HapMapII_GRCh38 have zero-padding instead of missing data on the right flank

[nspope]

MWE:

import stdpopsim

print("version:", stdpopsim.__version__)                                                                                                 

homsap = stdpopsim.get_species("HomSap")
contig = homsap.get_contig("chr21", genetic_map="HapMapII_GRCh38")
print(  # rate in first bin is NaN
    "First 3 rates/positions of chr21:\n", 
    contig.recombination_map.rate[:3], "\n",
    contig.recombination_map.position[:3], "\n",
)
print(  # rate in last bin is 0.0
    "Last 3 rate bins/positions of chr21:", "\n",
    contig.recombination_map.rate[-3:], "\n",
    contig.recombination_map.position[-3:], "\n",
)

engine = stdpopsim.get_engine("msprime")
demogr = stdpopsim.PiecewiseConstantSize(1000)
sample = {"pop_0": 1}

ts = engine.simulate(contig=contig, samples=sample, demographic_model=demogr)
print("First tree starts at:", ts.edges_left.min())  # left flank is missing
print("Last tree ends at:", ts.edges_right.max())  # ... but not on right flank

this is the output:

version: 0.3.0
First 3 rates/positions of chr21:
 [         nan 1.796830e-09 1.819887e-08] 
 [       0. 10326675. 10328257.]

Last 3 rate bins/positions of chr21: 
 [2.205359e-08 2.208893e-08 0.000000e+00] 
 [46677698. 46680243. 46709983.]

First tree starts at: 10326675.0
Last tree ends at: 46709983.0

Seems like the last bin should have NaN, not zero, here?

[nspope]

maybe related: tskit-dev/tskit#3252

[andrewkern]

hmmm. i thought we'd checked this once before?

[nspope]

I think it's come up before that the end of various p-arms had zero-padding rather than NaNs -- and that got fixed -- but I hadn't noticed this issue with the q-arms. Are you saying that the zero-recombination rate is intentional here, and doesn't represent missing data?