Add performance integration tests (astronomer#827)

## Description

This PR adds a step to our CI to measure how quickly Cosmos can run
models. This is part of a larger initiative to make the project more
performant now that it's reached a certain level of maturity.

How it works:
- We now have [a test that generates a dbt project with a certain number
of sequential
models](https://github.com/astronomer/astronomer-cosmos/blob/performance-int-tests/tests/perf/test_performance.py)
(based on a parameter that gets passed in), runs a simple DAG, and
measures task throughput (measured in terms of models run per second
- I've extended our CI to run this test for 1, 10, 50, and 100 models to
start
- This CI reports out a GitHub Actions output that gets shown in the
actions summary, [at the
bottom](https://github.com/astronomer/astronomer-cosmos/actions/runs/7894490582)

While this isn't perfect, it's a step in the right direction - we now
have some general visibility! Note that these numbers may not be
indicative of a production Airflow environment running something like
the Kubernetes Executor, because this runs a local executor on GH
Actions runners. Still, it's meant as a benchmark to demonstrate whether
we're moving in the right direction or not.

As part of this, I've also refactored our tests to call a script from
the pyproject file instead of embedding the scripts directly in the
file. This should make it easier to maintain and track changes.

<!-- Add a brief but complete description of the change. -->

## Related Issue(s)

<!-- If this PR closes an issue, you can use a keyword to auto-close.
-->
<!-- i.e. "closes #0000" -->
astronomer#800

## Breaking Change?

<!-- If this introduces a breaking change, specify that here. -->

## 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

.github/workflows/test.yml

@@ -11,10 +11,8 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-
   Authorize:
-    environment:
-      ${{ github.event_name == 'pull_request_target' &&
+    environment: ${{ github.event_name == 'pull_request_target' &&
       github.event.pull_request.head.repo.full_name != github.repository &&
       'external' || 'internal' }}
     runs-on: ubuntu-latest
@@ -30,8 +28,8 @@ jobs:
 
       - uses: actions/setup-python@v3
         with:
-          python-version: '3.9'
-          architecture: 'x64'
+          python-version: "3.9"
+          architecture: "x64"
 
       - run: pip3 install hatch
       - run: hatch run tests.py3.9-2.7:type-check
@@ -294,6 +292,55 @@ jobs:
       AIRFLOW_CONN_AIRFLOW_DB: postgres://postgres:[email protected]:5432/postgres
       PYTHONPATH: /home/runner/work/astronomer-cosmos/astronomer-cosmos/:$PYTHONPATH
 
+  Run-Performance-Tests:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.11"]
+        airflow-version: ["2.7"]
+        num-models: [1, 10, 50, 100]
+
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          ref: ${{ github.event.pull_request.head.sha || github.ref }}
+      - uses: actions/cache@v3
+        with:
+          path: |
+            ~/.cache/pip
+            .nox
+          key: perf-test-${{ runner.os }}-${{ matrix.python-version }}-${{ matrix.airflow-version }}-${{ hashFiles('pyproject.toml') }}-${{ hashFiles('cosmos/__init__.py') }}
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v4
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install packages and dependencies
+        run: |
+          python -m pip install hatch
+          hatch -e tests.py${{ matrix.python-version }}-${{ matrix.airflow-version }} run pip freeze
+
+      - name: Run performance tests against against Airflow ${{ matrix.airflow-version }} and Python ${{ matrix.python-version }}
+        id: run-performance-tests
+        run: |
+          hatch run tests.py${{ matrix.python-version }}-${{ matrix.airflow-version }}:test-performance-setup
+          hatch run tests.py${{ matrix.python-version }}-${{ matrix.airflow-version }}:test-performance
+
+          # read the performance results and set them as an env var for the next step
+          # format: NUM_MODELS={num_models}\nTIME={end - start}\n
+          cat /tmp/performance_results.txt > $GITHUB_STEP_SUMMARY
+        env:
+          AIRFLOW_HOME: /home/runner/work/astronomer-cosmos/astronomer-cosmos/
+          AIRFLOW_CONN_AIRFLOW_DB: postgres://postgres:[email protected]:5432/postgres
+          AIRFLOW__CORE__DAGBAG_IMPORT_TIMEOUT: 90.0
+          PYTHONPATH: /home/runner/work/astronomer-cosmos/astronomer-cosmos/:$PYTHONPATH
+          MODEL_COUNT: ${{ matrix.num-models }}
+
+    env:
+      AIRFLOW_HOME: /home/runner/work/astronomer-cosmos/astronomer-cosmos/
+      AIRFLOW_CONN_AIRFLOW_DB: postgres://postgres:[email protected]:5432/postgres
+      PYTHONPATH: /home/runner/work/astronomer-cosmos/astronomer-cosmos/:$PYTHONPATH
 
