Skip to content

Diophantos first paper release for Zenodo

Pre-release
Pre-release

Choose a tag to compare

View all tags
@B0Gec B0Gec released this 22 Jan 12:11
· 292 commits to oeis since this release
v1.0.1_d2024_02_26
8aa70dd

Description

This is the first GitHub release of Diofantos, corresponding to the Zenodo repository from my paper.

However, after paper revision, new release Diophantos - published revised paper release [updated Readme] is available, so I encourage its use instead.

How to set up Diofantos

To reproduce results, one could use container as an alternative to installing Python dependencies listed below.

Otherwise, go ahead and install the dependencies in a new python environment.

After that, you can clone the repository, but I recommend getting only (instead of all 7GB results) the following files:

exact_ed.py
diophantine_solver.py
doones.py
linear_database_newbl.csv
cores_test.csv
sindy_oeis.py

The list above is not checked so please make sure there are no import errors.

Get essential files via git

I find it easiest to use git to make an efficient clone to automatically download the essential files to try out the method. Run these commands in terminal:

git clone -n --depth=1 --filter=tree:0 https://github.com/B0Gec/Diofantos
cd Diofantos
git restore --source cf66eec6d45f6cbc0bc5d0b2da0361332a811484 exact_ed.py diophantine_solver.py doones.py cores_test.csv sindy_oeis.py

In the end download linear_database_newbl.csv manually from Zenodo repository (look in the paper), since it is stored as git lfs (large files) and they seem to be hard to download as a single file.

And ignore files under the GitHub "Assets" section of the GitHub's release page.

Simple example of Diofantos' execution in terminal:

python doones.py --task_id 14 --exper_id output_dir

will produce the output file results/output_dir/00014_A000045.txt with similar content:

While total time consumed by now, scale:15/164, seq_id:A000045, order:10 took:
 0.6 seconds, i.e. 0.01 minutes or 0.0 hours.
CORELIST True, SINDy False, GROUND_TRUTH False, SINDy_default True
Library: n, max_order 10, threshold: 0.1

by degree: 1 and order: 2. 
A000045: 
a(n) = a(n - 2) + a(n - 1)
truth: 


  -  checked against website ground truth.     
True  -  "manual" check if equation is correct.

Apptainer/Singularity container:

  • Results from paper can be reproduced by running the doones.py file from python from the Singularity container obtained
    from the Singularity Hub in the following way:
  • apptainer remote add --no-login SylabsCloud cloud.sycloud.io
  • singularity remote use SylabsCloud
  • singularity pull library://bogec/diofantos/oeis:latest
  • run e.g.: ~/ProGED_oeis$ singularity exec oeis_latest.sif python3 doones.py --task_id 13 --exper_id reproduced_experiment

Experiments

  • database of linrec sequences: linear_database_newbl.csv
  • database of core sequences: cores_test.csv
  • script for running Diofantos and SINDy-based approaches: doones.py
  • Diofantos code: exact_ed.py
  • SINDy based approaches: sindy_oeis.py
  • Results: directories results and results_oeis

Features

  • algebraic equations with variables n, a(n-k) for all k up to chosen order and
    their combinations up to degree d.

Dependencies

Usage examples of script doones.py in first version (26.2.2024)

  • Diofantos for Fibonacci sequence in core database:
    python doones.py --task_id 14 --exper_id output_dir
Assets 2
Loading