Add ability to customize BaseFeaturizer pbar by passing dict

[janosh]

I would like to set

tqdm(lst, position=0, leave=True)

so that slurm log files aren't filled with 1000s of pbar lines but just the last iteration's line showing the total wall time.

With this PR, that'll be possible by passing

df_features = featurizer.featurize_dataframe(
    df, input_col, pbar=dict(position=0, leave=True)
)

Old logs

...
MultipleFeaturizer: 100%|█████████▉| 12859/12874 [1:27:44<00:07,  2.14it/s]
MultipleFeaturizer: 100%|█████████▉| 12860/12874 [1:27:44<00:06,  2.04it/s]
MultipleFeaturizer: 100%|█████████▉| 12861/12874 [1:27:45<00:06,  2.00it/s]
MultipleFeaturizer: 100%|█████████▉| 12862/12874 [1:27:45<00:06,  1.97it/s]
MultipleFeaturizer: 100%|█████████▉| 12863/12874 [1:27:46<00:05,  1.95it/s]
MultipleFeaturizer: 100%|█████████▉| 12864/12874 [1:27:46<00:05,  1.94it/s]
MultipleFeaturizer: 100%|█████████▉| 12865/12874 [1:27:47<00:04,  1.93it/s]
MultipleFeaturizer: 100%|█████████▉| 12866/12874 [1:27:47<00:04,  1.92it/s]
MultipleFeaturizer: 100%|█████████▉| 12867/12874 [1:27:48<00:03,  1.98it/s]
MultipleFeaturizer: 100%|█████████▉| 12868/12874 [1:27:48<00:03,  1.95it/s]
MultipleFeaturizer: 100%|█████████▉| 12869/12874 [1:27:49<00:02,  2.25it/s]
MultipleFeaturizer: 100%|█████████▉| 12870/12874 [1:27:52<00:05,  1.35s/it]
MultipleFeaturizer: 100%|█████████▉| 12871/12874 [1:27:53<00:03,  1.03s/it]
MultipleFeaturizer: 100%|█████████▉| 12872/12874 [1:27:53<00:01,  1.26it/s]
MultipleFeaturizer: 100%|█████████▉| 12873/12874 [1:27:53<00:00,  1.59it/s]
MultipleFeaturizer: 100%|██████████| 12874/12874 [1:27:53<00:00,  1.77it/s]
MultipleFeaturizer: 100%|██████████| 12874/12874 [1:27:53<00:00,  2.44it/s]

New logs

MultipleFeaturizer: 100%|██████████| 12874/12874 [1:27:53<00:00,  2.44it/s]

[ardunn]

Looks ok to me, think we can get most of the tests passing though?

[janosh]

I think they're all timing out after 6h. Not sure there's anything I can do about that. Don't think any of my changes would break existing tests.