gh-154744: Improve detection of skipinitialspace in csv.Sniffer

[serhiy-storchaka]

skipinitialspace is now detected by parsing the sample both with and without it and comparing the two readings, instead of testing whether every field which follows a delimiter starts with a space.

• If the readings differ in the number of fields, the one which splits the sample into rows with the more consistent widths wins. This is what recognizes the space before a quoted field: a quote is only a quote at the very start of a field, so not skipping the space splits the quoted field.
• Otherwise a field differs between the readings only if a space was actually skipped at its start. This tells the padding apart from the spaces inside quoted or escaped fields, which are never skipped and therefore say nothing about the format.
• The first field of a row is judged separately, because it is commonly not padded (a, b, c).

Measured on the 560 files of the CSVsniffer corpora, skipinitialspace changes the number of fields in 7 of them, and in 3 of those the correct reading collapses the rows to a single width.

• Issue: csv.Sniffer misdetects skipinitialspace #154744