hdlib 2.0.0__tar.gz → 2.1.0__tar.gz

{hdlib-2.0.0 → hdlib-2.1.0}/PKG-INFO

@@ -1,11 +1,10 @@
 Metadata-Version: 2.4
 Name: hdlib
-Version: 2.0.0
+Version: 2.1.0
 Summary: Hyperdimensional Computing Library for building Vector Symbolic Architectures in Python
-Home-page: http://github.com/cumbof/hdlib
-Author: Fabio Cumbo
-Author-email: fabio.cumbo@gmail.com
+Author-email: Fabio Cumbo <fabio.cumbo@gmail.com>
 License: MIT
+Project-URL: Homepage, https://github.com/cumbof/hdlib
 Project-URL: Issues, https://github.com/cumbof/hdlib/issues
 Project-URL: Source, https://github.com/cumbof/hdlib
 Project-URL: Wiki, https://github.com/cumbof/hdlib/wiki
@@ -14,31 +13,36 @@ Classifier: Intended Audience :: Information Technology
 Classifier: Intended Audience :: Science/Research
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Scientific/Engineering
-Requires-Python: >=3
+Requires-Python: >=3.11
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: numpy>=2.2.5
-Requires-Dist: scikit-learn>=1.6.1
+Requires-Dist: mthree>=3.0.0
+Requires-Dist: numpy>=2.3.4
+Requires-Dist: qiskit>=2.2.1
+Requires-Dist: qiskit-aer>=0.17.2
+Requires-Dist: qiskit-ibm-runtime>=0.42.0
+Requires-Dist: scikit-learn>=1.7.2
+Requires-Dist: scipy>=1.16.2
 Requires-Dist: tabulate>=0.9.0
-Dynamic: author
-Dynamic: author-email
-Dynamic: classifier
-Dynamic: description
-Dynamic: description-content-type
-Dynamic: home-page
-Dynamic: license
+Provides-Extra: examples
+Requires-Dist: qiskit_machine_learning>=0.8.4; extra == "examples"
+Provides-Extra: test
+Requires-Dist: pytest>=8; extra == "test"
+Provides-Extra: dev
+Requires-Dist: build>=1; extra == "dev"
+Requires-Dist: twine>=5; extra == "dev"
+Requires-Dist: pytest>=8; extra == "dev"
 Dynamic: license-file
-Dynamic: project-url
-Dynamic: requires-dist
-Dynamic: requires-python
-Dynamic: summary
 
 # hdlib
 
 Hyperdimensional Computing Library for building Vector-Symbolic Architectures in Python 3.
 
-![Conda](https://img.shields.io/conda/dn/conda-forge/hdlib?label=hdlib%20on%20Conda)
+![Conda](https://img.shields.io/conda/dn/conda-forge/hdlib?label=hdlib%20in%20Conda)
 [![DOI](https://joss.theoj.org/papers/10.21105/joss.05704/status.svg)](https://doi.org/10.21105/joss.05704)
 [![DOI](https://zenodo.org/badge/485488487.svg)](https://doi.org/10.5281/zenodo.7996502)
 
@@ -67,6 +71,14 @@ Here is the table of content:
   - [Clustering](https://github.com/cumbof/hdlib/wiki/Examples#clustering)
 - [Support and contributions](https://github.com/cumbof/hdlib/wiki/Support-and-contributions)
 
+## Agent Skills
+
+Want to use `hdlib` with an LLM coding assistant? The [**hdlib-skills**](https://github.com/cumbof/hdlib-skills) repository provides a collection of [Agent Skills](https://github.com/cumbof/hdlib-skills) that teach LLM agents how to use the library effectively.
+
+The skills cover foundational concepts (vectors, space, arithmetic, and distance operations), machine learning models (classification, clustering, regression, graph encoding, and feature selection), quantum hyperdimensional computing, and common usage patterns (analogical reasoning, data encoding, and troubleshooting).
+
+They are compatible with any tool that supports the Agent Skills standard. See the [hdlib-skills](https://github.com/cumbof/hdlib-skills) repository for installation and usage instructions.
+
 ## Credits
 
