Swan

Swan is a semantic modeling and query optimization framework for DuckDB. It lets you define database constraints, schemas, advanced algebraic properties (such as symmetry, transitivity, equivalence, and disjointness), and mathematical optimization problems (LP/MIP via the HiGHS solver) as an executable semantic layer.

Unlike passive metadata catalogs, Swan's semantic definitions are actively compiled and enforced:

Python Client-Side AST Rewrite Passes: Declarative PyRel queries are automatically analyzed and rewritten to prune logical contradictions, optimize optional matches, push down filters, swap symmetric traversals, and materialize shared query blocks.
C++ Native Optimizer Extension: Integrates directly with DuckDB's query compilation pipeline to dynamically inject check constraints, eliminate redundant joins, and replace expensive recursive joins with custom physical structures (like connected components and transitive reachability indexes).
Prescriptive Reasoner (HiGHS Integration): Formulates mathematical optimization problems directly over database records, compiling variables, bounds, and indicator constraints to DuckDB C++ queries for native execution speed.
Predictive Reasoner (GNN + ONNX Runtime): Trains Graph Neural Networks over graph-structured database records and serves predictions natively in DuckDB, with built-in explainability for feature/edge importance and attention weights.

By treating semantics as executable code, Swan compiles highly-declarative traversals into ultra-efficient, vectorized SQL executions.

Installation

pip install rai-swan

This automatically detects your operating system and CPU architecture and downloads the matching prebuilt wheel (Linux x86_64/ARM64, macOS Intel/Apple Silicon, Windows x64) -- no compiler required.

Verify the install (this also loads the bundled native extension):

from pyrel_duckdb import Model

model = Model()
print("Swan successfully initialized!")

Quick Start

import duckdb
from pyrel_duckdb import Model, String

con = duckdb.connect("my_database.db", config={"allow_unsigned_extensions": "true"})
model = Model(name="my_model", connection=con)

# Declare a concept (maps to a physical table) with an identity scheme (primary key)
Customer = model.Concept("Customer", identify_by={"name": String})

See the full documentation for modeling concepts/properties/relationships, rules and constraints, graph algorithms, and the prescriptive/predictive reasoners.

Documentation

Full guides, API reference, and build templates: https://relationalai.github.io/swan

License

Swan Proprietary Software License

Copyright (c) 2026 RelationalAI, Inc. All rights reserved.

1. Ownership. Swan (the "Software"), including its source code, compiled
   binaries, and documentation, is the confidential and proprietary
   property of RelationalAI, Inc. ("RelationalAI"). No rights to the
   Software are granted except as expressly set out below.

2. Limited License Grant. Subject to the terms of a separate written
   agreement between you and RelationalAI (or, absent such an agreement,
   solely for internal evaluation by authorized RelationalAI personnel and
   design partners), RelationalAI grants you a non-exclusive,
   non-transferable, non-sublicensable license to install and run the
   compiled distributions of the Software (e.g. the packages published to
   PyPI) for that authorized purpose.

3. Restrictions. Except as expressly permitted above, you may not: (a)
   copy, modify, or create derivative works of the Software; (b) reverse
   engineer, decompile, or disassemble any compiled binary; (c)
   redistribute, sublicense, rent, lease, or otherwise make the Software
   available to any third party; or (d) remove or alter any proprietary
   notices.

4. No Warranty. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY
   KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
   MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NONINFRINGEMENT.

5. Limitation of Liability. IN NO EVENT SHALL RELATIONALAI BE LIABLE FOR
   ANY CLAIM, DAMAGES, OR OTHER LIABILITY ARISING FROM, OUT OF, OR IN
   CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
   SOFTWARE.

6. Third-Party Components. The Software statically links certain
   third-party open-source components under their own permissive licenses;
   see THIRD_PARTY_LICENSES.md for the full list and required attributions.

