sg_info_file = os.environ['SHOTGUN_INFO']

[maximcapra]

When is this environment variable set and what is the content of this file? Is this a custom file or does it come with the shotgun installation?

[fdan]

Hi, we need to update the readme to cover this, but I'll answer here first.

$SHOTGUN_INFO is set to point to a json file, which contains some information about project paths and id's, here's an example:

{
    "install":
    {
        "s118": "/proj/s118",
        "s119": "/proj/s119",
        "s120": "/proj/s120"
    },
    "platform":
    {
        "windows": "win32",
        "linux": "linux2",
        "osx": "darwin"
    },
    "id":
    {
        "s118": "129",
        "s119": "164",
        "s120": "167"
    }
}

The idea is the user decided when to set this env var on their system. We actually set all this via rez packages, however the idea behind this env var / json file was to make it not dependant on the package manager we use in any way.

Hope that helps,

cheers

[fdan]

turret_resolver also relies on an env var being set called $SG_API_KEY, which specifies your own API key (keep that out of public source control obviously).

I'll try to update the documentation to cover these env vars soon.

[fdan]

Oh and one more, we mostly use a shared tk config across all our shotgun projects, so we use this env var to point to the "tank" module location in that shared config.

[maximcapra]

turret_resolver also relies on an env var being set called $SG_API_KEY, which specifies your own API key (keep that out of public source control obviously).

I'll try to update the documentation to cover these env vars soon.

I see. The authentification is pretty straight forward, I was just wondering about the other custom stuff. Is there a reason, why you did not choose to authenticate and get the project, tank and sgtk via the shotgun_api3 python module? This seems pretty complicated to me and maybe I don't fully understand why it is done this way, but I would like to avoid creating additional files (and update them) just for project and path information ... btw, using conda here as a package manager instead of rez ....

[maximcapra]

Thanks so much for the json example and the explanation, wasn't obvious for me what's going on in the background, this helps a lot!

[fdan]

Generally as something developed at a university with a very small crew, there's probably some areas that could be improved such as authentication as you suggested. We had a pretty aggressive timeframe to put all of this into a working state for student projects, so it's entirely possible there's a better way to handle the authentication. If you have any examples of the shotgun_api3 approach I'd be keen to adopt that approach, cheers.

[maximcapra]

You did an absolute outstanding job! Even better, you put it on Github for others, I would not be able to pull this off by myself. I'm currently rewriting the resolver to fit our needs here in our studio. Basically you can get a lot of info about your projects with the api3 module like this:

    import shotgun_api3
    import os

    sg = shotgun_api3.Shotgun(os.environ["SG_SITE"], script_name=os.environ["SG_SCRIPT_NAME"], api_key=os.environ["SG_SCRIPT_KEY"])

    # Find all projects
    fields = [
        "id",
        "code",
        "windows_path",
        "linux_path",
        "mac_path",
        "project",
        "project.Project.tank_name",
        "project.Project.code",
    ]

    projects = sg.find("PipelineConfiguration",[],fields)

    for project in projects:
        print project

So I am basically trying to get the sgtk and the tank paths from there. We have no shared tank setup here, every Project has it's own sgtk and pipeline configuration. Like this:

    import shotgun_api3
    import os

    class Shotgun_Data():

        def __init__(self):
            self.sg = shotgun_api3.Shotgun(os.environ["SG_SITE"], script_name=os.environ["SG_SCRIPT_NAME"], api_key=os.environ["SG_SCRIPT_KEY"])

        def close_connection(self):
            self.sg.close()

        def get_pipeline_config(self, tank_name):
            filters = [["project.Project.tank_name", "is", tank_name]]
            fields = ["id",
                    "code",
                    "windows_path",
                    "linux_path",
                    "mac_path",
                    "project",
                    "project.Project.tank_name",
                    "project.Project.code",
                    "project.Project.sg_resolution"]

            pipeline_config = self.sg.find_one("PipelineConfiguration", filters, fields)
            return pipeline_config

    def import_sgtk():

        sg_connection = Shotgun_Data()

        # get pipeline config
        pipeline_path = None
        project_config = sg_connection.get_pipeline_config("INSERT_tank_name_HERE")

        # Generate Pipeline Path
        if project_config:
            windows_path = project_config["windows_path"]
            linux_path = project_config["linux_path"]
            mac_path = project_config["mac_path"]
            os_name = platform.system()
            if os_name == "Windows":
                pipeline_path = windows_path
            if os_name == "Linux":
                pipeline_path = linux_path
            if os_name == "Darwin":
                pipeline_path = mac_path
        else:
            return

        # Add SGTK-Module path to sys
        sgtk_module_path = os.path.join(os.path.normpath(pipeline_path), "install", "core", "python")
        if not sgtk_module_path in sys.path:
            sys.path.insert(0, sgtk_module_path)

        # Import Module and Create TANK Instance
        import sgtk