examples

Examples

Standalone, runnable scripts — each demonstrates a facet of hypervector computing with the published kongming-rs-hv package.

Prerequisite: pip install kongming-rs-hv (or pip install -r ../requirements.txt). Every script imports from kongming import hv. matplotlib is optional, and only for weighted_schedule.py --plot.

Example	What it shows
Mexican Dollar	Analogical reasoning — "What's the Dollar of Mexico?": bind/bundle as the math behind analogy.
Word Indexer	Encode 5,000 English words and query by positional suffix via multi-attractor NNS.
Bulk Storage	Populate in-memory / embedded substrates with thousands of chunks and measure retrieval.
Operators from Scratch	Reimplement bind and bundle in pure Python — the core math underneath the library.
LISP Interpreter	A full LISP where every atom, cons cell, and environment is a hypervector.
Learner Revisited	What a Learner holds: each segment is a weighted reservoir, and the write weights choose whether occupancy measures $P$, $\log P$, or $P/N$.

Running

python examples/mexican_dollar/mexican_dollar.py          # also: mexican_dollar_memory.py
python examples/word_indexer/word_indexer.py
python examples/bulk_storage/bulk_storage.py -n 10000     # --backend embedded, --model, ...
python examples/operators/operators.py
python examples/learner_revisited/learner_revisited.py   # then weighted_schedule.py; --quick for fewer trials
python examples/learner_revisited/weighted_schedule.py --plot /tmp/figs   # + figures, needs matplotlib
pytest examples/pylisp/                                    # LISP interpreter tests; see pylisp/README.md

Full walkthroughs for each example live in the docs: yangzh.github.io/hv/examples.