Download files

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

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

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.

rai_swan-0.0.10-cp314-cp314-win_arm64.whl (17.0 MB view details)

Uploaded Aug 9, 2026 CPython 3.14Windows ARM64

rai_swan-0.0.10-cp314-cp314-win_amd64.whl (18.0 MB view details)

Uploaded Aug 9, 2026 CPython 3.14Windows x86-64

rai_swan-0.0.10-cp314-cp314-pyemscripten_2026_0_wasm32.whl (6.2 MB view details)

Uploaded Aug 9, 2026 CPython 3.14PyEmscripten 2026.0 wasm32

rai_swan-0.0.10-cp314-cp314-manylinux_2_28_x86_64.whl (19.6 MB view details)

Uploaded Aug 9, 2026 CPython 3.14manylinux: glibc 2.28+ x86-64

rai_swan-0.0.10-cp314-cp314-manylinux_2_28_aarch64.whl (18.0 MB view details)

Uploaded Aug 9, 2026 CPython 3.14manylinux: glibc 2.28+ ARM64

rai_swan-0.0.10-cp314-cp314-macosx_11_0_arm64.whl (14.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.14macOS 11.0+ ARM64

rai_swan-0.0.10-cp314-cp314-macosx_10_15_x86_64.whl (15.5 MB view details)

Uploaded Aug 9, 2026 CPython 3.14macOS 10.15+ x86-64

rai_swan-0.0.10-cp313-cp313-win_arm64.whl (16.1 MB view details)

Uploaded Aug 9, 2026 CPython 3.13Windows ARM64

rai_swan-0.0.10-cp313-cp313-win_amd64.whl (17.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.13Windows x86-64

rai_swan-0.0.10-cp313-cp313-manylinux_2_28_x86_64.whl (19.4 MB view details)

Uploaded Aug 9, 2026 CPython 3.13manylinux: glibc 2.28+ x86-64

rai_swan-0.0.10-cp313-cp313-manylinux_2_28_aarch64.whl (17.7 MB view details)

Uploaded Aug 9, 2026 CPython 3.13manylinux: glibc 2.28+ ARM64

rai_swan-0.0.10-cp313-cp313-macosx_11_0_arm64.whl (14.2 MB view details)

Uploaded Aug 9, 2026 CPython 3.13macOS 11.0+ ARM64

rai_swan-0.0.10-cp313-cp313-macosx_10_15_x86_64.whl (15.5 MB view details)

Uploaded Aug 9, 2026 CPython 3.13macOS 10.15+ x86-64

rai_swan-0.0.10-cp312-cp312-win_arm64.whl (16.1 MB view details)

Uploaded Aug 9, 2026 CPython 3.12Windows ARM64

rai_swan-0.0.10-cp312-cp312-win_amd64.whl (17.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.12Windows x86-64

rai_swan-0.0.10-cp312-cp312-manylinux_2_28_x86_64.whl (19.4 MB view details)

Uploaded Aug 9, 2026 CPython 3.12manylinux: glibc 2.28+ x86-64

rai_swan-0.0.10-cp312-cp312-manylinux_2_28_aarch64.whl (17.7 MB view details)

Uploaded Aug 9, 2026 CPython 3.12manylinux: glibc 2.28+ ARM64

rai_swan-0.0.10-cp312-cp312-macosx_11_0_arm64.whl (14.1 MB view details)

Uploaded Aug 9, 2026 CPython 3.12macOS 11.0+ ARM64

rai_swan-0.0.10-cp312-cp312-macosx_10_15_x86_64.whl (15.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.12macOS 10.15+ x86-64

rai_swan-0.0.10-cp311-cp311-win_arm64.whl (16.1 MB view details)

Uploaded Aug 9, 2026 CPython 3.11Windows ARM64

rai_swan-0.0.10-cp311-cp311-win_amd64.whl (17.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.11Windows x86-64

rai_swan-0.0.10-cp311-cp311-manylinux_2_28_x86_64.whl (19.1 MB view details)

Uploaded Aug 9, 2026 CPython 3.11manylinux: glibc 2.28+ x86-64

rai_swan-0.0.10-cp311-cp311-manylinux_2_28_aarch64.whl (17.4 MB view details)

Uploaded Aug 9, 2026 CPython 3.11manylinux: glibc 2.28+ ARM64

rai_swan-0.0.10-cp311-cp311-macosx_11_0_arm64.whl (14.0 MB view details)

