Hello!
I’ve recently started university placement working on fortran-lang, specifically stuff surrounding stdlib and fpm, under supervision of @fxm, @hkvzjal and @FedericoPerini.
I’ll update any PR’s or discussion points in this thread, hopefully often.
I plan to start looking into the statistical functions of stdlib, specifically expanding upon the existing distribution functions. Whilst fairly new to fortran I look forward to getting to know the codebase and developing!
Despite a fairly busy week academically I managed to work on my first PR to stdlib and begin work locally on what will be my second, including reading up a bit more on more specificities of beta distribution for such. I’ve gotten to know the stats area of the codebase quite well and I’m getting to terms with the different standards of the library.
Hopefully I’ll be able to continue to work on beta distribution this coming week whilst finishing off gamma distribution which seems in a good spot as is, just needs some slight adjustments probably.
(and a bit)
Last week I largely continued to iterate the gamma PR specifically commits on docs, I’ve been able to progress my beta distribution using the feedback for the gamma PR. Given the documentation and structures will be the main differing point between the two it’s quite easy to work locally and have a beta PR ready once my gamma PR is finished!
I also did a fair bit of reading into Student’s T and F distributions numerical stability and the best way to go abouts mathematically approaching them. Their foundations are laid out so I’m able to work on them when I’m not working on gamma and beta.
This week I hope to continue iterating gamma.
Whilst having a fairly busy week I managed to work on making the loc parameter optional and adjusting all the documentation for it accordingly. I have my beta distribution in a similar position as gamma but realised that largely due to my lack of experience in large-scale github projects that I’ve PR’d from my Master fork so will hold on making a new PR for Beta distribution until I can fully finish gamma. I’d intended to do this either way as I’m now working beta, Student’s t, and F distribution along-side gamma so they all get worked on at once to best make sure they’re all in line.
I’ve done a fair bit of external reading on Student’s t distribution over the week to best understand how to approach it numerically as there are more ways; than the other distributions I’ve looked at so far, to achieve numerical stability.
After making some final changes to the Gamma Distribution PR it was able to be merged. It did have a bug on merge specifically with the gamma special functions fypp template which I had used to make the distribution function but not updated the test for properly so I should hopefully have that fixed in this PR.
Aside from those two I’ll be able to start creating branches this week for each of the distribution functions I’ve been working on along-side gamma. I’m just seeing through this gamma function change as beta distribution relies on it too!
This week I was able to fix the issue with the fypp template for the gamma special functions, and also added a check in the fpm deployment to catch the error if it were to happen again. Having this fixed I was able to open a new PR for the beta distribution I’ve been working on alongside gamma. I’ll be looking to polish and finalise this over the next week.
Student’s t and F distribution I’ve been able to locally insert the maths for now so are getting to a point where I can have PR’s for them too!
