diff options
Diffstat (limited to 'python/README.md')
| -rw-r--r-- | python/README.md | 36 |
1 files changed, 36 insertions, 0 deletions
diff --git a/python/README.md b/python/README.md new file mode 100644 index 0000000..8a061c3 --- /dev/null +++ b/python/README.md @@ -0,0 +1,36 @@ +This repository is the source code accompanying the following papers. +- [An O(n) algorithm for generating uniform random vectors in n-dimensional cones](https://arxiv.org/abs/2101.00936) +- [An algorithm for estimating volumes and other integrals in n dimensions](https://arxiv.org/abs/2007.06808) + +However, most of the code for [An O(n) algorithm for generating +uniform random vectors in n-dimensional +cones](https://arxiv.org/abs/2101.00936) has been released as a +separate library—sambal. See [PyPI +page](https://pypi.org/project/sambal/) and +[source](https://git.systemreboot.net/sambal). + +# Dependencies + +The code depends on the following dependencies. Please install them +before running the scripts. A python virtual environment may be an +easy way to do this. + +- Python 3 +- matplotlib +- numpy +- scipy +- sambal + +# Run + +The core functions are in `nsmc.py`. The experiments are in the +scripts `distribution-volumes.py`, `distribution-integrals.py` and +`spheroid.py`. Run those scripts as +``` +$ python3 distribution-volumes.py +$ python3 distribution-integrals.py +$ python3 spheroid.py +``` +Plots of the number of samples versus the dimension get written to the +current directory. Play around with the number of trials, dimensions, +and relative error tolerances in the scripts. |