Warn users when simulating campaigns without an objective

[AVHopp]

Fixes #749 by warning users that there are campaign without an objective instead of crashing.

simulate_scenarios will now run for the case described above, but warn the users about the missing objective(s). The code from the original bug now produces the following warning

(baybe) mXXXXX@LTDE0243486 baybe % python bug_test.py
/Users/mXXXXX/Programming/BayesianBE/baybe/baybe/simulation/scenarios.py:215: UserWarning: The following scenario(s) have campaigns without an objective and will not track any targets during simulation: ['Random Search']. If this was not intended, make sure to define an objective for each campaign.

and the following result dataframe:

        Scenario  Random_Seed  Initial_Data  Iteration  Num_Experiments
0  Random Search         1337           NaN          0                1
1  Random Search         1337           NaN          1                2
2  Random Search         1337           NaN          2                3

When also using a campaign that actually has an objective, the resulting dataframe looks as follows:

         Scenario  Random_Seed  Initial_Data  Iteration  Num_Experiments t_Measurements  t_IterBest  t_CumBest
0   Random Search         1337           NaN          0                1            NaN         NaN        NaN
1   Random Search         1337           NaN          1                2            NaN         NaN        NaN
2   Random Search         1337           NaN          2                3            NaN         NaN        NaN
3  With Objective         1337           NaN          0                1          [0.5]         0.5        0.5
4  With Objective         1337           NaN          1                2          [0.5]         0.5        0.5
5  With Objective         1337           NaN          2                3          [0.5]         0.5        0.5

[Copilot]

Pull request overview

This PR fixes issue #749 by enabling simulate_scenarios to work with campaigns that don't have an objective (e.g., campaigns using RandomRecommender). The fix automatically propagates an objective from the first campaign that has one to any campaigns lacking an objective, allowing target tracking during simulation.

Changes:

• Modified simulate_scenarios to propagate objectives from campaigns that have them to campaigns that don't
• Added comprehensive tests for both successful objective propagation and failure when no objective exists
• Updated CHANGELOG to document the fix

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
baybe/simulation/scenarios.py	Added objective propagation logic using attrs.evolve() and updated docstring to document the new behavior
tests/test_simulation.py	Added two tests: one for successful simulation with mixed objective/no-objective campaigns, and one to verify proper error handling when all campaigns lack objectives. Also updated import from baybe.targets.numerical to baybe.targets
CHANGELOG.md	Documented the fix in the "Fixed" section

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Scienfitz]

i think the warning should be thrown in simulate_experiment where the original error was thrown. This place here can catch and count them and rethrow a combined warning