[BD-24][TNL-7647] BB-3128: Score linking ADR (#114)

* Write score linking ADR

* Formatting fixes

* Nit

* Addressing review comments

* Final implementation

* Nit

* Updating ADR given discussions

docs/decisions/0003-lti-1p3-score-linking.rst

@@ -0,0 +1,67 @@
+LTI Advantage AGS Score Linking
+-------------------------------
+
+Status
+======
+
+In Review
+
+Context
+=======
+
+LTI Advantage provides new ways for LTI tools to push grades back into the platform through the `Assignment and Grades Services (AGS)`_,
+which don't map 1:1 with the grading and gradebook structure present in Open edX.
+
+There's two models of interaction to pushing grades to the platform in the LTI AGS services:
+
+1. Declarative: the platform creates a LineItem (equivalent of a gradebook line/grade) and tools can only push results to that item.
+2. Programmatic: the tool uses the AGS endpoints to manage it's own line items and grades. The tool is responsible for linking each line item to the resourceLinks, which means that a tool might not link a grade to it's respective problem.
+
+See a more detailed description in the `Coupled vs decoupled line items`_ section of the spec.
+
+.. _`Assignment and Grades Services (AGS)`: https://www.imsglobal.org/spec/lti-ags/v2p0
+.. _`Coupled vs decoupled line items`: https://www.imsglobal.org/spec/lti-ags/v2p0#coupled-vs-decoupled-line-items
+
+
+Decisions
+=========
+
+Given the platform's fixed gradebook structure, we'll implement the declarative interaction model detailed in the
+`LTS-AGS Spec - Declarative interation model`_.
+
+In the future, we want to enable the platform to have full LTI Advantage compatibility, so we need to allow both the declarative and programmatic
+interaction models to happen. There's already full support for the programmatic approach, but we decided to restrict the tool's access since we
+don't want to deal with mixing and averaging multiple grades in the current implementation.
+
+.. _`LTS-AGS Spec - Declarative interation model`: https://www.imsglobal.org/spec/lti-ags/v2p0#declarative-
+
+Declarative grade handling
+~~~~~~~~~~~~~~~~~~~~~~~~~~
+When the LTI configuration is created, a signal should trigger the creation of a LineItem if the tool is configured to use the declarative
+model. The LineItem created will have the following attributes:
+
+.. list-table::
+   :widths: auto
+   :header-rows: 1
+
+   * - Attribute
+     - Value
+   * - lti_configuration
+     - LTI configuration just created.
+   * - resource_id
+     - Blank, this is used by LTI tools in the programmatic interaction model.
+   * - label
+     - The problem title, derived from the block's attributes.
+   * - score_maximum
+     - Maximum score for this given problem, derived from the block's attributes.
+   * - tag
+     - Blank, this is used by LTI tools in the programmatic interaction model.
+   * - start_date_time
+     - The problem's start date, if available in the block's attributes.
+   * - end_date_time
+     - The problem's end date, if available in the block's attributes.
+
+Consequences
+============
+
+This will NOT make the platform LTI compliant since it doesn't allow the programmatic grade interaction.