python model reference f"{dbt.ref('test5')}"

[FishtownBuildBot]

f"{dbt.ref('test5')}"

is now a supported way to reference a model in a python model

cc: @ChenyuLInx

[runleonarun]

@runleonarun adding Copilot summary

This pull request primarily focuses on fixing the parsing of f-strings in Python models. The changes include adding support for dbt function calls in f-strings, modifying test cases to include f-string usage, and adding a new test case specifically for f-string configuration.

Fixes and improvements:

.changes/unreleased/Fixes-20231128-102111.yaml: Announced the fix for parsing f-strings in Python models.
Changes to the Python model parser:

core/dbt/parser/models.py: Added support for dbt function calls in f-strings. This change allows the parser to correctly interpret dbt function calls that are embedded within f-strings.
Changes to the test cases:

tests/unit/test_parser.py: Modified the model function to include f-string usage. This change tests the parser's ability to correctly interpret f-strings.
tests/unit/test_parser.py: Added a new Python model that uses f-strings. This model is used in a new test case that verifies the correct parsing of f-strings.
tests/unit/test_parser.py: Modified the test_python_model_parse function to include a new reference argument that uses f-strings. This change tests the parser's ability to correctly interpret reference arguments within f-strings.
tests/unit/test_parser.py: Added a new test case test_python_model_f_string_config to verify the correct parsing of f-string configuration in Python models.

[mirnawong1]

thank you @runleonarun ! working on this now!

[mirnawong1]

hey @ChenyuLInx are we saying there is now another way to ref python models? so:

• dbt.ref("upstream_model_name") -- existing way
• f"{dbt.ref('test5')}" -- new change per Fix python model parsing from meder dbt-core#9162

or is the f string the new and only way? also is there any change when referencing in downstream models (e.g. select * from {{ ref('my_python_model') }})?

[mirnawong1]

hey @ChenyuLInx , friendly nudge if you have any clarification on this pls? thank you!

[ChenyuLInx]

Hey @mirnawong1 , the change means that when using a config in a python f string, the config will be available in python models. Here's an example:
If my_var config for the current model is 5, before this change,

print(f"{dbt.config.get('my_var')}")

will be None, but after this change, it will actually print 5.