aelzenaar / bella Goto Github PK

Right now the examples are kind of messy. Write specific well-documented examples to showcase particular features, or at least document the existing ones clearer.

MWE:

from bella import slices

# Compute a whole bunch of level sets, and stick them into one pandas dataframe
depth = 20
p = 3
q = 6
elliptic_slice = slices.elliptic_exterior(p, q, depth)

print(elliptic_slice[43])

The polynomial solver fails to converge, and we see:

bella.slices.ConvergenceFailedException: Convergence failed at r/s = 2/3, polynomial 3.0 -
12.12435565298214105469212439054110656859963677667266439639064885616353111836160025956802330730074003·x +
15.0·x² -
5.196152422706631880582339024517617100828415761431141884167420938355799050726400111243438560271745727·x³

But:

>>> print(farey.farey_polynomial_classic(2,3,3,6)+2)
3.732050807568877293527446341505872366942805253810380628055806979451933016908800037081146186757248576 -
1.0·x +
1.732050807568877293527446341505872366942805253810380628055806979451933016908800037081146186757248576·x² -
1.0·x³

so the wrong Farey polynomial is computed by slices.elliptic_exterior. (It even fails to be monic!)

The following is supposed to be the translates of the 2/3 cusp group's peripheral discs by elements of the group...

Use the power of dask and/or pandas to parallelise limit set computations.

Implement computation of approximations of Riley slices following https://github.com/aelzenaar/riley/blob/main/riley.py.

One needs to find a good (fast) method for solving the polynomials produced by the (existing) bella.farey module.

• generator walking down r/s pleating ray from -10 to -2 (say) with a step either uniform in [-10,2] or in the moduli space (general for all polynomials)
• selection of correct cusp from roots of Farey polynomial
• continued fraction approximation of cusp groups
• add subclass CuspGroup of ClassicalRileyGroup

Currently riley.riley_slice_exterior is very slow. This is a factor of three variables:

• mp.dps = 100 (__init__.py)
• maxsteps = 500 (in farey.solve_polynomial)
• extraprec = 1000 (in farey.solve_polynomial)

First of all, we should have an internal mpmath context to avoid setting the dps globally. We should then choose maxsteps and extraprec to be minimal to ensure convergence. This could be done as follows:

1. Try with maxsteps = 20, extraprec = 20
2. If no convergence:
   1. Increase maxsteps by (say) 1.5x and extraprec by 2x
   2. try again
3. Store minimal working maxsteps and extraprec in global context and use them as starting point next time solve_polynomial is called.

Add functions to:

• construct regular polygons
• compute geometry of pairs of geodesics [B, section starting p.162]
• Find pencils fixed by mobius transforms and (inversely) find the mobius transform fixing a pencil

Add functions to:

• approximate arbitrary numbers with continued fractions
• convert continued fraction expansions to and from rational numbers

If the parameters p, q to RileyGroup are non-integral then an error should not occur.

• Use trX,trY,trXY not alpha and beta
• Remove all but numpy polynomial generators
• Add utility functions to convert <-> numpy and mpmath representations
• Add utility function to solve polynomials

GroupCache.subgroup() should take a list of words, i.e. [(0,), (3,),...]; instead it takes just matrices so is essentially useless.

RileyGroup should also expose a subgroup method that takes XY strings.

MPmath does not support using numeric types from the pyadic package.

In farey.py implement:

• function to check if a word is a conjugate of some generator
• cyclic permutation generator
• combination of the above to compute the two splittings of the Farey word producing peripheral groups
• update the peripherals.py example to use this.

• the "tracing" algorithm for quasi-Fuchsian group limit sets (box 25 of [IP]
• depth-first search

Add a method to GroupCache which can produce a set of isometric circles, similar in interface to GroupCache.coloured_limit_set().

Add back in this function removed in de0855b and add a new "left-to-right" mode. This is of interest in computing zooms of limit sets and cannot be done in the "fast" way so it is worth having the function back even though it is slow at "normal" limit sets.

@ariymarkowitz please copy the error message here thx

• unit test that 4*i lies on the 1/2 pleating ray and 37 lies on the 0/1 pleating ray and -43 lies on the 1/1 pleating ray (parabolic/parabolic case)
• unit test that riley_string * X * riley_string^-1 Y = farey_string
• unit test fixed points and words are equal to finite initial list
• verify trace of farey words equals value of farey polynomials

Use holoviews.Graph to plot

• Stern-Brocot tree
• Farey graph

• draw circles fast using isometric_circles.py method but tweaked to make thinner circles
• for non-dynamic groups add global switch (e.g. command line parameter) to swap between panel serve or png output

Traceback (most recent call last):
  File "/Users/amar630/anaconda3/envs/py3-11/lib/python3.11/site-packages/dask/backends.py", line 136, in wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/Users/amar630/anaconda3/envs/py3-11/lib/python3.11/site-packages/dask/dataframe/io/csv.py", line 760, in read
    return read_pandas(
           ^^^^^^^^^^^^
  File "/Users/amar630/anaconda3/envs/py3-11/lib/python3.11/site-packages/dask/dataframe/io/csv.py", line 533, in read_pandas
    raise OSError(f"{urlpath} resolved to no files")
OSError: atom/*.csv resolved to no files

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/Users/amar630/Downloads/bella-main/examples/atom.py", line 70, in <module>
    df = dd.read_csv("atom/*.csv")
         ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/amar630/anaconda3/envs/py3-11/lib/python3.11/site-packages/dask/backends.py", line 138, in wrapper
    raise type(e)(
OSError: An error occurred while calling the read_csv method registered to the pandas backend.

I tried uncommenting the commented code, but then I get an error that G is not defined.

I am running Python 3.11.