perf: improve performance of build_cellset_from_pandas_dataframe

[Kevin-Dekker]

• Use dict(zip(keys, values)) instead of a for loop

• Use itertuples to make each row until last column of the dataframe into a tuple

Efficiency Statistics (1000 iterations)

Larger dataframe:

Comparison in markdown tables:

Current (before commit/merge)

	length: 1000 width: 3	length: 1000 width: 13	length: 5000 width: 3	length: 5000 width: 13	length: 15000 width: 3	length: 15000 width: 13
mean	0.0049	0.0179	0.0477	0.1101	0.1464	0.3459
std	0.0059	0.0079	0.0166	0.026	0.0308	0.0602
max	0.0472	0.0642	0.1056	0.2145	0.2607	0.6604

Suggested (after commit/merge)

	length: 1000 width: 3	length: 1000 width: 13	length: 5000 width: 3	length: 5000 width: 13	length: 15000 width: 3	length: 15000 width: 13
mean	0.0016	0.0069	0.0157	0.0425	0.0473	0.1294
std	0.0036	0.0046	0.0053	0.0121	0.0123	0.0269
max	0.0167	0.0282	0.038	0.0934	0.094	0.2652

[MariusWirtz]

Nice work @Kevin-Dekker. Especially the larger TM1py implementations will benefit from this.

I think we may even want to create a minor release on PyPI after the merge.