   Code-Coverage:
     if: github.event.action != 'labeled'
@@ -309,7 +356,7 @@ jobs:
       - name: Set up Python 3.11
         uses: actions/setup-python@v3
         with:
-          python-version: '3.11'
+          python-version: "3.11"
       - name: Install coverage
         run: |
           pip3 install coverage

dev/dags/dbt/perf/.gitignore

@@ -0,0 +1,4 @@
+
+target/
+dbt_packages/
+logs/

dev/dags/dbt/perf/README.md

@@ -0,0 +1,3 @@
+dbt project for running performance tests.
+
+The `models` directory gets populated by an integration test defined in `tests/perf`.

dev/dags/dbt/perf/dbt_project.yml

@@ -0,0 +1,17 @@
+# Name your project! Project names should contain only lowercase characters
+# and underscores. A good package name should reflect your organization's
+# name or the intended use of these models
+name: "perf"
+version: "1.0.0"
+config-version: 2
+
+model-paths: ["models"]
+analysis-paths: ["analyses"]
+test-paths: ["tests"]
+seed-paths: ["seeds"]
+macro-paths: ["macros"]
+snapshot-paths: ["snapshots"]
+
+clean-targets: # directories to be removed by `dbt clean`
+  - "target"
+  - "dbt_packages"

dev/dags/dbt/perf/profiles.yml

@@ -0,0 +1,11 @@
+simple:
+  target: dev
+  outputs:
+    dev:
+      type: sqlite
+      threads: 1
+      database: "database"
+      schema: "main"
+      schemas_and_paths:
+        main: "{{ env_var('DBT_SQLITE_PATH') }}/imdb.db"
+      schema_directory: "{{ env_var('DBT_SQLITE_PATH') }}"

dev/dags/performance_dag.py

@@ -0,0 +1,36 @@
+"""
+A DAG that uses Cosmos to render a dbt project for performance testing.
+"""
+
+import airflow
+from datetime import datetime
+import os
+from pathlib import Path
+
+from cosmos import DbtDag, ProjectConfig, ProfileConfig, RenderConfig
+
+DEFAULT_DBT_ROOT_PATH = Path(__file__).parent / "dbt"
+DBT_ROOT_PATH = Path(os.getenv("DBT_ROOT_PATH", DEFAULT_DBT_ROOT_PATH))
+DBT_SQLITE_PATH = str(DEFAULT_DBT_ROOT_PATH / "data")
+
+profile_config = ProfileConfig(
+    profile_name="simple",
+    target_name="dev",
+    profiles_yml_filepath=(DBT_ROOT_PATH / "simple/profiles.yml"),
+)
+
+cosmos_perf_dag = DbtDag(
+    project_config=ProjectConfig(
+        DBT_ROOT_PATH / "perf",
+        env_vars={"DBT_SQLITE_PATH": DBT_SQLITE_PATH},
+    ),
+    profile_config=profile_config,
+    render_config=RenderConfig(
+        dbt_deps=False,
+    ),
+    # normal dag parameters
+    schedule_interval=None,
+    start_date=datetime(2024, 1, 1),
+    catchup=False,
+    dag_id="performance_dag",
+)

