the lines:

from shotgun_api3.lib.httplib2 import Http, ProxyInfo, socks
from shotgun_api3.lib.sgtimezone import SgTimezone
from shotgun_api3.lib.xmlrpclib import Error, ProtocolError, ResponseError

This means that the shotgun_api3 module has to be directly in a path in $PYTHONPATH, and can't be inside another module.

Can I suggest two things.

Firstly, that the "shotgun_api3." is removed from each of those lines, giving:

from lib.httplib2 import Http, ProxyInfo, socks
from lib.sgtimezone import SgTimezone
from lib.xmlrpclib import Error, ProtocolError, ResponseError

And secondly that, as this could conflict with something else called "lib", this folder (and these import lines) are changed to something like "shotgun_lib" or "sg_lib".

In my case, I have a module called "pipeline_core" that contains all of the core bits and pieces of my pipeline. The Shotgun API lives within this, so I need to be able to do:

from pipeline_core import shotgun_api3

Right now, this doesn't work, as the shotgun.py module inside it can't do the import. However, my suggested change will fix this.

I'm more than happy to submit a pull request for this if it is deemed appropriate.

Cheers

Hugh

UnitTests on Windows - 'tuple' object has no attribute 'replace'

Error occurring when running the unit tests on Windows with Python 2.7.8:

Error
Traceback (most recent call last):
  File "C:\dev\shotgun\python_api3\tests\test_api.py", line 378, in test_share_thumbnail
    shot_path = _get_path(shot_url)
  File "C:\dev\shotgun\python_api3\tests\test_api.py", line 1639, in _get_path
    return os.path.join(url[:4])
  File "C:\Python27\lib\ntpath.py", line 64, in join
    result_drive, result_path = splitdrive(path)
  File "C:\Python27\lib\ntpath.py", line 115, in splitdrive
    normp = p.replace(altsep, sep)
AttributeError: 'tuple' object has no attribute 'replace'

error found when install shotgun-api3 by pip

In our company, we have our own pypi source, so I build shotgun-api3 package first.

tar -xvf python-api-3.2.1.tar.gz
cd python-api-3.2.1
python setup.py sdist

Then I copy shotgun_api3-3.2.1.tar.gz to our pypi source directory.

cp dist/shotgun_api3-3.2.1.tar.gz pypi/source/shotgun_api3/

Finally, I could install shotgun-api3 by pip. But it raise an error.

pip install shotgun-api3

FileNotFoundError: [Errno 2] No such file or directory: 'LICENSE'

I think the reason is you didn't add LICENSE file to package_data but read it in setup.py.

EDIT

If I use git clone to download the source package, there is no problem. But if I download the released package, it can't work. So the problem comes from include_package_data, because the released package doesn't have .git folder.

Can't install python-api in virtual environment

I am using a python virtual environment.

I am using pip to install python-api from git:
pip install git+https://github.com/shotgunsoftware/python-api.git
....
Collecting shotgun_api3 from git+https://github.com/shotgunsoftware/python-api.git#egg=shotgun_api3
Cloning https://github.com/shotgunsoftware/python-api.git to c:\users\abcxyz\appdata\local\temp\pip-install-klwu73\shotgun-api3
Building wheels for collected packages: shotgun-api3
Building wheel for shotgun-api3 (setup.py) ... done
etc ....

The install fails with the following error:
Could not install packages due to an EnvironmentError: [Error 2] The system cannot find the file specified: 'c:\users\abcxyz\appdata\local\temp\pip-install-klwu73\shotgun-api3\tests\Noel.jpg'

When I check the directory for the Noel.jpg, I notice that the file isn't spelled Noel.jpg.
Instead it is spelled Noël.jpg . With an <ë>.
Thus, the path for \Noel.jpg fails.

Please fix this, so that I can install it into my python virtual environment.