 Please credit our work in your manuscript by citing:
@@ -82,6 +94,16 @@ Please credit our work in your manuscript by citing:
   year    = {2023},
   doi     = {10.21105/joss.05704}
 }
+
+@misc{cumbo2026hdlib,
+  title         = {hdlib 2.0: Extending Machine Learning Capabilities of Vector-Symbolic Architectures}, 
+  author        = {Fabio Cumbo and Kabir Dhillon and Daniel Blankenberg},
+  year          = {2026},
+  eprint        = {2601.02509},
+  archivePrefix = {arXiv},
+  primaryClass  = {cs.LG},
+  url           = {https://arxiv.org/abs/2601.02509}
+}
 ```
 
 ## Other publications
@@ -96,7 +118,13 @@ Please credit our work in your manuscript by citing:
 >
 > Cumbo et al., (2025). Hyperdimensional computing in biomedical sciences: a brief review. _PeerJ Computer Science_, 11, e2885. https://doi.org/10.7717/peerj-cs.2885
 >
-> Cumbo et al., (2025). A novel Vector-Symbolic Architecture for graph encoding and its application to viral pangenome-based species classification. _bioRxiv_, 2025-09. https://doi.org/10.1101/2025.09.08.674958
+> Cumbo et al., (2026). A novel Vector-Symbolic Architecture for graph encoding and its application to viral pangenome-based species classification. _BioData Mining_, 2026-05. https://doi.org/10.1186/s13040-026-00561-1
+>
+> Cumbo et al., (2026). Quantum Hyperdimensional Computing: a foundational paradigm for quantum neuromorphic architectures. _npj Unconventional Computing_, 3(1), 21. https://doi.org/10.1038/s44335-026-00064-6
+>
+> Cumbo et al., (2026). Designing vector-symbolic architectures for biomedical applications: ten tips and common pitfalls. _PeerJ Computer Science_, 12, e3682. https://doi.org/10.7717/peerj-cs.3682
+>
+> Cumbo et al., (2026). Predicting the toxicity of chemical compounds via Hyperdimensional Computing. _Molecular Informatics_, 45(9), e70052. https://doi.org/10.1002/minf.70052
 
 ## Support and contributions
 

hdlib-2.0.0/hdlib.egg-info/PKG-INFO → hdlib-2.1.0/README.md

@@ -1,44 +1,8 @@
-Metadata-Version: 2.4
-Name: hdlib
-Version: 2.0.0
-Summary: Hyperdimensional Computing Library for building Vector Symbolic Architectures in Python
-Home-page: http://github.com/cumbof/hdlib
-Author: Fabio Cumbo
-Author-email: fabio.cumbo@gmail.com
-License: MIT
-Project-URL: Issues, https://github.com/cumbof/hdlib/issues
-Project-URL: Source, https://github.com/cumbof/hdlib
-Project-URL: Wiki, https://github.com/cumbof/hdlib/wiki
-Classifier: Intended Audience :: Developers
-Classifier: Intended Audience :: Information Technology
-Classifier: Intended Audience :: Science/Research
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Programming Language :: Python :: 3
-Classifier: Topic :: Scientific/Engineering
-Requires-Python: >=3
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: numpy>=2.2.5
-Requires-Dist: scikit-learn>=1.6.1
-Requires-Dist: tabulate>=0.9.0
-Dynamic: author
-Dynamic: author-email
-Dynamic: classifier
-Dynamic: description
-Dynamic: description-content-type
-Dynamic: home-page
-Dynamic: license
-Dynamic: license-file
-Dynamic: project-url
-Dynamic: requires-dist
-Dynamic: requires-python
-Dynamic: summary
-
 # hdlib
 
 Hyperdimensional Computing Library for building Vector-Symbolic Architectures in Python 3.
 
-![Conda](https://img.shields.io/conda/dn/conda-forge/hdlib?label=hdlib%20on%20Conda)
+![Conda](https://img.shields.io/conda/dn/conda-forge/hdlib?label=hdlib%20in%20Conda)
 [![DOI](https://joss.theoj.org/papers/10.21105/joss.05704/status.svg)](https://doi.org/10.21105/joss.05704)
 [![DOI](https://zenodo.org/badge/485488487.svg)](https://doi.org/10.5281/zenodo.7996502)
 
@@ -67,6 +31,14 @@ Here is the table of content:
   - [Clustering](https://github.com/cumbof/hdlib/wiki/Examples#clustering)
 - [Support and contributions](https://github.com/cumbof/hdlib/wiki/Support-and-contributions)
 
+## Agent Skills
+
+Want to use `hdlib` with an LLM coding assistant? The [**hdlib-skills**](https://github.com/cumbof/hdlib-skills) repository provides a collection of [Agent Skills](https://github.com/cumbof/hdlib-skills) that teach LLM agents how to use the library effectively.
+
+The skills cover foundational concepts (vectors, space, arithmetic, and distance operations), machine learning models (classification, clustering, regression, graph encoding, and feature selection), quantum hyperdimensional computing, and common usage patterns (analogical reasoning, data encoding, and troubleshooting).
+
+They are compatible with any tool that supports the Agent Skills standard. See the [hdlib-skills](https://github.com/cumbof/hdlib-skills) repository for installation and usage instructions.
+
 ## Credits
 
 Please credit our work in your manuscript by citing:
@@ -82,6 +54,16 @@ Please credit our work in your manuscript by citing:
   year    = {2023},
   doi     = {10.21105/joss.05704}
 }
+
+@misc{cumbo2026hdlib,
+  title         = {hdlib 2.0: Extending Machine Learning Capabilities of Vector-Symbolic Architectures}, 
+  author        = {Fabio Cumbo and Kabir Dhillon and Daniel Blankenberg},
+  year          = {2026},
+  eprint        = {2601.02509},
+  archivePrefix = {arXiv},
+  primaryClass  = {cs.LG},
+  url           = {https://arxiv.org/abs/2601.02509}
+}
 ```
 
