feat(rust): Trim leading and trailing whitespaces while parsing csv data

[xgillard]

I often find myself in the situation where I want to parse csv data which have been vertically aligned in the text file (because it makes the files prettier thus easier to read in terminal). However, these leading and trailing spaces come in the way of properly parsing numeric values with polars. This is why I propose to add a switch that lets us configure a way to drop these leading and trailing spaces.

Example: Parsing the following data would not work before edit but it could be accepted with this PR

INSTANCE     | STATUS | UB   | LB                   | DURATION | PEAK-MEMORY
05items/005 | Proved   | 1471 | 1470.998622 | 0                 | 0.004846898839
05items/033 | Proved   | 1104 | 1103.999        | 0                 | 0.004846531898
05items/024 | Proved   | 1308 | 1307.998815 | 0                 | 0.004889146425
05items/018 | Proved   |  1315 | 1314.998748 | 0                 | 0.004864221439
05items/006 | Proved   | 1386 | 1385.901602 | 0                 | 0.004862984642
05items/038 | Proved   |  1113 | 1112.999091  | 0                 | 0.004888778552
05items/034 | Proved   | 1108 | 1107.999018  | 0                 | 0.004848460667
05items/030 | Proved   |  588 | 587.9995413 | 0                 | 0.004298817366
05items/010 | Proved   | 1952 | 1951.998264 | 0                 | 0.004889526404

[ricglz]

@xgillard you have not updated this branch. I think you updated another one in your fork

[xgillard]

ouch... I'm so sorry for wasting your time on this one :-/

[ritchie46]

Thanks for the PR. Can you expose those options also to the python api and the lazy scanner?

[xgillard]

thx for your reply, I'll do it 👍

[orlp]

We've decided that this is not a feature we want, for more details see https://github.com/pola-rs/polars/pull/10333#issuecomment-1669291968.