Cyclebane

About

Transform directed acyclic graphs using map-reduce and groupby operations

This library is an attempt to merge the concepts of directed acyclic graphs (DAG) with array-like objects such as NumPy arrays, Pandas DataFrames, or Xarray/Scipp DataArrays. This could be useful for describing tasks graphs, e.g., when a series of tasks is applied to chunks of an array. These tasks also have an array structure. After an reduction operation of chunks, the graph loses this structure, i.e., only a subset of the graph's nodes has array structure. What if we could work with this structure, even though only parts of the graph follows it? And what if we could use the power of array slicing with named dimensions, or select by label? This is what Cyclebane tries to do.

Our initial goal is to support:

map operations of a DAG's source nodes over an array-like (https://docs.dask.org/en/latest/high-level-graphs.html). Cyclebane will effectively copy all descendants of those nodes, once for each array element. Cyclebane will support joint mappings of multiple source nodes by mapping over, e.g., a DataFrame with multiple columns, as well as chaining independent map operations at different source nodes. In the latter case this will effectively broadcast at descendant nodes that depend on multiple such source nodes.
reduce operations at descendants of mapped nodes. This will add a new node with edges to all copies of the mapped node being reduced. Cyclebane will support reducing only individual axes or all axes, similar to Numpy.
groupby operations similar to Pandas and Xarray (albeit more limited).
Positional and label-based indexing. Cyclebane will support selecting branches that were creating during map (or groupby) operations based on their indices. The graph structure will be left untouched, i.e., nodes after a reduce operation will be preserved, but fewer edges will lead to the reduce node.

See also Dask's High Level Graphs for a related concept (without the direct support for any such operations).

Installation

python -m pip install cyclebane

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

cyclebane-26.9.0.tar.gz (54.6 kB view details)

Uploaded Sep 2, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

cyclebane-26.9.0-py3-none-any.whl (17.3 kB view details)

Uploaded Sep 2, 2026 Python 3

File details

Details for the file cyclebane-26.9.0.tar.gz.

File metadata

Download URL: cyclebane-26.9.0.tar.gz
Upload date: Sep 2, 2026
Size: 54.6 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.13

File hashes

Hashes for cyclebane-26.9.0.tar.gz
Algorithm	Hash digest
SHA256	`60ad6bf20cf0e4fd0217c3f56aaf83ed8e90e14497dca5875f84a4b81cab4c1a`
MD5	`a8e68ad4c1ed3f1a5615cdbc43d6372c`
BLAKE2b-256	`b085ee7312ffdce098b26afa03b54a81b2ecaddd53fb2149208f790fed6c1e79`

See more details on using hashes here.

Provenance

The following attestation bundles were made for cyclebane-26.9.0.tar.gz:

Publisher: release.yml on scipp/cyclebane

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: cyclebane-26.9.0.tar.gz
- Subject digest: 60ad6bf20cf0e4fd0217c3f56aaf83ed8e90e14497dca5875f84a4b81cab4c1a
- Sigstore transparency entry: 2686952496
- Sigstore integration time: Sep 2, 2026
Source repository:
- Permalink: scipp/cyclebane@b7e215216f83852d1ba02ea88836016ec838a743
- Branch / Tag: refs/tags/26.9.0
- Owner: https://github.com/scipp
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: release.yml@b7e215216f83852d1ba02ea88836016ec838a743
- Trigger Event: release