`timeout` related fixes and error propogation improvements by amruthesht · Pull Request #120 · Becksteinlab/imdclient

amruthesht · 2025-11-19T01:30:40Z

Fixes #111

Changes made in this Pull Request:

timeout defualt value chnaged to 600
Imporvements to exception chaining and error propogation
test-cases for timeout
warning for timeout <= 1

PR Checklist

Tests?
Docs?
CHANGELOG updated?
Issue raised/referenced?

codecov · 2025-11-19T01:39:08Z

Codecov Report

❌ Patch coverage is 52.94118% with 16 lines in your changes missing coverage. Please review.
✅ Project coverage is 79.50%. Comparing base (73d54b7) to head (51dd426).

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

orbeckst

👍 on the timeout=600.

I have technical comments on the tests, please see inline.

@HeydenLabASU do you want to review and give your blessing?

@ljwoods2 do you have comments?

imdclient/tests/test_imdclient.py

orbeckst · 2025-11-19T17:15:35Z

imdclient/tests/test_imdclient.py

+        server = InThreadIMDServer(universe.trajectory)
+        server.set_imdsessioninfo(imdsinfo)
+        server.handshake_sequence("localhost", first_frame=False)
+        client = IMDClient(
+            f"localhost",
+            server.port,
+            universe.trajectory.n_atoms,
+            timeout=timeout_val,
+        )
+        server.join_accept_thread()
+
+        # Sleep for less than timeout before sending frames
+        time.sleep(timeout_val - 1)
+        server.send_frames(0)


This code is repeated - does it make sense to turn it into a fixture or at least a method that can be used by multiple tests?

Yes, I can make a function that does the "server setup and connection" part of the code

orbeckst · 2025-11-19T17:16:01Z

imdclient/tests/test_imdclient.py

+        server = InThreadIMDServer(universe.trajectory)
+        server.set_imdsessioninfo(imdsinfo)
+        server.handshake_sequence("localhost", first_frame=False)
+        client = IMDClient(
+            f"localhost",
+            server.port,
+            universe.trajectory.n_atoms,
+            timeout=timeout_val,
+        )
+        server.join_accept_thread()
+
+        # Sleep for longer than timeout without sending any frames
+        time.sleep(timeout_val + 1)


repeated code

ljwoods2 · 2025-11-19T17:57:13Z

@orbeckst super backed up atm, will need ~1 week to get to this. apologies, feel free to move forward without me if you're hoping to push a release

orbeckst · 2025-11-21T16:03:27Z

@amruthesht please resolve the conflict that's holding up the CI. Thanks.

orbeckst

The refactoring of the client/server in the tests looks good to me (and all tests continue to pass).

Coverage is low because you added many more cases for exceptions and we're not testing all these different fail cases separately. There also appears to be some code (InThread client?) that is not really tested at all (?)

I don't think it's worthwhile right now to create test cases for each separate failure case (unless it's easy to do, e.g. with mocking) but we should be clear what is untested code. Could you please

check the [coverage]((https://app.codecov.io/gh/Becksteinlab/imdclient/pull/120) for the code.
mark untested code blocks with #pragma: nocover
raise issues for larger code blocks that should be tested eventually

orbeckst · 2025-11-21T23:08:06Z

imdclient/tests/server.py

+        logger.debug(
+            f"InThreadIMDServer: Listening on {host}:{self._bound_port}"
+        )


Did black want to reformat this??

orbeckst · 2025-11-21T23:09:51Z

imdclient/IMDClient.py

-            raise EOFError
+            raise EOFError("Consumer has finished")

        return self._empty_q.get()


This return is never tested according to coverage. Do you know why and does it matter?

orbeckst · 2025-11-21T23:15:44Z

imdclient/IMDClient.py

            try:
                return self._producer._get_imdframe()
-            except EOFError:
+            except EOFError as e:
+                logger.debug(f"IMDClient: Single-threaded connection ended")
                self._disconnect()
-                raise EOFError
+                raise EOFError from e


This block is never tested. Do we not test the single-threaded client?

I think we should do test it at least with a simple test. Otherwise we're advertising code for which we don't even know if it runs correctly.

orbeckst · 2025-11-22T00:14:10Z

imdclient/IMDClient.py

        try:
            self._parse_imdframe()
        except EOFError as e:
-            raise EOFError
+            logger.debug(f"IMDProducer: No more frames to read: {e}")
+            raise EOFError from e
        except Exception as e:
            raise RuntimeError("An unexpected error occurred") from e


None of these lines are covered.

amruthesht self-assigned this Nov 19, 2025

amruthesht requested review from ljwoods2 and orbeckst November 19, 2025 04:35

amruthesht force-pushed the socket-timeout branch from 079696d to 0596e06 Compare November 19, 2025 07:41

amruthesht marked this pull request as ready for review November 19, 2025 07:44

orbeckst requested changes Nov 19, 2025

View reviewed changes

orbeckst requested a review from HeydenLabASU November 19, 2025 17:17

amruthesht requested a review from orbeckst November 21, 2025 08:31

amruthesht added 7 commits November 21, 2025 13:05

Update: timeout default

35b2c86

Update: Error propogation and exception chaining improvements

c470322

Add: small timeout warning message and Black formatting

f51a86a

tests for timeout

3fa4551

Remove: unnecessary yields

1b810c8

Add: more error logging in queue

5068f24

Overhaul: tests with fixture factory for server_client

51dd426

amruthesht force-pushed the socket-timeout branch from 00b7edb to 51dd426 Compare November 21, 2025 20:09

orbeckst requested changes Nov 22, 2025

View reviewed changes

Conversation

amruthesht commented Nov 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Checklist

Uh oh!

codecov bot commented Nov 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

orbeckst left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

orbeckst Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

amruthesht Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

orbeckst Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

ljwoods2 commented Nov 19, 2025

Uh oh!

orbeckst commented Nov 21, 2025

Uh oh!

orbeckst left a comment

Choose a reason for hiding this comment

Uh oh!

orbeckst Nov 21, 2025

Choose a reason for hiding this comment

Uh oh!

orbeckst Nov 21, 2025

Choose a reason for hiding this comment

Uh oh!

orbeckst Nov 21, 2025

Choose a reason for hiding this comment

Uh oh!

orbeckst Nov 22, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

amruthesht commented Nov 19, 2025 •

edited

Loading

codecov bot commented Nov 19, 2025 •

edited

Loading