From 7d898c6297372d97916e7db354b15db8d3f4d570 Mon Sep 17 00:00:00 2001
From: Ricard Illa <ricard@trkkn.com>
Date: Mon, 26 Jun 2023 10:04:28 +0200
Subject: [PATCH] fix: airflow's BashOperator's cwd cannot be templated

---
 dags/sustainability_score/README.md   | 5 ++++-
 dags/sustainability_score/__init__.py | 3 +--
 2 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/dags/sustainability_score/README.md b/dags/sustainability_score/README.md
index ebd1ffb..1cdbfcb 100644
--- a/dags/sustainability_score/README.md
+++ b/dags/sustainability_score/README.md
@@ -16,8 +16,11 @@ The following parameters are available:
 
 * `input`: location of the CSV input file
 * `beam_etl_path`: location of the apache beam pipeline
-* `dbt_path`: location of the dbt project
 * `products_table`: products_table table name
 
 I decided not to configure the rest of the table locations because that makes
 more sense to be defined in DBT.
+
+Ideally, I would parametrize the dbt path as well. But the `cwd` parameter of
+`BashOperator` is not a template and implementing that is not worth it for such
+a minor improvement.
diff --git a/dags/sustainability_score/__init__.py b/dags/sustainability_score/__init__.py
index f1072a9..ead9ef9 100644
--- a/dags/sustainability_score/__init__.py
+++ b/dags/sustainability_score/__init__.py
@@ -23,7 +23,6 @@ CSV_FNAME = (
 CONFIG = {
     "input": f"{ HOME }/gcs/data/{ CSV_FNAME }",
     "beam_etl_path": "/etl/main.py",
-    "dbt_path": "/dbt",
     "products_table": "sustainability_score.products",
 }
 
@@ -34,7 +33,7 @@ def dbt(cmd: str, attach_dag: DAG) -> BashOperator:
         dag=attach_dag,
         task_id=f"dbt_{ cmd }",
         bash_command=f"dbt { cmd }",
-        cwd="{{ params.dbt_path }}",
+        cwd="/dbt",
         env={
             "POSTGRES_HOST": "{{ conn.get('pg_db').host }}",
             "POSTGRES_USER": "{{ conn.get('pg_db').login }}",