Files
clickstream-ch-kafka-supers…/generator/tests/test_generator_control_dag_contract.py
T
ddadmin 080d5ae313 fix(generator): исключено смешение миров при запуске
- Зачем:
  - старый state и ручной live-генератор могли тихо писать события в новый мир данных.
- Что:
  - добавлена общая проверка чистого стенда для Airflow и host-путей.
  - защищены backfill, import, continue и clean-start сценарии live-generator.
  - описан миграционный отказ старого state.
- Проверка:
  - make generator-test; docker compose config --quiet; py_compile; bash -n.
2026-07-05 20:21:13 +03:00

137 lines
5.4 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""
Контракт DAG generator_control без запуска Airflow.
"""
import ast
from pathlib import Path
from clickstream_generator.launch import PROFILES
DAG_PATH = Path(__file__).parents[2] / "airflow" / "dags" / "generator_control_dag.py"
REPO_ROOT = Path(__file__).parents[2]
def _tree():
return ast.parse(DAG_PATH.read_text(encoding="utf-8"))
def test_dag_file_exists_and_uses_dynamic_profiles():
"""DAG берёт варианты профилей из PROFILES, а не из ручного списка."""
text = DAG_PATH.read_text(encoding="utf-8")
assert "dag_id=\"generator_control\"" in text
assert "sorted(PROFILES)" in text
for profile in PROFILES:
assert profile not in {"hardcoded-profile"}
def test_trigger_form_has_expected_param_enums():
"""Форма запуска ограничивает операции и профили."""
text = DAG_PATH.read_text(encoding="utf-8")
assert 'enum=["backfill", "import", "check"]' in text
assert "enum=sorted(PROFILES)" in text
assert '"duration": Param(' in text
assert '"artifact_path": Param(' in text
def test_dag_branches_and_waits_for_etl_completion():
"""Backfill/import запускают ETL и ждут его завершения перед check."""
text = DAG_PATH.read_text(encoding="utf-8")
assert "BranchPythonOperator" in text
assert "TriggerDagRunOperator" in text
assert 'trigger_dag_id="etl_pipeline"' in text
assert "wait_for_completion=True" in text
assert 'allowed_states=["success"]' in text
assert 'failed_states=["failed"]' in text
def test_no_docker_or_continue_operation_in_dag():
"""Пульт не управляет Docker и не содержит операцию continue."""
text = DAG_PATH.read_text(encoding="utf-8")
assert "docker" not in text.lower()
assert '"continue"' not in text
def test_compose_mounts_generator_code_without_socket_and_gates_live_service():
"""Airflow видит код генератора, но не получает Docker socket."""
text = (REPO_ROOT / "docker-compose.yml").read_text(encoding="utf-8")
assert "PYTHONPATH: /opt/airflow/generator_src" in text
assert "./generator/src:/opt/airflow/generator_src:ro" in text
assert "/var/run/docker.sock" not in text
assert "profiles:\n - live-generator" in text
def test_down_and_clean_include_live_generator_profile():
"""Остановка стенда удаляет профильный live-генератор."""
text = (REPO_ROOT / "Makefile").read_text(encoding="utf-8")
assert "down:\n\t$(COMPOSE) --profile live-generator down" in text
assert "clean:\n\t$(COMPOSE) --profile live-generator down -v --remove-orphans" in text
def test_console_generator_paths_precheck_clean_stand_before_writes():
"""Backfill/import с хоста проверяют чистый стенд до записи."""
run_generator = (REPO_ROOT / "scripts" / "run_generator.sh").read_text(
encoding="utf-8"
)
import_artifact = (
REPO_ROOT / "scripts" / "import_startup_history_artifact.sh"
).read_text(encoding="utf-8")
assert "bash \"${SCRIPT_DIR}/assert_stand_clean.sh\" clean" in run_generator
assert run_generator.index("assert_stand_clean.sh") < run_generator.index(
"run --rm --no-deps"
)
assert "bash \"${SCRIPT_DIR}/assert_stand_clean.sh\" clean" in import_artifact
assert import_artifact.index("assert_stand_clean.sh") < import_artifact.index(
"startup_history_artifact import"
)
def test_console_continue_checks_state_or_clean_stand_before_live_start():
"""Continue не стартует новый live-мир на грязном стенде без state."""
run_generator = (REPO_ROOT / "scripts" / "run_generator.sh").read_text(
encoding="utf-8"
)
assert "elif [[ \"${VERB}\" == \"continue\" ]]" in run_generator
assert "bash \"${SCRIPT_DIR}/assert_stand_clean.sh\" continue" in run_generator
assert run_generator.index("assert_stand_clean.sh\" continue") < run_generator.index(
"up -d --build generator"
)
def test_clean_start_paths_include_live_generator_profile():
"""Все чистые сбросы видят профильный live-генератор."""
generated_history = (
REPO_ROOT / "scripts" / "run_generated_history_analytics.sh"
).read_text(encoding="utf-8")
assert "--profile live-generator down -v --remove-orphans" in generated_history
def test_host_clean_precheck_checks_live_before_kafka_and_stg():
"""Host backfill/import сначала проверяет, что live-генератор не запущен."""
text = (REPO_ROOT / "scripts" / "assert_stand_clean.sh").read_text(
encoding="utf-8"
)
assert "--mode live" in text
assert text.index("--mode live") < text.index("clickhouse-client")
assert text.index("clickhouse-client") < text.index("--mode \"${CHECK_MODE}\"")
def test_airflow_and_generator_kafka_dependency_versions_match():
"""Airflow и генератор используют одну версию kafka-python."""
airflow_req = (REPO_ROOT / "airflow" / "requirements.txt").read_text(encoding="utf-8")
generator_req = (REPO_ROOT / "generator" / "requirements.txt").read_text(encoding="utf-8")
assert "kafka-python==2.0.6" in airflow_req
assert "kafka-python==2.0.6" in generator_req
assert "prometheus-client==0.21.1" in airflow_req