LTI 1.3 Deep Linking Assignment No Speedgrader Preview

[FunkyPajamas]

Hi,

I'm currently working on LTI 1.3 integration with Canvas. I created the Developer Key using the manual entry, with the Assignment Selection placement set to LtiDeepLinkingRequest. This seems to work fine: I can go from my Canvas development instance to my site by selecting "External Tool" in the Submission Type, choose my tool, pick the LTI content for deep linking, go back to Canvas, save the assignment, preview it, etc. I see the Line Item created in the Gradebook and I can switch to a student, load the LTI content from my site into Canvas, and submit the content in my site, which I can see sends the score back to the Canvas gradebook by POSTing to the create score endpoint.

When I switch back to a teacher and go to the Speedgrader, I get a "No Preview Available" blank page. I don't see any LTI requests to my site so I think I'm missing something in the configuration of the LTI key, or maybe in the creation of the assignment itself.

I checked the AGS documentation on Canvas (https://canvas.instructure.com/doc/api/file.assignment_tools.html) and this is the only mention I found for "Speedgrader":

The Score service has been extended to allow an external tool to submission data back to the Canvas Gradebook. This data is then exposed in the Submission Details and Speedgrader Views so that both students and teachers can see what was submitted to the external tool without leaving Canvas. Support for basic urls, text, and LTI links are supported.

So I checked the Score service documentation (https://canvas.instructure.com/doc/api/score.html#method.lti/ims/scores.create) and I don't see anything there that relates to Speedgrader. 

Any ideas on how to get my LTI content to show in the Speedgrader preview? Thanks!

[michellepm]

Hello! Did you figure this out? I have this same exact question!

[JohnOpferkuch]

I got this working... (I think... still developing our solution) using https://github.com/dmitry-viskov/pylti1.3

Simplified code below.  The bit that works is documented here:

https://canvas.instructure.com/doc/api/score.html#method.lti/ims/scores.create

import datetime

from pylti1p3.contrib.django import (
    DjangoDbToolConf,
    DjangoMessageLaunch,
)
from pylti1p3.grade import Grade

message_launch = DjangoMessageLaunch(...)
lti_1_3_data = message_launch.get_launch_data()
earned_score = 49 # arbitrary.... 

resource_link_id = (
    lti_1_3_data
    .get("https://purl.imsglobal.org/spec/lti/claim/resource_link", {})
    .get("id")
)
sub = lti_1_3_data.get("sub")
timestamp = datetime.datetime.utcnow().isoformat() + "Z"
ags = message_launch.get_ags() 

sc = Grade()
sc.set_score_given(earned_score).set_score_maximum(100).set_timestamp(
   timestamp
).set_activity_progress("Completed").set_grading_progress(
   "FullyGraded"
).set_user_id(sub)


# see https://canvas.instructure.com/doc/api/score.html#method.lti/ims/scores.create
sc.set_extra_claims(
    {
        "https://canvas.instructure.com/lti/submission": {
            "submission_type": "basic_lti_launch",
            "submission_data": f"http://localhost:8000/lti/1.3/launch/?assignment_id={assignment_id}&submission_id={submission_id}",
        }
    }
)

result = ags.put_grade(sc)