Take care of deprecation warnings from dependencies

[timmens]

In this PR, we take care of deprecation warnings from estimagic's dependencies.

In particular, we

• Call df.map() instead of df.applymap() if pandas version 2.1.0 or greater is installed
• Add new module compat.py, in which we implement functions with compatibility issues
• Ignore deprecation warning about MemoizeJac, which is raised in cyipopt. It seems that there are working on this as well (see here)
• Use preferred np.prod instead of np.product, which is deprecated with numpy version 1.25.0 (see release notes here. In earlier versions of numpy np.prod was already implemented and np.product would simply call np.prod; therefore we can just swap np.product for np.prod. (There were some subtle differences, but they do not apply to our use cases; see here)
• Use pd.Series(list).unique() instead of pd.unique(list), as the latter ist deprecated. Also use pd.Series(list).value_counts() instead of pd.value_counts(list).
• Remove maxiter option from scipy_truncated_newton algorithm, since it is not a valid option.
• Use DataFrame.replace() properly for categorical columns. The old implementation raised the following errors:

  estimagic/tests/inference/test_bootstrap.py:70: FutureWarning: Downcasting behavior in `replace` is deprecated and 
  will be removed in a future version. To retain the old behavior, explicitly call `result.infer_objects(copy=False)`. To opt-in 
  to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`
  df = raw.replace({"time": replacements})
  
  estimagic/tests/inference/test_bootstrap.py:70: FutureWarning: The behavior of Series.replace (and 
  DataFrame.replace) with CategoricalDtype is deprecated. In a future version, replace will only be used for cases that 
  preserve the categories. To change the categories, use ser.cat.rename_categories instead.
  df = raw.replace({"time": replacements})

• Remove usage of DataFrameGroupBy.grouper, due to following warning:

   estimagic/src/estimagic/parameters/consolidate_constraints.py:595: FutureWarning: DataFrameGroupBy.grouper is deprecated and will be removed in a future version of pandas.
    weights["dupl_group"] = weights.groupby(list(weights.columns)).grouper.group_info[0]

  Current solution:

  weights["dupl_group"] = weights.groupby(list(weights.columns)).ngroup()

• Set dtype to float before overwriting certain values with np.inf, due to the following warning:

  estimagic/src/estimagic/visualization/profile_plot.py:169: Setting an item of incompatible dtype is deprecated and will raise an error in a future version of pandas. Value 'inf' has dtype incompatible with int64,  please explicitly cast to a compatible dtype first.
  solution_times[~converged_info] = np.inf

  Current solution:
  • Cast data frame to float for all columns, then assign np.inf.
  • Previously all columns where we assigned np.inf where casted to float anyways, resulting in a data frame with some float and some int columns. Now they are all float.
• Conversion of an array with ndim > 0 to a scalar is deprecated. Specifically, this occurs in
  • tranquilo (see issue #15 in tranquilo)
  • simopt strong optimize (todo: what to do about simopt)
    Current solution: Ignore the warning.
• For CI run certain tests with pandas version < 2.0.0
  • Select tests we want to run
  • Create new CI task
  • Add new environment to .envs
• Add Python 3.12 to linux test matrix

---

Closes #470.

[codecov]

Codecov Report

Attention: Patch coverage is 96.55172% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 92.99%. Comparing base (3fba00e) to head (50c2c24).

Files	Patch %	Lines
src/estimagic/compat.py	83.33%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #482      +/-   ##
==========================================
+ Coverage   92.98%   92.99%   +0.01%     
==========================================
  Files         193      194       +1     
  Lines       14638    14649      +11     
==========================================
+ Hits        13611    13623      +12     
+ Misses       1027     1026       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[hmgaudecker]

Looks great to me -- just a couple of small remarks!