Uploaded Aug 9, 2026 CPython 3.11macOS 11.0+ ARM64

rai_swan-0.0.10-cp311-cp311-macosx_10_15_x86_64.whl (15.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.11macOS 10.15+ x86-64

rai_swan-0.0.10-cp310-cp310-win_amd64.whl (17.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.10Windows x86-64

rai_swan-0.0.10-cp310-cp310-manylinux_2_28_x86_64.whl (18.9 MB view details)

Uploaded Aug 9, 2026 CPython 3.10manylinux: glibc 2.28+ x86-64

rai_swan-0.0.10-cp310-cp310-manylinux_2_28_aarch64.whl (17.3 MB view details)

Uploaded Aug 9, 2026 CPython 3.10manylinux: glibc 2.28+ ARM64

rai_swan-0.0.10-cp310-cp310-macosx_11_0_arm64.whl (14.0 MB view details)

Uploaded Aug 9, 2026 CPython 3.10macOS 11.0+ ARM64

rai_swan-0.0.10-cp310-cp310-macosx_10_15_x86_64.whl (15.2 MB view details)

Uploaded Aug 9, 2026 CPython 3.10macOS 10.15+ x86-64

File details

Details for the file rai_swan-0.0.10-cp314-cp314-win_arm64.whl.

File metadata

Download URL: rai_swan-0.0.10-cp314-cp314-win_arm64.whl
Upload date: Aug 9, 2026
Size: 17.0 MB
Tags: CPython 3.14, Windows ARM64
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.13

File hashes

Hashes for rai_swan-0.0.10-cp314-cp314-win_arm64.whl
Algorithm	Hash digest
SHA256	`ec00859cf5b6a7703dda599b2b297f5352c4a3a4d70af0bd3e37a3aadab1770d`
MD5	`ac87d62ac6b38f6b0ef0ce58a33f6e49`
BLAKE2b-256	`2f1db117ef8526a3614459ba52f070dc5d487779296b5d0fa8d446d6743b7c0e`

Algorithm	Hash digest
SHA256	`8c60a9f9021235dbc8c25a23aeeb8041d17bc6309ff57253f43837fd20bffdc4`
MD5	`9e008b08220a0d3176ca1463be890797`
BLAKE2b-256	`d1aca95340d446459bea0de434784b84d63d80af84940fe41c0a353c6190c619`

Algorithm	Hash digest
SHA256	`7869cd257af4f30190b37c869f32257e300d7d7239381c25d8142b559054519a`
MD5	`9141c5c86d972de6c6af56c3587ec27e`
BLAKE2b-256	`e127d7ba870265e050498047777adb91b641f3b977bf435e0463919522d5d771`

Algorithm	Hash digest
SHA256	`e72d2fa826fa39b635f9f15bc3b2829b9e53bbc65c361e5dcfe1d47192caf8aa`
MD5	`2dffffdb55cf4bd98e4392f4549f455a`
BLAKE2b-256	`fa35a98fe7087f0d9b92abbf8d676dce736e5055b1068da5984b88e79498d12f`

Algorithm	Hash digest
SHA256	`fc3ab5789825b28b5d3f9e4b02ee53e93a9bee7cba64a74869455103322258fa`
MD5	`84599d8ad0ea41a780f7313e23ff9f17`
BLAKE2b-256	`fa7ec8c212fba2661cc4a48f663790057cf59ff689e862f45544d727b7dc3d23`

Algorithm	Hash digest
SHA256	`4994596b0ef7a8e6afe87bf72d2b21957a2566b12a5410e9188b971665a33198`
MD5	`603b48b8ffddd9c1936d1802f9e575b3`
BLAKE2b-256	`bdd4dd83463dca3dfb0d61180977722dc4f3086b084e3c9f1c3c7e70da98db82`

Algorithm	Hash digest
SHA256	`d688618da81d49391a386982be93ebe117e4c8bf343f69a73535ed3bcbbe5882`
MD5	`a7424dcdf22f461dfc5cc8590c3bee7b`
BLAKE2b-256	`ee74f4ecb3d68f24b906616114850008753947af96e014b7a6f9953491165519`

Algorithm	Hash digest
SHA256	`97f991c0dbc646e1105730621ac3dc888927f19fe909e29f59c57cf30499c992`
MD5	`b7cc66e441cc8f59e032d438925fa3f1`
BLAKE2b-256	`52e3b4c589b31d7f795db8d80d85f886cc3e0d0ba99c7adab4c6ad685fce1ea6`