 ## Other publications
@@ -96,7 +78,13 @@ Please credit our work in your manuscript by citing:
 >
 > Cumbo et al., (2025). Hyperdimensional computing in biomedical sciences: a brief review. _PeerJ Computer Science_, 11, e2885. https://doi.org/10.7717/peerj-cs.2885
 >
-> Cumbo et al., (2025). A novel Vector-Symbolic Architecture for graph encoding and its application to viral pangenome-based species classification. _bioRxiv_, 2025-09. https://doi.org/10.1101/2025.09.08.674958
+> Cumbo et al., (2026). A novel Vector-Symbolic Architecture for graph encoding and its application to viral pangenome-based species classification. _BioData Mining_, 2026-05. https://doi.org/10.1186/s13040-026-00561-1
+>
+> Cumbo et al., (2026). Quantum Hyperdimensional Computing: a foundational paradigm for quantum neuromorphic architectures. _npj Unconventional Computing_, 3(1), 21. https://doi.org/10.1038/s44335-026-00064-6
+>
+> Cumbo et al., (2026). Designing vector-symbolic architectures for biomedical applications: ten tips and common pitfalls. _PeerJ Computer Science_, 12, e3682. https://doi.org/10.7717/peerj-cs.3682
+>
+> Cumbo et al., (2026). Predicting the toxicity of chemical compounds via Hyperdimensional Computing. _Molecular Informatics_, 45(9), e70052. https://doi.org/10.1002/minf.70052
 
 ## Support and contributions
 

{hdlib-2.0.0 → hdlib-2.1.0}/examples/chopin2/chopin2.py

@@ -3,10 +3,11 @@
 
 __author__ = ("Fabio Cumbo (fabio.cumbo@gmail.com)")
 
-__version__ = "1.1.0"
-__date__ = "Jul 16, 2023"
+__version__ = "1.1.1"
+__date__ = "Nov 7, 2025"
 
 import argparse as ap
+import contextlib
 import errno
 import os
 import statistics
@@ -102,6 +103,12 @@ def read_params():
         default=1,
         help="Make it parallel when possible",
     )
+    p.add_argument(
+        "--output",
+        type=os.path.abspath,
+        default=None,
+        help="Path to the output file (full path). If not specified, output is printed to stdout",
+    )
     p.add_argument(
         "-v",
         "--version",
@@ -388,7 +395,7 @@ def chopin2():
 
     print("Class labels: {}\n".format(class_labels))
 
