Use the Arrow C and PyCapsule data interfaces to share data with Python #98
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
See: - https://arrow.apache.org/docs/format/CDataInterface.html - https://arrow.apache.org/docs/format/CDataInterface/PyCapsuleInterface.html This basically allows us to do the following: - Statically link any version of Arrow into _hictkpy. This is used by hictk to return interactions through a `std::share_ptr<arrow::Table>`. - Use the Arrow C data interface (`<arrow/c/abi.h>`) and C bridge (`<arrow/c/bridge.h>`) to export data generated with the C++ API (which is not ABI stable) through the C ABI (which is stable). - Wrap the exported data using `PyCapsule`s. We don't even need to depend on nanoarrow to do this, as all we need to do is return Python objects that expose the `PyCapsule`s through the `__arrow_c_schema__` or `__arrow_c_stream__` attributes. - Use pyarrow's Python API to construct a `pyarrow.Table` given the `ArrowSchema` and `ArrowArrayStream` objects returned by the C data interface. As a bonus, we no longer have to depend on pyarrow (we only need it when fetching pixels or bins as DataFrames). The only downside of this approach is that pyarrow versions <16 are not supported (because `pyarrow.Table.from_arrays()` does not recognize objects exposing the `__arrow_c_stream__` attribute.
robomics
added
enhancement
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
See:
This basically allows us to do the following:
std::share_ptr<arrow::Table>.<arrow/c/abi.h>) and C bridge (<arrow/c/bridge.h>) to export data generated with the C++ API (which is not ABI stable) through the C ABI (which is stable).PyCapsules. We don't even need to depend on nanoarrow to do this, as all we need to do is return Python objects that expose thePyCapsules through the__arrow_c_schema__or__arrow_c_stream__attributes.pyarrow.Tablegiven theArrowSchemaandArrowArrayStreamobjects returned by the C data interface.As a bonus, we no longer have to depend on pyarrow (we only need it when fetching pixels or bins as DataFrames).
The only downside of this approach is that pyarrow versions <16 are not supported (because
pyarrow.Table.from_arrays()does not recognize objects exposing the__arrow_c_stream__attribute).Closes #91.