fix(solve): Fix planning_horizons in solve_network

[coroa]

Follow-up to #1537 .

Closes #1551 .

Changes proposed in this Pull Request

In perfect foresight, the planning_horizons param remains the list
of all planning horizons (to trigger re-runs).

The current planning horizons is always taken from the planning_horizons
wildcard and passed into extra_functionality as a partial kwarg.

Checklist

• I tested my contribution locally and it works as intended.
• Code and workflow changes are sufficiently documented.

[yechenyan]

In this PR, if CCL is added to the perfect forecast, errors will occur as shown in the following image：

This is because the network is missing planning_horizons.

config.perfect.yaml is configured as follows:

solving:
  solver:
    name: highs
    options: highs-simplex
  mem: 4000

  check_objective:
    enable: true
    expected_value: 1.3778753459e+10
  
  agg_p_nom_limits:
    agg_offwind: true
    include_existing: false
    file: data/agg_p_nom_minmax.csv
  
  constraints: 
    CCL: true

Run the command: snakemake make_summary_perfect --configfile config/test/config.perfect.yaml

I think it doesn't affect the merging of this PR; this is just a reminder that there are still issues when adding CCL to the perfect forecast.

[coroa]

@yechenyan you are right, i added now a very concrete error message, that CCL has not been implemented for perfect foresight, yet

[coroa]

the n.planning_horizons was actually an intermediate idea that leaked into the commit. sorry for that.