support for poisson errors on data

[aminnj]

support for poisson errors on data

[aminnj]

Looks like that’s an approximation for clipper Pearson based on the name? For binomial errors there’s this:

https://github.com/aminnj/yahist/blob/a6a906923c6de557ffad56211c13c9e6b1fc5565/yahist/utils.py#L133

[sgnoohc]

Ugh. I hate statistics.

Let me choose the faster/lighter version already in FHist, and then provide an option for user to apply their favorite lambda.

[Moelf]

what about:
https://github.com/Moelf/FHist.jl/blob/2ea9ea2036b416b96042b173751be3e69d2c4347/src/FHist.jl#L13

as a faster/lighter alternative

[aminnj]

Dunno why scipy uses “ppf” instead of just “quantile” 🤷

[sgnoohc]

#24
For first half of the implementation

[sgnoohc]

someone would have to factor out "pearson_err" function call in get_data_with_pearson_err in utils.jl (and possiblt change name of the function)

[sgnoohc]

#26
#27
The latter half of the implementation
Done!