-    for fold, (y_indices, y_pred, retrainings, error_rate, _) in enumerate(predictions):
+    for fold, (y_indices, y_pred, _, retrainings, error_rate, _) in enumerate(predictions):
         # Produce the confusion matrix for each fold
         print("Fold {} ({} retrainings)".format(fold + 1, retrainings))
         retraining_iterations.append(retrainings)
@@ -447,9 +454,20 @@ def chopin2():
 
 
 if __name__ == "__main__":
-    t0 = time.time()
+    # Pre-parse --output only so redirection is set up before full argument parsing
+    _pre = ap.ArgumentParser(add_help=False)
+    _pre.add_argument("--output", default=None)
+    _output_arg, _ = _pre.parse_known_args()
 
-    chopin2()
+    t0 = time.time()
 
-    t1 = time.time()
-    print("Total elapsed time: {}s".format(int(t1 - t0)))
+    if _output_arg.output:
+        with open(_output_arg.output, "w") as _outfile, contextlib.redirect_stdout(_outfile):
+            chopin2()
+            t1 = time.time()
+            print("Total elapsed time: {}s".format(int(t1 - t0)))
+        print("Output written to: {}".format(_output_arg.output))
+    else:
+        chopin2()
+        t1 = time.time()
+        print("Total elapsed time: {}s".format(int(t1 - t0)))

{hdlib-2.0.0 → hdlib-2.1.0}/hdlib/__init__.py

@@ -12,5 +12,5 @@ modules and how to use the library."""
 
 __author__ = ("Fabio Cumbo (fabio.cumbo@gmail.com)")
 
-__version__ = "2.0.0"
-__date__ = "Sep 23, 2025"
+__version__ = "2.1.0"
+__date__ = "Mar 10, 2026"

hdlib-2.0.0/hdlib/arithmetic.py → hdlib-2.1.0/hdlib/arithmetic/__init__.py

@@ -1,4 +1,4 @@
-"""Implementation of the arithmetic operators.
+"""Implementation of the MAP arithmetic operators.
 
 This library also provides the same set of arithmetic functions also accessible as _Vector_'s class methods. 
 However, while the result of calling these functions from a _Vector_ object woule be applied in place, 
@@ -71,7 +71,13 @@ def bind(vector1: Vector, vector2: Vector) -> Vector:
     if vector1.vtype != vector2.vtype:
         raise Exception("Vector types are not compatible")
 
-    vector = vector1.vector * vector2.vector
+    if vector1.vtype == "bipolar":
+        # Element-wise multiplication
+        vector = vector1.vector * vector2.vector
+
+    elif vector1.vtype == "binary":
+        # Element-wise XOR
+        vector = vector1.vector.astype(int) ^ vector2.vector.astype(int)
 
     tags = set(vector1.tags).union(set(vector2.tags))
 
@@ -140,7 +146,13 @@ def bundle(vector1: Vector, vector2: Vector) -> Vector:
     if vector1.vtype != vector2.vtype:
         raise Exception("Vector types are not compatible")
 
-    vector = vector1.vector + vector2.vector
+    if vector1.vtype == "bipolar":
+        # Element-wise addition
+        vector = vector1.vector + vector2.vector
+
+    elif vector1.vtype == "binary":
+        # Element-wise majority vote
+        vector = ((vector1.vector + vector2.vector) > 1).astype(int)
 
     tags = set(vector1.tags).union(set(vector2.tags))
 
@@ -165,7 +177,8 @@ def subtraction(vector1: Vector, vector2: Vector) -> Vector:
     Raises
     ------
     Exception
-        If vectors have different sizes or different vector types.
+        - If vectors have different sizes or different vector types;
+        - If vectors type is binary.
 
     Examples
     --------
@@ -201,7 +214,12 @@ def subtraction(vector1: Vector, vector2: Vector) -> Vector:
     if vector1.vtype != vector2.vtype:
         raise Exception("Vector types are not compatible")
 
-    vector = vector1.vector - vector2.vector
+    if vector1.vtype == "bipolar":
+        # Element-wise subtraction
+        vector = vector1.vector - vector2.vector
+
+    elif vector1.vtype == "binary":
+        raise Exception("Subtraction is not available for binary vectors")
 
     return Vector(size=vector1.size, vector=vector, tags=vector1.tags, vtype=vector1.vtype, seed=vector1.seed)