pyproject.toml

@@ -9,16 +9,8 @@ description = "Orchestrate your dbt projects in Airflow"
 readme = "README.rst"
 license = "Apache-2.0"
 requires-python = ">=3.8"
-authors = [
-    { name = "Astronomer", email = "[email protected]" },
-]
-keywords = [
-    "airflow",
-    "apache-airflow",
-    "astronomer",
-    "dags",
-    "dbt",
-]
+authors = [{ name = "Astronomer", email = "[email protected]" }]
+keywords = ["airflow", "apache-airflow", "astronomer", "dags", "dbt"]
 classifiers = [
     "Development Status :: 3 - Alpha",
     "Environment :: Web Environment",
@@ -56,48 +48,23 @@ dbt-all = [
     "dbt-spark",
     "dbt-vertica",
 ]
-dbt-athena = [
-    "dbt-athena-community",
-    "apache-airflow-providers-amazon>=8.0.0",
-]
-dbt-bigquery = [
-    "dbt-bigquery",
-]
-dbt-databricks = [
-    "dbt-databricks",
-]
-dbt-exasol = [
-    "dbt-exasol",
-]
-dbt-postgres = [
-    "dbt-postgres",
-]
-dbt-redshift = [
-    "dbt-redshift",
-]
-dbt-snowflake = [
-    "dbt-snowflake",
-]
-dbt-spark = [
-    "dbt-spark",
-]
-dbt-vertica = [
-    "dbt-vertica<=1.5.4",
-]
-openlineage = [
-    "openlineage-integration-common",
-    "openlineage-airflow",
-]
-all = [
-    "astronomer-cosmos[dbt-all]",
-    "astronomer-cosmos[openlineage]"
-]
-docs =[
+dbt-athena = ["dbt-athena-community", "apache-airflow-providers-amazon>=8.0.0"]
+dbt-bigquery = ["dbt-bigquery"]
+dbt-databricks = ["dbt-databricks"]
+dbt-exasol = ["dbt-exasol"]
+dbt-postgres = ["dbt-postgres"]
+dbt-redshift = ["dbt-redshift"]
+dbt-snowflake = ["dbt-snowflake"]
+dbt-spark = ["dbt-spark"]
+dbt-vertica = ["dbt-vertica<=1.5.4"]
+openlineage = ["openlineage-integration-common", "openlineage-airflow"]
+all = ["astronomer-cosmos[dbt-all]", "astronomer-cosmos[openlineage]"]
+docs = [
     "sphinx",
     "pydata-sphinx-theme",
     "sphinx-autobuild",
     "sphinx-autoapi",
-    "apache-airflow-providers-cncf-kubernetes>=5.1.1"
+    "apache-airflow-providers-cncf-kubernetes>=5.1.1",
 ]
 tests = [
     "packaging",
@@ -137,9 +104,7 @@ Documentation = "https://astronomer.github.io/astronomer-cosmos"
 path = "cosmos/__init__.py"
 
 [tool.hatch.build.targets.sdist]
-include = [
-    "/cosmos",
-]
+include = ["/cosmos"]
 
 [tool.hatch.build.targets.wheel]
 packages = ["cosmos"]
@@ -175,51 +140,20 @@ matrix.airflow.dependencies = [
 [tool.hatch.envs.tests.scripts]
 freeze = "pip freeze"
 type-check = "mypy cosmos"
-test = 'pytest -vv --durations=0 . -m "not integration" --ignore=tests/test_example_dags.py --ignore=tests/test_example_dags_no_connections.py'
-test-cov = """pytest -vv --cov=cosmos --cov-report=term-missing --cov-report=xml --durations=0 -m "not integration" --ignore=tests/test_example_dags.py --ignore=tests/test_example_dags_no_connections.py"""
-# we install using the following workaround to overcome installation conflicts, such as:
-# apache-airflow 2.3.0 and dbt-core [0.13.0 - 1.5.2] and jinja2>=3.0.0 because these package versions have conflicting dependencies
-test-integration-setup = """pip uninstall -y dbt-postgres dbt-databricks dbt-vertica; \
-rm -rf airflow.*; \
-airflow db init; \
-pip install  'dbt-core'  'dbt-databricks' 'dbt-postgres' 'dbt-vertica' 'openlineage-airflow'"""
-test-integration = """rm -rf dbt/jaffle_shop/dbt_packages;
-pytest -vv \
---cov=cosmos \
---cov-report=term-missing \
---cov-report=xml \
---durations=0 \
--m integration  \
--k 'not (sqlite or example_cosmos_sources or example_cosmos_python_models or example_virtualenv)'"""
-test-integration-expensive = """pytest -vv \
---cov=cosmos \
---cov-report=term-missing \
---cov-report=xml \
---durations=0 \
--m integration  \
--k 'example_cosmos_python_models or example_virtualenv'"""
-test-integration-sqlite-setup = """pip uninstall -y dbt-core dbt-sqlite openlineage-airflow openlineage-integration-common; \
-rm -rf airflow.*; \
-airflow db init; \
-pip install 'dbt-core==1.4' 'dbt-sqlite<=1.4' 'dbt-databricks<=1.4' 'dbt-postgres<=1.4' """
-test-integration-sqlite = """
-pytest -vv \
---cov=cosmos \
---cov-report=term-missing \
---cov-report=xml \
---durations=0 \
--m integration  \
--k 'example_cosmos_sources or sqlite'"""
+test = 'sh scripts/test/unit.sh'
+test-cov = 'sh scripts/test/unit-cov.sh'
+test-integration-setup = 'sh scripts/test/integration-setup.sh'
+test-integration = 'sh scripts/test/integration.sh'
+test-integration-expensive = 'sh scripts/test/integration-expensive.sh'
+test-integration-sqlite-setup = 'sh scripts/test/integration-sqlite-setup.sh'
+test-integration-sqlite = 'sh scripts/test/integration-sqlite.sh'
+test-performance-setup = 'sh scripts/test/performance-setup.sh'
+test-performance = 'sh scripts/test/performance.sh'
 
