• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

quaquel / EMAworkbench / 43945984220
94%

Build:
DEFAULT BRANCH: master
Ran 16 Nov 2026 08:11PM UTC
Jobs 1
Files 81
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

16 Nov 2026 08:04PM UTC coverage: 94.265% (+0.05%) from 94.214%
43945984220

push

github

web-flow
Fix for handling shaped parameters correctly in nd (#499)

* Store shaped levers as one column per element in optimization archives

Since 3.0, levers can have a shape. Optimization already treated each
element of a shaped lever as a separate decision variable, but the
archives returned by optimize held a single column with an array for
each shaped lever. rebuild_platypus_population expected one column for
each decision variable, so epsilon_nondominated and the convergence
metrics failed on these archives, and the stored archives contained the
arrays as strings, because they are written as csv.

Store a shaped lever as one column for each element, named
<name>_<index>, as is already done for the experiments returned by
perform_experiments. Problem.latent_parameter_names gives these names.
Stored archives are numeric again, and archives from several seeds can
be merged and scored.

To turn rows back into Sample instances with arrays, e.g. to re-evaluate
the policies of an archive, from_experiments now takes an optional
parameters argument.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RbrUQUbjwdPmHx5YcBBzjc

* Pass parameters with a multidimensional shape to the model in that shape

The values of a parameter with a shape are sampled, and handled by
platypus, as one value for each element. sample_generator and
Sample._from_platypus_solution turned these into a flat array, without
restoring the shape. So, a parameter with shape (2, 2) reached the
model as an array with shape (4,), both with perform_experiments and
with optimize. One dimensional shapes were not affected.

Reshape the values to the shape of the parameter. The experiments still
have one column for each element.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RbrUQUbjwdPmHx5YcBBzjc

* Correct the documentation of the shape of parameters

The docstring of Parameter said that s... (continued)

124 of 124 new or added lines in 4 files covered. (100.0%)

10881 of 11543 relevant lines covered (94.26%)

0.94 hits per line

Jobs
ID Job ID Ran Files Coverage
1 43945984220.1 16 Nov 2026 08:11PM UTC 81
94.26
GitHub Action Run
Source Files on build 43945984220
  • Tree
  • List 81
  • Changed 5
  • Source Changed 5
  • Coverage Changed 5
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #43945984220
  • 56f07066 on github
  • Prev Build on master (#43927833321)
  • Next Build on master (#44060573428)
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc