[MNT] Intermediate test plan

.github/workflows/test.yml

@@ -101,6 +101,12 @@ jobs:
         echo "BEFORE=$git_status" >> $GITHUB_ENV
         echo "Repository status before tests: $git_status"
 
+    - name: Checkout server-api and patch Docker path
+      shell: bash
+      run: |
+        git clone --depth 1 https://github.com/openml/server-api.git server-api
+        sed -i 's|\.\./server-api|./server-api|g' docker-compose.yml
+
     - name: Show installed dependencies
       run: python -m pip list
 
@@ -117,7 +123,7 @@ jobs:
           marks="not production and not uses_test_server"
         fi
 
-        pytest -n 4 --durations=20 --dist load -sv $codecov -o log_cli=true -m "$marks"
+        pytest --durations=20 --dist load -sv $codecov -o log_cli=true -m "$marks"
 
     - name: Run tests on Ubuntu Production
       if: matrix.os == 'ubuntu-latest'
@@ -132,12 +138,19 @@ jobs:
           marks="production and not uses_test_server"
         fi
 
-        pytest -n 4 --durations=20 --dist load -sv $codecov -o log_cli=true -m "$marks"
+        pytest --durations=20 --dist load -sv $codecov -o log_cli=true -m "$marks"
 
     - name: Run tests on Windows
       if: matrix.os == 'windows-latest'
       run: |  # we need a separate step because of the bash-specific if-statement in the previous one.
-        pytest -n 4 --durations=20 --dist load -sv --reruns 5 --reruns-delay 1 -m "not uses_test_server"
+        pytest --durations=20 --dist load -sv --reruns 5 --reruns-delay 1 -m "not uses_test_server"
+
+    - name: Cleanup Docker setup
+      if: always()
+      shell: bash
+      run: |
+        rm -rf server-api
+        git checkout docker-compose.yml
 
     - name: Check for files left behind by test
       if: matrix.os != 'windows-latest' && always()

docker-compose.yml

@@ -0,0 +1,52 @@
+services:
+  database:
+    image: "openml/test-database:20240105"
+    container_name: "openml-test-db-ci"
+    environment:
+      MYSQL_ROOT_PASSWORD: ok
+    ports:
+      - "33060:3306"
+    healthcheck:
+      test: ["CMD", "mysqladmin" ,"ping", "-h", "localhost"]
+      start_period: 30s
+      interval: 5s
+      retries: 10
+
+  database-setup:
+    image: mysql
+    container_name: "openml-test-setup-ci"
+    volumes:
+      - ./docker/update.sh:/database-update.sh
+    command: /bin/sh -c "/database-update.sh"
+    depends_on:
+      database:
+        condition: service_healthy
+
+# V1 API (PHP)
+  php-api:
+    image: "openml/php-rest-api:v1.2.2"
+    container_name: "openml-php-api-ci"
+    ports:
+      - "9002:80"
+    depends_on:
+      database:
+        condition: service_started
+    environment:
+      - DB_HOST_OPENML=database:3306
+      - DB_HOST_EXPDB=database:3306
+      - BASE_URL=http://localhost:9002/
+      - INDEX_ES_DURING_STARTUP=false
+
+  # V2 API (PYTHON)
+  python-api:
+    container_name: "openml-python-api-ci"
+    build:
+          # TODO: replace with image when available
+          context: ../server-api
+          dockerfile: docker/python/Dockerfile
+    ports:
+      - "9001:8000"
+    depends_on:
+      - database
+    environment:
+      - DATABASE_URL=mysql://root:ok@database:3306/openml