 [tool.pytest.ini_options]
-filterwarnings = [
-    "ignore::DeprecationWarning",
-]
+filterwarnings = ["ignore::DeprecationWarning"]
 minversion = "6.0"
-markers = [
-    "integration",
-    "sqlite"
-]
+markers = ["integration", "sqlite", "perf"]
 
 ######################################
 # DOCS
@@ -233,7 +167,7 @@ dependencies = [
     "sphinx-autobuild",
     "sphinx-autoapi",
     "openlineage-airflow",
-    "apache-airflow-providers-cncf-kubernetes>=5.1.1"
+    "apache-airflow-providers-cncf-kubernetes>=5.1.1",
 ]
 
 [tool.hatch.envs.docs.scripts]

scripts/test/integration-expensive.sh

@@ -0,0 +1,8 @@
+pytest -vv \
+    --cov=cosmos \
+    --cov-report=term-missing \
+    --cov-report=xml \
+    --durations=0 \
+    -m integration  \
+    --ignore=tests/perf \
+    -k 'example_cosmos_python_models or example_virtualenv'

scripts/test/integration-setup.sh

@@ -0,0 +1,6 @@
+# we install using the following workaround to overcome installation conflicts, such as:
+# apache-airflow 2.3.0 and dbt-core [0.13.0 - 1.5.2] and jinja2>=3.0.0 because these package versions have conflicting dependencies
+pip uninstall -y dbt-postgres dbt-databricks dbt-vertica; \
+rm -rf airflow.*; \
+airflow db init; \
+pip install 'dbt-core' 'dbt-databricks' 'dbt-postgres' 'dbt-vertica' 'openlineage-airflow'

scripts/test/integration-sqlite-setup.sh

@@ -0,0 +1,4 @@
+pip uninstall -y dbt-core dbt-sqlite openlineage-airflow openlineage-integration-common; \
+rm -rf airflow.*; \
+airflow db init; \
+pip install 'dbt-core==1.4' 'dbt-sqlite<=1.4' 'dbt-databricks<=1.4' 'dbt-postgres<=1.4'

scripts/test/integration-sqlite.sh

@@ -0,0 +1,8 @@
+pytest -vv \
+    --cov=cosmos \
+    --cov-report=term-missing \
+    --cov-report=xml \
+    --durations=0 \
+    -m integration  \
+    --ignore=tests/perf \
+    -k 'example_cosmos_sources or sqlite'

scripts/test/integration.sh

@@ -0,0 +1,9 @@
+rm -rf dbt/jaffle_shop/dbt_packages;
+pytest -vv \
+    --cov=cosmos \
+    --cov-report=term-missing \
+    --cov-report=xml \
+    --durations=0 \
+    -m integration  \
+    --ignore=tests/perf \
+    -k 'not (sqlite or example_cosmos_sources or example_cosmos_python_models or example_virtualenv)'

scripts/test/performance-setup.sh

@@ -0,0 +1,4 @@
+pip uninstall -y dbt-core dbt-sqlite openlineage-airflow openlineage-integration-common; \
+rm -rf airflow.*; \
+airflow db init; \
+pip install 'dbt-core==1.4' 'dbt-sqlite<=1.4' 'dbt-databricks<=1.4' 'dbt-postgres<=1.4'

scripts/test/performance.sh

@@ -0,0 +1,5 @@
+pytest -vv \
+    -s \
+    -m 'perf' \
+    --ignore=tests/test_example_dags.py \
+    --ignore=tests/test_example_dags_no_connections.py

scripts/test/unit-cov.sh

@@ -0,0 +1,10 @@
+pytest \
+    -vv \
+    --cov=cosmos \
+    --cov-report=term-missing \
+    --cov-report=xml \
+    --durations=0 \
+    -m "not (integration or perf)" \
+    --ignore=tests/perf \
+    --ignore=tests/test_example_dags.py \
+    --ignore=tests/test_example_dags_no_connections.py

scripts/test/unit.sh

@@ -0,0 +1,7 @@
+pytest \
+    -vv \
+    --durations=0 \
+    -m "not (integration or perf)" \
+    --ignore=tests/perf \
+    --ignore=tests/test_example_dags.py \
+    --ignore=tests/test_example_dags_no_connections.py