Anthony Wilder Wohns
Anthony Wilder Wohns
I'm having a similar issue. Did anyone ever resolve this error?
I've written a qc script that returns the number of close breakpoints and potential phasing switch errors for each individual. It runs like this: `python phasing_qc.py inferredtreesequence output_filename_prefix how_close_are_breakpoints individual_id_name_in_metadata`...
Yan and I just discussed this further and came to an agreement on three points: 1. Four parameters seem necessary: (1) original tree sequence, (2) ancestors_ts, (3) original sample_data, (4)...
Just cleaning it up a bit and will post as draft PR momentarily!
Ok I'll just put the rough version here then
Here's a version of the time-aware windowing only (i.e. no positional, span-based windows) This version returns nan when a time window has no information. ```python def genealogical_nearest_neighbours_time(ts, focal, reference_sets, time_windows=None):...
Great, thanks @jeromekelleher. I think the way to check the normalisation is that for each focal node/time window value, the GNN values for the _m_ reference sets should sum to...
> You could argue that we should create an extra bin for everything above the oldest time window in this case, but then the number of time windows could be...
This is super @hyanwong!! How about some simple metrics like MSE and spearman's to benchmark?
@hyanwong, one of the changes in the inside pass has dramatically slowed it down, I'm not sure which one