feat(bigtable): Rerouted RowSet and RowRange to use ReadRows from data client

[daniel-sanche]

Migrating over @gkevinzheng PR from bigtable monorepo googleapis/python-bigtable#1296

Original description:

Changes made:

• Refactored _MappableAttributesMixin to helpers.py for use in RowRange.
• Used RowRange and ReadRowsQuery from the data client as backing data sources for RowRange and RowSet respectively.

Note to reviewers: This PR has already been reviewed and merged to a staging branch, with the intention of doing a single merge to main. We are now planning to slowly rollout these changes back to the main branch. Minimal re-review should be necessary

[gemini-code-assist]

Code Review

This pull request refactors RowSet and RowRange to use ReadRowsQuery and BaseRowRange from the data client, and extracts _MappableAttributesMixin into a shared helper module. The review feedback suggests addressing potential infinite recursion in _MappableAttributesMixin's magic methods and avoiding fragile private attribute access in RowRange.get_range_kwargs by using public properties.

[gemini-code-assist]

Accessing self._attribute_map inside __getattr__ can lead to infinite recursion (maximum recursion depth exceeded) if _attribute_map is not defined on the class or instance, or if it is accessed before initialization. To prevent this, add a guard to check if the requested attribute name is _attribute_map and raise an AttributeError immediately. Similarly, guard __setattr__ to handle _attribute_map safely.

Suggested change

	def __getattr__(self, name):
	if name not in self._attribute_map:
	raise AttributeError
	return getattr(self, self._attribute_map[name])
	def __setattr__(self, name, value):
	attribute = self._attribute_map.get(name, name)
	super(_MappableAttributesMixin, self).__setattr__(attribute, value)
	def __getattr__(self, name):
	if name == "_attribute_map":
	raise AttributeError
	if name not in self._attribute_map:
	raise AttributeError
	return getattr(self, self._attribute_map[name])
	def __setattr__(self, name, value):
	if name == "_attribute_map":
	super(_MappableAttributesMixin, self).__setattr__(name, value)
	return
	attribute = self._attribute_map.get(name, name)
	super(_MappableAttributesMixin, self).__setattr__(attribute, value)

[gemini-code-assist]

Accessing the private attributes self._pb._pb is fragile and violates encapsulation. Since RowRange inherits from BaseRowRange (which exposes public properties start_key, end_key, start_is_inclusive, and end_is_inclusive), we can construct the dictionary using these public properties. This is safer, more robust, and avoids relying on the internal protobuf implementation details.

Suggested change

	def get_range_kwargs(self):
	"""Convert row range object to dict which can be passed to
	google.bigtable.v2.RowRange add method.
	"""
	range_kwargs = {}
	if self.start_key is not None:
	start_key_key = "start_key_open"
	if self.start_inclusive:
	start_key_key = "start_key_closed"
	range_kwargs[start_key_key] = _to_bytes(self.start_key)
	if self.end_key is not None:
	end_key_key = "end_key_open"
	if self.end_inclusive:
	end_key_key = "end_key_closed"
	range_kwargs[end_key_key] = _to_bytes(self.end_key)
	return range_kwargs
	return {
	descriptor.name: value for descriptor, value in self._pb._pb.ListFields()
	}
	def get_range_kwargs(self):
	"""Convert row range object to dict which can be passed to
	google.bigtable.v2.RowRange add method.
	"""
	range_kwargs = {}
	if self.start_key is not None:
	key = "start_key_closed" if self.start_is_inclusive else "start_key_open"
	range_kwargs[key] = self.start_key
	if self.end_key is not None:
	key = "end_key_closed" if self.end_is_inclusive else "end_key_open"
	range_kwargs[key] = self.end_key
	return range_kwargs