Table subclass with parameters

par_table.py

import numpy as np
from astropy.table import Table, Column
from astropy.table.ndarray_mixin import NdarrayMixin, NdarrayMixinInfo
from astropy.io.registry import UnifiedReadWriteMethod
from astropy.table.connect import TableRead, TableWrite
from astropy.table.info import serialize_method_as


def fmt_func(val):
    return f'{val["par"]:.2f} ({val["pmn"]:.2f}, {val["pmx"]:.2f})'


PAR_DTYPE = np.dtype([('par', 'f8'), ('pmn', 'f8'), ('pmx', 'f8')])


class ParTableRead(TableRead):
    def __call__(self, *args, **kwargs):
        out = super().__call__(*args, **kwargs)
        for col in out.columns.values():
            if col.dtype == PAR_DTYPE:
                col.info.format = fmt_func
        return out


class ParTableWrite(TableWrite):
    """Something like this SHOULD work, but there is something obscure and
    weird going on that makes this fail. So instead just override the
    write() method directly."""
    def __call__(self, *args, serialize_method=None, **kwargs):
        instance = self._instance
        par_cols = []
        try:
            for col in instance.columns.values():
                if col.info.format is fmt_func:
                    par_cols.append(col.info.name)
                    col.info.format = None
            return super().__call__(*args, **kwargs)
        finally:
            for par_col in par_cols:
                self[par_col].info.format = fmt_func


class ParTable(Table):
    read = UnifiedReadWriteMethod(ParTableRead)
    # write = UnifiedReadWriteMethod(ParTableWrite)

    def _convert_data_to_col(self, *args, **kwargs):
        col = super()._convert_data_to_col(*args, **kwargs)
        if col.dtype == PAR_DTYPE:
            col.info.format = fmt_func
        return col

    def write(self, *args, **kwargs):
        new_self = Table(self)
        for col in new_self.columns.values():
            if col.info.format is fmt_func:
                col.info.format = None
        return new_self.write(*args, **kwargs)


name = ['par1', 'par2']
a = np.array([(np.pi, 2, 4.5), (np.pi / 2, 1, 3.1)], dtype=PAR_DTYPE)
t = ParTable([name, a], names=['name', 'par'])
t.write('pars.ecsv', overwrite=True)
t2 = ParTable.read('pars.ecsv')