OP dataset generator (for use in Kaggle Dataset publishing)

[comcon1]

• Generate OP dataset from experiments.
• Generate OP dataset from simulations.

The PR is a shorter version of the all-dataset version #476. It's much shorter for review.
This PR: NMRLipids/BilayerData#388 already uses this branch.

It has additional dependencies: tables and unidecode, and it has very simple usage:

(Databank) comcon1@tanteliss:~/repo/Databank/developer$ ./gen-op-dataset.py --help
fairmd-lipids 1.5.1.dev4+g37cd713e8 by NMRlipids Open Collaboration - GPL-3.0-or-later
FAIRMD Lipids is initialized from the folder: /home/comcon1/repo/Databank/Data
---------------------------------------------------------------
usage: gen-op-dataset.py [-h] [--exps] [--sims]

options:
  -h, --help  show this help message and exit
  --exps      Run experiments
  --sims      Run simulations

It will generate one or another dataset in HDF5 format.

---

📚 Documentation preview 📚: https://databank--491.org.readthedocs.build/

[batukav]

Might be worth fixing before merge is the dataframe preallocation in _prepare_df_common, which can leave NaN/-1 rows in the output when not every opdict entry has a SMILEIDX mapping. Beyond that there's a fair amount of duplication between the two store_to_hdf5 implementations that the abstract base could absorb. Minor cleanups noted inline.