Algorithm	Hash digest
SHA256	`491d89570002ef27d5baa5065fdd5e151fa51b1e75355945086444413e33d5df`
MD5	`9ad5a8250270b3a412aabf8d3de26497`
BLAKE2b-256	`4c78fcd0d40c3f689ec16b4bfb1ee1eedd213645b9bf728d12336d54e66432f7`

Algorithm	Hash digest
SHA256	`0fad2bccbbb5c6453b851910a13fb09c7082a98d01a9ec73e9bc81c19eaa282e`
MD5	`1331b0f0f460caf1c87fac44f3152c34`
BLAKE2b-256	`8814485ee8940062b71ed76b878e42231aa1697e38dc1db1207e0d6c789fdcc7`

Algorithm	Hash digest
SHA256	`4bf40c86d6eee64fc43130efae70b1877f4ff6a9b497938e4af06fc6c51bbc67`
MD5	`fbb704b8de4f86f98d4156f07bc2c6aa`
BLAKE2b-256	`ab2f2a0e0381374916d0a4033fbe26d5bb283d3dd6f685039ec23e175d7d6b29`

Algorithm	Hash digest
SHA256	`ef01dc71beb605a4a9d1df2715f83d4d0e4d31529dfad72dd3ece67462f5f328`
MD5	`6af2d148e3df7fa3f7eb5c09efc536ba`
BLAKE2b-256	`b1a6d42d0d3aa980a09fecf0a3d8a761184daca8b94bf0a0aad27d0f3a35371e`

Algorithm	Hash digest
SHA256	`2dc1f67aa9b335a0307f117bf20c5186d36207cdf0afc1627ea226a5b989dde5`
MD5	`b0ba61ce5feed0cfb3b87e8cf7e85dc0`
BLAKE2b-256	`909edc44c3e07590829fbabb78dc90d6f1c89913ce818ef518683f36de284078`

Algorithm	Hash digest
SHA256	`d4fcb1c49a9b0bbd153378a51ddb790b4f2ffe8c8c6acd1971620bbe6c66f672`
MD5	`0193c83758c614da9d81537a8a351b21`
BLAKE2b-256	`a61b42fd603a52c0e6e8f190288e92c1b26d794bf9e7bde827457fb884e0222e`

Algorithm	Hash digest
SHA256	`c331794588f3af13bf80be26cfc7814938dc4d0c48e316b9dbdfe1e06f08c5b6`
MD5	`cda90d430f5615ce1cf7bf9efebfa5e3`
BLAKE2b-256	`b7b1308cf1eaedc8f225ed795e9b5fe6d2f7ab67dd04d16148ae9855affff3d9`

Algorithm	Hash digest
SHA256	`3c8ead539d50f3ef0be8e8bd7abcbe545b5e2d44e48ee5a13b3147abccc2ecb0`
MD5	`9b1ac7e5821a5fe3865a995784dc3ff2`
BLAKE2b-256	`408556da7e6d793c2c56ac03c4413e31783eb0350443e772f17b05a23e6e3141`

Algorithm	Hash digest
SHA256	`84e23346753214be1af9acdb2f41377a6af02f89a92b93af50b597a66385e36a`
MD5	`bd11976bfbc2a69722a849094c59e4da`
BLAKE2b-256	`65e9f4f587d4f9a1312d2293668d82a9cca07a028c80e732d64f061f77b2f66c`

Algorithm	Hash digest
SHA256	`3c6f184a52b6790396677ae14c680d8b393522bc16f62f59248243aeaa412739`
MD5	`60fcd4ca90d7388e4d1b022da2e2f0e2`
BLAKE2b-256	`4bbf3eb4ee43e9f821532085366a1fce3e1001f2356077ae8ea82cdb80f90439`

Algorithm	Hash digest
SHA256	`e9b5144788bbd3188d78bb041caf243e656daae933153596763cb3716e5dec5a`
MD5	`1123444dab0f13905a10c2fdab53bbb8`
BLAKE2b-256	`55ff8c93159411fc8408434358dd214ac0e05be06d855867171fbdb926620255`

