Fix error when analyzing CSVs with empty time metrics

[hbierlee]

If you analyze a CSV file with an empty time_* metric (which happens for me currently, only total_time is recorded for gurobi, which may be a different issue), the following error occurs:

> python cpmpy/tools/xcsp3/analyze.py results/expr.csv/xcsp3_2025_COP25_gurobi_20260416_185126.csv --time_limit 600  --output analysis/
Traceback (most recent call last):
  File "/cw/dtailocal/henk/Projects/cpmpy-master/cpmpy/tools/xcsp3/analyze.py", line 235, in <module>
    main()
  File "/cw/dtailocal/henk/Projects/cpmpy-master/cpmpy/tools/xcsp3/analyze.py", line 220, in main
    xcsp3_stats(merged_df)
  File "/cw/dtailocal/henk/Projects/cpmpy-master/cpmpy/tools/xcsp3/analyze.py", line 177, in xcsp3_stats
    slowest_idx = df[f'time_{phase}'].idxmax()
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/cw/dtailocal/henk/miniforge3/envs/cpmpy/lib/python3.12/site-packages/pandas/core/series.py", line 2545, in idxmax
    iloc = self.argmax(axis, skipna, *args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/cw/dtailocal/henk/miniforge3/envs/cpmpy/lib/python3.12/site-packages/pandas/core/base.py", line 814, in argmax
    result = nanops.nanargmax(delegate, skipna=skipna)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/cw/dtailocal/henk/miniforge3/envs/cpmpy/lib/python3.12/site-packages/pandas/core/nanops.py", line 1158, in nanargmax
    result = _maybe_arg_null_out(result, axis, mask, skipna)  # type: ignore[arg-type]
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/cw/dtailocal/henk/miniforge3/envs/cpmpy/lib/python3.12/site-packages/pandas/core/nanops.py", line 1473, in _maybe_arg_null_out
    raise ValueError("Encountered all NA values")
ValueError: Encountered all NA values

This PR fixes this error.