Merge pull request #132 from jessemapel/unmangle

            point_spec.id = _set_pid(i)

            point_spec.type = g.iloc[0].pointType

            for attr, attrtype in self.point_attrs:

                if attr in g.columns:

                # Un-mangle common attribute names between points and measures

                df_attr = self.point_field_map.get(attr, attr)

                if df_attr in g.columns:

                    # As per protobuf docs for assigning to a repeated field.

                    if attr == 'aprioriCovar' or attr == 'adjustedCovar':

                        arr = g.iloc[0][attr]

                    if df_attr == 'aprioriCovar' or df_attr == 'adjustedCovar':

                        arr = g.iloc[0][df_attr]

                        if isinstance(arr, np.ndarray):

                            arr = arr.ravel().tolist()

                        point_spec.aprioriCovar.extend(arr)

                    else:

                        setattr(point_spec, attr, attrtype(g.iloc[0][attr]))

                        setattr(point_spec, attr, attrtype(g.iloc[0][df_attr]))

            # The reference index should always be the image with the lowest index

            point_spec.referenceIndex = 0

                measure_spec = point_spec.Measure()

                # For all of the attributes, set if they are an dict accessible attr of the obj.

                for attr, attrtype in self.measure_attrs:

                    if attr in g.columns:

                        setattr(measure_spec, attr, attrtype(m[attr]))

                measure_spec.serialnumber = m.serialnumber

                    # Un-mangle common attribute names between points and measures

                    df_attr = self.measure_field_map.get(attr, attr)

                    if df_attr in g.columns:

                        setattr(measure_spec, attr, attrtype(m[df_attr]))

                # ISIS pixels are centered on (0.5, 0.5). NDArrays are (0,0) based.

                measure_spec.sample = m['sample'] + 0.5

                measure_spec.line = m['line'] + 0.5

                measure_spec.type = m.measureType

                measure_iterable.append(measure_spec)

                self.nmeasures += 1