Algorithm	Hash digest
SHA256	`84eed19679cffb7f2b0f8e815427a4a7151ca6ae57f848f57eb1f792e219ddd0`
MD5	`7326304f0e17d0ead674d7f388526a4c`
BLAKE2b-256	`0b40df8e54c11f7df613115b16f3d30c64f56a42e21f871ae18ad49b81d89658`

Algorithm	Hash digest
SHA256	`11f0d39f8ebc6854a6852539b72a11f9ae7214103cb3fcbc57c287433d2dbd41`
MD5	`048d382ba274b368f6437a8b3c219561`
BLAKE2b-256	`1a407d48d114535d9384f0540be86e182e30efb2b07852ea5d2a3c573f267e4e`

rai-swan 0.0.10

Swan

Installation

Quick Start

Documentation

License

Project links

Key dates

Owner

1 maintainer

GitHub Statistics

License

Requires

Provides Extra

Classifiers

Download files

Source Distributions

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

Algorithm	Hash digest
SHA256	`13b2ec008940f7f52b3bafa6d10658251881271bff67da8207d535d7243aa48a`
MD5	`4ceb09e22c84ae94137650be293c3bf9`
BLAKE2b-256	`6648880c318c46184139f2c0c2276983ad2d6f45c7f944a10f66df1fa086c648`

Algorithm	Hash digest
SHA256	`23f8cd9aeefb25590b5b701e0089184ad0d7fee3abf5335e5c7652067f537ca4`
MD5	`120e4edfc9c3b93f7078debace44a0e0`
BLAKE2b-256	`09eec7142e04fb883cc3ca62300b6e7c359132ad35e0c4de3db7d787a03e079f`

Algorithm	Hash digest
SHA256	`6013b3ee6acf8c29a87c817b824c6bfddac8511d55b9dcb8676408b9b424b782`
MD5	`5c7962646dcfa60ee9d01b64ba6dc9ab`
BLAKE2b-256	`0d0b92bf6f2b612d44b5c0f775e52cad594e0e85bf61c1ce39d446febc172d9b`

Algorithm	Hash digest
SHA256	`426481034ee80edeca97361a8bfe95a246067f3d7ad903bab72fb7c2e64d4401`
MD5	`e562d9f8b06e9cc0cc865f1e2fa67ba1`
BLAKE2b-256	`e3719305553fd75ab43fdb0dc5fb76ba68ce21fe5b30ee14e766ea39db843064`

Algorithm	Hash digest
SHA256	`8754017bf6d3be7d35d8e7b1e1c8fe4c5a2d8a767ef16461a99a267176869687`
MD5	`fc6eaf8ecc633ea7e712e1efc55c9449`
BLAKE2b-256	`ed24e7c0221f505c45a7b6e40a963801b4b89707fd77644a27c19f22e57d8d7d`

Algorithm	Hash digest
SHA256	`0356fb4b5549aa7065db029267ba08d9350c6094cbb73182ca39d64e044da796`
MD5	`040128ce573abfda0061955ca2b4e6bc`
BLAKE2b-256	`0348fcf3c4c7ddc3ab7749eb71eafd83beb39b836db322c925fedfa9c481168d`

Algorithm	Hash digest
SHA256	`829e0e6de925f270c652a20798c7c91da846ea2393e621908918cd02b6ce266a`
MD5	`0ac36bfd704a94d74532ca9c0051b524`
BLAKE2b-256	`220946976b734608c027eba000fb7abd336b4dc132408030f5b72a1f28ffbca3`

Algorithm	Hash digest
SHA256	`687827dba2f6525dc6c805de25a0b90e656b55589cc6a5a7ea88f36bdd3d9d6d`
MD5	`756ea23b4341629b3eb5ddb4d6a4aa71`
BLAKE2b-256	`5e9d2e84b464aedbb0df30a9a17e049072390dc00582c747779a0bdb8aa9b3a3`

Algorithm	Hash digest
SHA256	`08b114882eb4c5f33cf25bf316fc3718ca92a3948a5fb7869467cb77b97c83a0`
MD5	`62cf0b4889daa6f0702a7594ca40c401`
BLAKE2b-256	`b77773172043ab1468159ca911950e60d5c900c775051175de825e6ae2667588`