Documentation: Deprecated protocol in installation instructions

Issue

The installation documentation contains a method that is no longer supported. This is the resulting error:

$ pip install git+git://github.com/shotgunsoftware/python-api.git
Collecting git+git://github.com/shotgunsoftware/python-api.git
...
fatal: remote error:
The unauthenticated git protocol on port 9418 is no longer supported.
Please see https://github.blog/2021-09-01-improving-git-protocol-security-github/ for more information

Solution

Installing over https works fine:

$ pip install git+https://github.com/shotgunsoftware/python-api.git
Collecting git+https://github.com/shotgunsoftware/python-api.git
...
Building wheels for collected packages: shotgun-api3
  Building wheel for shotgun-api3 (setup.py) ... done
  Created wheel for shotgun-api3: filename=shotgun_api3-3.3.3-py3-none-any.whl size=495076 sha256=7b85ce0030933099890695744157c7750e8532323fe5fc127d2dff959787d465
...
Successfully installed shotgun-api3-3.3.3

References

GitHub blog post detailing deprecation: https://github.blog/2021-09-01-improving-git-protocol-security-github/

Make easy_install compatible

I've been using virtualenv a lot lately to manage my web projects, and it's useful for keeping environments clean with only the required dependencies and lib versions for a project. One thing that makes using virtualenv easy is if any third party modules are installable using easy_install.

Do you think others will want to run:
easy_install shotgun

To install the shotgun module with its dependencies? If it's done this way, there won't need to be a lib directory with third party modules hard-linked there, you can define certain modules (and versions) as dependencies, and they'll get installed as the shotgun module is installed.

Here are the virtualenv docs: http://www.virtualenv.org/en/latest/index.html

And some info on creating python packages: http://packages.python.org/distribute/setuptools.html

Note that if this is done right, users will still be able to add it to their PYTHONPATH like it's done now.

If I get some time, I'll do some work on documenting what it would to take to turn this module into a package. What do you think?

Exception includes user login and password as plaintext

User password and login are included in exceptions as plain text if the shotgun_api encounters an "internal error". This seems to be used in several different functions in shotgun.py. This is obviously a huge issue when storing exceptions for debugging or when using the api in programs that automatically log to an external file. An example use of it is on line 1367 - 1382:

        params.update(self._auth_params())

        except urllib2.HTTPError, e:
            if e.code == 500:
                raise ShotgunError("Server encountered an internal error. "
                    "\n%s\n(%s)\n%s\n\n" % (url, params, e))

Is there something i'm missing here or something I'm doing wrong?

UnitTests on Windows - '.cache' leads to No such file or directory

".cache" doesn't always lead to a correct path on windows which in turn means the file isn't found:

h = Http(".cache")

see: https://github.com/shotgunsoftware/python-api/blob/master/tests/test_api.py#L339


Error
Traceback (most recent call last):
  File "C:\dev\wspy\sandbox_shared\src\shotgun\python_api3\tests\test_api.py", line 309, in test_upload_thumbnail_for_task
    thumb_resp, content = h.request(task_with_thumbnail.get('image'), "GET")
  File "C:\dev\shotgun\python_api3\shotgun_api3\lib\httplib2\__init__.py", line 1448, in request
    (response, content) = self._request(conn, authority, uri, request_uri, method, body, headers, redirections, cachekey)
  File "C:\dev\shotgun\python_api3\shotgun_api3\lib\httplib2\__init__.py", line 1256, in _request
    _updateCache(headers, response, content, self.cache, cachekey)
  File "C:\dev\shotgun\python_api3\shotgun_api3\lib\httplib2\__init__.py", line 438, in _updateCache
    cache.set(cachekey, text)
  File "C:\dev\shotgun\python_api3\shotgun_api3\lib\httplib2\__init__.py", line 707, in set
    f = file(cacheFullPath, "wb")
IOError: [Errno 2] No such file or directory: '.cache\\pixomondodev.shotgunstudio.com,thumbnail,api_image,150,AccessKeyId=d&Expires=xxx&Signature=xxx'

URLError doesnt have code attribute

I dont believe URLError have a code attribue. HTTPError does though. HTTPError inherit from URLError. So I have doubt this case can happen here :
https://github.com/shotgunsoftware/python-api/blob/master/shotgun_api3/shotgun.py#L2668

https://docs.python.org/3/library/urllib.error.html#urllib.error.URLError

Bug in download_attachment() in the JSON branch

Hi,

I've just been getting very confused as to why I'm not getting the right files when I call download_attachment()

It turns out the json version is slightly different from the master branch.