docker/update.sh

@@ -0,0 +1,31 @@
+#/bin/bash
+# Change the filepath of openml.file
+# from "https://www.openml.org/data/download/1666876/phpFsFYVN"
+# to "http://minio:9000/datasets/0000/0001/phpFsFYVN"
+mysql -hdatabase -uroot -pok -e 'UPDATE openml.file SET filepath = CONCAT("http://minio:9000/datasets/0000/", LPAD(id, 4, "0"), "/", SUBSTRING_INDEX(filepath, "/", -1)) WHERE extension="arff";'
+
+# Update openml.expdb.dataset with the same url
+mysql -hdatabase -uroot -pok -e 'UPDATE openml_expdb.dataset DS, openml.file FL SET DS.url = FL.filepath WHERE DS.did = FL.id;'
+
+
+
+
+
+# Create the data_feature_description TABLE. TODO: can we make sure this table exists already?
+mysql -hdatabase -uroot -pok -Dopenml_expdb -e 'CREATE TABLE IF NOT EXISTS `data_feature_description` (
+  `did` int unsigned NOT NULL,
+  `index` int unsigned NOT NULL,
+  `uploader` mediumint unsigned NOT NULL,
+  `date` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
+  `description_type` enum("plain", "ontology") NOT NULL,
+  `value` varchar(256) NOT NULL,
+  KEY `did` (`did`,`index`),
+  CONSTRAINT `data_feature_description_ibfk_1` FOREIGN KEY (`did`, `index`) REFERENCES `data_feature` (`did`, `index`) ON DELETE CASCADE ON UPDATE CASCADE
+)'
+
+# SET dataset 1 to active (used in unittests java)
+mysql -hdatabase -uroot -pok -Dopenml_expdb -e 'INSERT IGNORE INTO dataset_status VALUES (1, "active", "2024-01-01 00:00:00", 1)'
+mysql -hdatabase -uroot -pok -Dopenml_expdb -e 'DELETE FROM dataset_status WHERE did = 2 AND status = "deactivated";'
+
+# Temporary fix in case the database missed the kaggle table. The PHP Rest API expects the table to be there, while indexing.
+mysql -hdatabase -uroot -pok -Dopenml_expdb -e 'CREATE TABLE IF NOT EXISTS `kaggle` (`dataset_id` int(11) DEFAULT NULL, `kaggle_link` varchar(500) DEFAULT NULL)'

tests/conftest.py

@@ -24,6 +24,9 @@
 from __future__ import annotations
 
 import multiprocessing
+import sys
+
+import fasteners
 
 multiprocessing.set_start_method("spawn", force=True)
 
@@ -35,6 +38,9 @@
 import pytest
 import openml_sklearn
 
+import time
+import subprocess
+import requests
 import openml
 from openml.testing import TestBase
 
@@ -296,6 +302,38 @@ def with_test_cache(test_files_directory, request):
     if tmp_cache.exists():
         shutil.rmtree(tmp_cache)
 
+def _is_server_responding():
+    """Check if the Docker API is already listening."""
+    try:
+        requests.get("http://localhost:9001/api/v2/", timeout=1)
+        return True
+    except (requests.exceptions.ConnectionError, requests.exceptions.Timeout):
+        return False
+
+def _start_docker():
+    """Logic to spin up the containers and wait for initialization."""
+    subprocess.run(["docker", "compose", "up", "-d"], check=True, capture_output=True, text=True)
+    subprocess.run(["docker", "wait", "openml-test-setup-ci"], check=True)
+
+@pytest.fixture(scope="session", autouse=True)
+def openml_docker_stack(tmp_path_factory, worker_id):
+    # For local development with single worker
+    if worker_id == "master":
+        _start_docker()
+        yield
+        subprocess.run(["docker", "compose", "down", "-v"], check=True)
+        return
+
+    # For CI with multiple workers (xdist)
+    root_tmp_dir = tmp_path_factory.getbasetemp().parent
+    lock_file = root_tmp_dir / "docker_setup.lock"
+
+    lock = fasteners.InterProcessLock(str(lock_file))
+    with lock:
+        if not _is_server_responding():
+            _start_docker()
+
+    yield
 
 @pytest.fixture
 def static_cache_dir():