DAGE-14: Integration tests - working on Docker + Solr testing by Intrinsical-AI · Pull Request #172 · SeaseLtd/rated-ranking-evaluator

Intrinsical-AI · 2025-07-21T13:42:29Z

Relocated some files + cleaning (Docker Compose to project root folder; solr-init also)
Test stack added
Added tests/integration/docker-compose.yml (Solr 9.8.1 + solr-precreate testcore) to keep CI fast and isolated. Simpler than the original
Pytest fixtures (tests/conftest.py):

docker_compose_file – tells pytest-docker where the compose lives.
solr_url – waits for testcore to be responsive and returns the base URL.
seed_dataset – idempotently seeds dataset.json when the core is empty.

Integration tests (tests/integration/test_docker_solr.py)

See: https://sease.atlassian.net/browse/DAGE-14?focusedCommentId=16343

nicolo-rinaldi · 2025-07-21T14:04:29Z

+from src.search_engine.solr_search_engine import SolrSearchEngine
+
+def test_core_exists(solr_url):
+    r = requests.get(solr_url + "../admin/cores?action=STATUS&wt=json")


what is this ../?

solr_url # → "http://127.0.0.1:8983/solr/testcore/"

But we want to access: "http://127.0.0.1:8983/solr/admin/cores?action=STATUS"

Web directory traversal: .. it's the same as cd .. in Bash

We're going up 1 level in the path

why we want to access that url? I was assuming that we test if the specific testcore was in the collection

nicolo-rinaldi · 2025-07-21T14:05:08Z

Why do we need to change this? what are the benefits?

The last version was a bit heavy for the pytest (as we decided to do the migration to a specilized python library), I assume it's supposed to be the real docker-compose.yml for the application, in that case that version should live in the root folder, and we need to add a simpler version for the testing in the integration folder. That's what I've done. I'm sure we can discuss a better solution. Check:
https://sease.atlassian.net/jira/software/projects/DAGE/boards/84?selectedIssue=DAGE-14
and feel free to ping me on Slack!

nicolo-rinaldi · 2025-07-21T14:26:54Z

+
+def test_search_engine_fetch(solr_url):
+    engine = SolrSearchEngine(solr_url)
+    docs   = engine.fetch_for_query_generation(None, 3, ["title","body"])


I think we don't have "body" in our solr documents

I did not realize due to silent behaviour coming from engine.fetch_for_query_generation - that accept every string as field, even it's not registered in the index, and does not lof any warning message / error nor exception

reminder to change it

nseidan · 2025-07-23T14:03:53Z

+      context: solr-init
+    depends_on:
+      - solr
+    command: ["sh", "-c", "/rre-dataset-generator/solr-init/solr-init.sh"]


@Intrinsical-AI we shouldn't move the docker-compose file to a root folder, it's a test env; not a service we produce, it is coming from user side; also our project is not a docker-based service, it is just CLI, no need to put it into the root folder. cc @dantuzi

Hey Naz, hanks for the feedback! this is a research task. It's a proposal for using python-docker.

I had already tagged you in the Jira explaining it.
https://sease.atlassian.net/browse/DAGE-14?focusedCommentId=16413

Please, check the board before review! 👍

nseidan · 2025-07-23T14:05:41Z

-      - solr
-    command: ["sh", "-c", "/opt/rre-dataset-generator/solr-init.sh"]
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8983/solr/testcore/admin/ping"]


@Intrinsical-AI we don't need a healthcheck here, already handled in solr-init.sh

This research proposal suggest using python-docker, the one you mentioned during a call. if so, the checks of solr-init.sh will be moved to a proper pytests file that uses fixtures. The information it's here: https://sease.atlassian.net/browse/DAGE-14?focusedCommentId=16413

This is not yet accepted - it's the proposal on "how" to do Integration tests of Solr - Docker.

In general docker compose would be better to be on the root.

in general: it's a good practice to implement healthheck on the docker compose, even you "handle" it outside (too).

I guess we will have to discuss in detail, but if you can propose a solution to easy orchestrating all the tests, would be nice!

I would pick one way to do things and stick to that across all integration tests of all search engines. TBD

I agree. This is exploratory for the reasearch

nicolo-rinaldi

I definitely like this library to set up integration tests, but we need to remember that we need to find a common way to set up folders. I think this is a key to improve the process of building integration tests for the search engines supported.

nicolo-rinaldi · 2025-07-28T15:52:56Z

+
+def test_search_engine_fetch(solr_url):
+    engine = SolrSearchEngine(solr_url)
+    docs   = engine.fetch_for_query_generation(None, 3, ["title","body"])


reminder to change it

nicolo-rinaldi · 2025-07-28T15:54:41Z

-      - solr
-    command: ["sh", "-c", "/opt/rre-dataset-generator/solr-init.sh"]
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8983/solr/testcore/admin/ping"]


I would pick one way to do things and stick to that across all integration tests of all search engines. TBD

MrCabss69 added 2 commits July 21, 2025 15:40

DAGE-14: Integration tests - working on Docker + Solr testing

3f51d0a

Minor fixes - added pytests-docker package to pyproject.toml

45bd013

Intrinsical-AI requested review from dantuzi and nicolo-rinaldi July 21, 2025 13:47

nicolo-rinaldi reviewed Jul 21, 2025

View reviewed changes

fixed styling - added whitelines

ec65e68

Intrinsical-AI requested a review from nicolo-rinaldi July 21, 2025 15:07

nseidan reviewed Jul 23, 2025

View reviewed changes

nicolo-rinaldi reviewed Jul 28, 2025

View reviewed changes

Conversation

Intrinsical-AI commented Jul 21, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nicolo-rinaldi Jul 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Intrinsical-AI Jul 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Intrinsical-AI Jul 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nseidan Jul 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Intrinsical-AI Jul 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nseidan Jul 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Intrinsical-AI Jul 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nicolo-rinaldi left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

nicolo-rinaldi Jul 28, 2025 •

edited

Loading

Intrinsical-AI Jul 21, 2025 •

edited

Loading

Intrinsical-AI Jul 21, 2025 •

edited

Loading

nseidan Jul 23, 2025 •

edited

Loading

Intrinsical-AI Jul 24, 2025 •

edited

Loading

nseidan Jul 23, 2025 •

edited

Loading

Intrinsical-AI Jul 24, 2025 •

edited

Loading