[query] Fix old hl.plot.histogram log argument#11268
Conversation
| changes = { | ||
| "bin_freq": [math.log10(x) if x > 0.0 else x for x in data.bin_freq], | ||
| "n_larger": math.log10(data.n_larger) if data.n_larger > 0.0 else data.n_larger, | ||
| "n_smaller": math.log10(data.n_smaller) if data.n_smaller > 0.0 else data.n_smaller | ||
| } | ||
| data = data.annotate(**changes) |
There was a problem hiding this comment.
I think this is wrong, because it plots count-0 and count-1 bins the same. I think the right thing to do is log(x) if x > 0.0 else shift, where shift < 0 is arbitrary, but shift=-1 is as good as anything.
johnc1231
changed the title
|
@patrick-schultz , shift = -1 can lead to some pretty ugly graphs when the range of log transformed y values is small (i.e. between 0 and 1) |
|
The negative 1 doesn't seem right, but we should probably give a warning or something when a user does this. I could log a warning telling user how many points were 0 and so weren't plotted? |
|
That seems good. Perhaps we could also suggest using |
|
|
…am would never work. Also have to handle the case where there are 0s
johnc1231
force-pushed
the
johnc-fix-old-plot-hist
branch
from
7683536 to
a016948
Compare
3 participants
Struct fields are not mutable, so the log argument of hl.plot.histogram would never work. Also have to handle the case where there are 0s