Metric Views: doc() is undefined when compiling metric_view models in 1.12

[ondrapaul]

Describe the bug

The metric_view materialization renders its model body (the YAML containing comment: / display_name: fields) through the model-SQL Jinja context which does not include the doc() function. As a result, {{ doc('my_block') }} inside a metric view fails to compile:

 Compilation Error in model my_metric_view
      'doc' is undefined. This can happen when calling a macro that does not exist.

Every other Jinja construct works in the same metric-view body — {{ ref() }}, {{ source() }}, {{ config() }}, {{ var() }}, and custom macros all render correctly. Only doc() is missing from the context. A macro that internally calls {{ doc(name) }} fails the same way, confirming doc is simply absent from the render context (not a doc-block lookup problem).

This breaks documentation parity: for normal models, column description: fields in schema YAML support doc() (and dbt doc blocks). Metric-view comment: fields cannot, so a description maintained once as a doc block cannot be reused in a metric view — forcing duplicated, drift-prone copy.

Steps To Reproduce

1. Define a doc block:

   {% docs id %}
   Unique identifier of the entity.
   {% enddocs %}
   

2. Create a metric view models/example_mv.sql:

   {{ config(materialized="metric_view") }}
   
   version: 1.1
   
   source: {{ ref('my_fact')}}
   
   dimensions:
     - name: id
       expr: id
       display_name: "Entity ID"
       comment: "{{ doc('enterprise_id_doc') }}" --quoted or unquoted version both fail
   

3. dbt parse → succeeds (does not catch the problem).

4. dbt compile --select example_mv → fails with 'doc' is undefined.

Swapping {{ doc('enterprise_id_doc') }} for {{ var('some_var') }} or a custom macro compiles and renders correctly — only doc() is unavailable.

Expected behavior

doc() should resolve inside metric_view documentation fields (comment:, display_name:) just as it does in schema-YAML description: fields, so a single doc block can document both regular-model columns and metric-view dimensions/measures, and persist_docs output stays consistent.

Screenshots and log output

�[0m10:30:09  Running with dbt=1.11.8
�[0m10:30:11  Registered adapter: databricks=1.12.0
�[0m10:30:14  Found 206 models, 832 data tests, 16 seeds, 4 snapshots, 69 sources, 1982 macros, 3 unit tests
�[0m10:30:14  
�[0m10:30:14  Concurrency: 10 threads (target='playground')
�[0m10:30:14  
�[0m10:30:16  1 of 1 START sql metric_view model ondra_paul.mv_ic_plus_gross_fees ............ [RUN]
�[0m10:30:16  1 of 1 ERROR creating sql metric_view model ondra_paul.mv_ic_plus_gross_fees ... [�[31mERROR�[0m in 0.02s]
�[0m10:30:16  
�[0m10:30:16  Finished running 1 metric view model in 0 hours 0 minutes and 2.64 seconds (2.64s).
�[0m10:30:17  
�[0m10:30:17  �[31mCompleted with 1 error, 0 partial successes, and 0 warnings:�[0m
�[0m10:30:17  
�[0m10:30:17  �[31mFailure in model mv_ic_plus_gross_fees (models/metric_views/fees/mv_ic_plus_gross_fees.sql)�[0m
�[0m10:30:17    Compilation Error in model mv_ic_plus_gross_fees (models/metric_views/fees/mv_ic_plus_gross_fees.sql)
  'doc' is undefined. This can happen when calling a macro that does not exist. Check for typos and/or install package dependencies with "dbt deps".
�[0m10:30:17  
�[0m10:30:17  Done. PASS=0 WARN=0 ERROR=1 SKIP=0 NO-OP=0 TOTAL=1

System information

The output of dbt --version:
Core:

• installed: 1.11.8
  Plugins:
• databricks: 1.12.0
• spark: 1.10.1

The operating system you're using: macOS (Darwin)

The output of python --version:
Python 3.10.11