Add elastic search indexing as a separate service

README.md

@@ -47,8 +47,8 @@ You can use different profiles:
 - `"elasticsearch"`: databases + nginx + elasticsearch
 - `"rest-api"`: databases + nginx + elasticsearch + REST API
 - `"frontend"`: databases + nginx + elasticsearch + REST API + frontend + email-server
-- `"minio"`: databases + nginx + elasticsearch + REST APP + MinIO + parquet and croissant conversion
-- `"evaluation-engine"`: databases + nginx + elastichsearc + REST API + MinIO + evaluation engine
+- `"minio"`: databases + nginx + elasticsearch +  REST APP + MinIO + parquet and croissant conversion
+- `"evaluation-engine"`: databases + nginx + elasticsearch + REST API + MinIO + evaluation engine
 - `"all"`: everything
 
 Usage examples:
@@ -68,6 +68,11 @@ docker compose down
 ```
 While it is up, the API is available under `http://localhost:8082`, e.g., `http://localhost:8082/datasets/1`.
 
+Similarly, you can start the php-api directly without a profile to start it without elasticsearch:
+```bash
+docker compose up php-api -d
+docker compose down
+```
 
 ## Known issues
 See the Github Issue list for the known issues.
@@ -128,10 +133,6 @@ To test if dataset processing works, we can then run:
 ```
 where `DATASET_ID` is replaced by the identifier printed in the previous step.
 
-It is also possible to run the php-api without elasticsearch, this is useful to avoid the considerable overhead of starting the elasticsearch container and creating indices.
-First, ensure that `INDEX_ES_DURING_STARTUP=false` in `config/php/.env` otherwise it will try to connect to elasticsearch and fail.
-Then explicitly specify the compose file so that the `docker-compose.override.yaml` file will be ignored: `docker compose -f docker-compose.yaml php-api -d`.
-
 ### PHP, Parquet and Croissant converter
 If you want to do local development on containers that are part of the docker-compose, you want those containers to change based on your code. You should have the relevant code somewhere on your system. You will need to tell the docker-compose where to find it. You can do so by setting environment variables.
 

config/php/.env

@@ -11,4 +11,5 @@ DB_USER_EXPDB_READ=root
 DB_PASS_EXPDB_READ=ok
 ES_URL=elasticsearch:9200
 ES_PASSWORD=default
-INDEX_ES_DURING_STARTUP=true
+INDEX_ES=false
+ES_ENABLED=false

docker-compose.yaml

@@ -1,4 +1,8 @@
 services:
+  x-php-image: &php-image
+    image: openml/php-rest-api:v1.3.1
+    env_file: config/php/.env
+
   database:
     image: "openml/test-database:v0.1.20260204"
     container_name: "openml-test-database"
@@ -48,6 +52,19 @@ services:
           cpus: '0.2'
           memory: 250M
 
+  elasticsearch-indexer:
+    profiles: ["all", "minio", "rest-api", "frontend", "elasticsearch", "evaluation-engine"]
+    <<: *php-image
+    container_name: "openml-elasticsearch-indexer"
+    environment:
+      ES_ENABLED: true
+      INDEX_ES: true
+    depends_on:
+      elasticsearch:
+        condition: service_healthy
+      database-setup:
+        condition: service_completed_successfully
+
   nginx:
     profiles: ["all", "minio", "rest-api", "frontend", "elasticsearch", "evaluation-engine"]
     build:
@@ -62,13 +79,10 @@ services:
 
   php-api:
     profiles: ["all", "minio", "rest-api", "frontend", "evaluation-engine"]
-    image: openml/php-rest-api:v1.2.10
+    <<: *php-image
     container_name: "openml-php-rest-api"
     ports:
       - "8080:80"  # also known as /api (nginx)
-    env_file: config/php/.env
-    environment:
-      ES_ENABLED: false
     depends_on:
       database-setup:
         condition: service_completed_successfully
@@ -149,6 +163,8 @@ services:
     depends_on:
       php-api:
         condition: service_healthy
+      elasticsearch-indexer:
+        condition: service_completed_successfully
 
   minio:
     profiles: ["all", "minio", "rest-api", "evaluation-engine"]

scripts/wait_for_run.sh

@@ -11,7 +11,7 @@ ELAPSED=0
 while [ "$ELAPSED" -lt "$TIMEOUT" ]; do
   RESPONSE=$(curl -sS "http://localhost:8000/api/v1/xml/run/${RUN_ID}")
 
-  if echo "$RESPONSE" | grep -q "area_under_roc_curve"; then
+  if grep -q "area_under_roc_curve" <<< "$RESPONSE"; then
     echo "Run $RUN_ID has evaluations."
     exit 0
   fi