In the master branch (line 889):

    url = urlparse.urlunparse((self.config.scheme, self.config.server, 
        "/file_serve/attachment/%s" % urllib.quote(str(attachment_id)), 
        None, None, None))

In the JSON version (line 880):

    url = urlparse.urlunparse((self.config.scheme, self.config.server, 
        "/file_serve/%s" % urllib.quote(str(attachment_id)), 
        None, None, None))

Note that the actual URL is missing 'attachment/' in the JSON version.

Without this, it gives me files, but I'm not sure what the connection is. With it, it gives me the files I would expect.

I'll put in a pull request at some point in the near future to roll this back to how it was...

If I look at a page of all of th

Mockgun: find_one / find does not support include_archived_projects param

The param is not defined so the mocked code will raise if some call are made with it.
As the default value is True, we actually have some of our core functions that set the value to false for optim purposes.

I fixed the issue in our version of Mockgun (by adding a silly include_archived_projects=True in the function proto), but thought it was worth mentioning :)
On the breaking API thematic, the same function is also missing the additional_filter_presets one.

Did I already said I ❤️ Mockgun ?

share_thumbnail() and filmstrip

I found that share_thumbnail() doesn't work with filmstrips unless you run it twice. Once for the static thumbnail and once for the filmstrip image :

sg.share_thumbnail(entities=[target], source_entity=source, filmstrip_thumbnail=False)
sg.share_thumbnail(entities=[target], source_entity=source, filmstrip_thumbnail=True)

If I leave out the first it won't work inside the shotgun interface.

Is that expected behaviour?

setup.py uses missing 'README' file

Trying to install from pip yields the following error:

$ pip install git+https://github.com/shotgunsoftware/python-api.git
Downloading/unpacking git+https://github.com/shotgunsoftware/python-api.git
  Cloning https://github.com/shotgunsoftware/python-api.git to /tmp/pip-j4u7Qw-build
  Running setup.py (path:/tmp/pip-j4u7Qw-build/setup.py) egg_info for package from git+https://github.com/shotgunsoftware/python-api.git
    Traceback (most recent call last):
      File "<string>", line 17, in <module>
      File "/tmp/pip-j4u7Qw-build/setup.py", line 5, in <module>
        f = open('README')
    IOError: [Errno 2] No such file or directory: 'README'
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):
  File "<string>", line 17, in <module>
  File "/tmp/pip-j4u7Qw-build/setup.py", line 5, in <module>
    f = open('README')
IOError: [Errno 2] No such file or directory: 'README'

The bug was introduced in 2dcb1b8 .

I recommend using travis-ci to check for such errors in future. One of my projects is breaking on travis now that one of it's dependencies cannot be installed anymore using the line above.

Missing relative import

One of the imports on the embedded version of httplib2 for Python 2 was not converted to use relative imports:

python-api/shotgun_api3/lib/httplib2/python2/__init__.py

Line 125 in e63f83a

from iri2uri import iri2uri

It should be from .iri2uri import iri2uri like in the Python 3 version.

Bug in _parse_records where it won't get a URL for a linked image field

Hi,

This is something I came across a while back, and just came across something similar in the GUI, so thought I'd report it.

When requesting a linked entity's image field, it won't convert the result to a URL

For example, if I want to get a Shot's details, and the linked Project's image at the same time, I feel that I should be able to do:

sg.find_one("Shot", [["id", "is", 2418]], fields = ["code", "image", "project.Project.image"])

This currently gives a result of:

{'image': 'https://protected/url/to/thumbnail.jpg', 'type': 'Shot', 'id': 2418, 'project.Project.image': 6867}

Rather than turning the Project's image into a URL

	if connect:
	self.server_caps

	# Check for api_max_entities_per_page in the server info and change the record per page value if it is supplied.
	self.config.records_per_page = self.server_info.get('api_max_entities_per_page') or self.config.records_per_page