File suffix to handle up to 1e5 projections

Currently, output files are written with a file suffix of four digits ("%04d"). This is true for volume files (partial HDF5) and single-frame files (EDF, Tiff, JP2) using start_index

That said, we increasingly have to cope with datasets with more than 10k projections. Perhaps a workaround could be to patch tomoscan class in the beginning:

from tomoscan.esrf.volume import TIFFVolume, ...
TIFFVolume.DEFAULT_DATA_DATA_PATH_PATTERN = '{volume_basename}_{index_zfill5}.{data_extension}'

For the record, nabu can normally handle more than 1e4 projections when still using the %04d suffix:

def variable_idxlen_sort(fname):
    return int(fname.split("_")[-1].split(".")[0])


[...]

        if files is None:
            files = list(self.results.values())
        try:
            files.sort(key=variable_idxlen_sort)
        except:
            self.logger.error(
                "Lexical ordering failed, falling back to default sort - it will fail for more than 10k projections"
            )
            files.sort()