fix: missed condition in adding vars args

[AlexandrKhabarov]

Description

DbtToAirflowConverter can pass dbt_vars to DbtGraph with help of ProjectConfig or operator_args. If operator_args is used in DbtToAirflowConverter then it will lead to the issue with absence of dbt_vars in dbt ls command (faced rendering issue during usage of cosmos with project level variables)

Related Issue(s)

Breaking Change?

No breaking changes

Checklist

• I have made corresponding changes to the documentation (if required)
• I have added tests that prove my fix is effective or that my feature works

[netlify]

✅ Deploy Preview for sunny-pastelito-5ecb04 canceled.

Name	Link
🔨 Latest commit	cf45616
🔍 Latest deploy log	https://app.netlify.com/sites/sunny-pastelito-5ecb04/deploys/67b35b28f03791000805664a

[AlexandrKhabarov]

@tatiana can you take a look, please?

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.30%. Comparing base (968b80e) to head (cf45616).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1543   +/-   ##
=======================================
  Coverage   97.30%   97.30%           
=======================================
  Files          80       80           
  Lines        4854     4856    +2     
=======================================
+ Hits         4723     4725    +2     
  Misses        131      131           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pankajkoti]

Suggested change

	if self.project.dbt_vars:
	cmd_args.extend(["--vars", json.dumps(self.project.dbt_vars, sort_keys=True)])
	elif self.dbt_vars:
	cmd_args.extend(["--vars", json.dumps(self.dbt_vars, sort_keys=True)])
	if self.dbt_vars:
	cmd_args.extend(["--vars", json.dumps(self.dbt_vars, sort_keys=True)])

This should yield the same result, right?

@tatiana Was there a specific reason we checked for dbt_vars only in self.project instead of directly using self.dbt_vars while working on #1114?

@AlexandrKhabarov While Tatiana(she is Out of Office this week) provides insights on this, logically speaking, wouldn’t it make more sense to use the variables set at the project level during parsing/graph building rather than the ones set for the operator(I would incline to think that operator args would be used for each of the operator tasks during execution)? What do you think?

[AlexandrKhabarov]

So, I saw this part of the code and decided that project level variables have higher priority than operator vars and here can be two different approaches to pass variables to dbt (we are passing dbt variables via operator_args)

And this PR was intended to follow to the same logic in listing resource of dbt

Also, project level variables can be the same as variables of operator, right? Listing resource of dbt will lead to duplicated logic on client side (I will need to pass the same variable in two places)