Peter Harrison

@alecpm implemented a couple of websocket chatroom demos in PsyNet: - https://gitlab.com/PsyNetDev/PsyNet/-/merge_requests/997 - https://gitlab.com/PsyNetDev/PsyNet/-/merge_requests/1016 The task here is to create playwright tests for these demos. UPDATE: I now already implemented...

Extend `psynet performance test` to consider export: (5 pts)

2

## Overview Currently the performance test focuses on HTTP request duration. This issue proposes additionally considering the export process. ## Context PsyNet has a custom export mechanism (`psynet export`) that...

The [asv](https://asv.readthedocs.io/en/stable/index.html) package illustrates a neat idea: you define certain tasks whose performance you care about, you measure performance on each pushed commit, and then display the performance profile over...

See outcome of discovery in #7802.

## Context PsyNet uses various JSON columns, in particular `vars`, `definition`, `answer` (are there others?). `vars` is used as a dynamic variable store: ```py participant.var.x = 123 ``` All three...

## Description Unify bot completion with the same post-submission worker-event path used by human participants by removing the dedicated bot completion event path. ## Motivation and Context Bots previously emitted...

Dallinger docker-ssh CI testing

1

## Description This PR adds end-to-end `dallinger docker-ssh` coverage in CI and hardens related deployment paths. ### What changed 1. **New `docker-ssh-smoke` CI job** in `.github/workflows/ci.yml`. - Runs pytest-based integration...

I propose removing the few dependencies required by the `ec2` extra. Here's an autogenerated scoping: ## Current state - In this branch, `ec2` currently includes `numpy` and `pandas` only (no...

## Overview Currently the performance test focuses on HTTP request duration. This issue proposes additionally considering async processes (see the `AsyncProcess` class in PsyNet). ## Suggested approach Inspect the `process`...

One signature of an overloaded experiment is that participants take far too few trials (because PsyNet skips them to the end of the experiment, because no trials are available for...