Next Release of lmfit

[Matthew Giammar]

Hi there,

Thank you for being responsive on my issue with loading json strings (issue #739). Running locally, everything works as expected!

As I'm trying to push the updated app to our hosting service I'm running into some limitations (mostly me being new to heroku). I need the bug fix for the app to run without errors but the current version on pip is older. Updating lmfit via pip would be much easier than some finagling with .egg files.

When is the next planned release of the lmfit library? Is there any possibility for a patch release?

Thanks,

Matthew

[Renee Otten]

Hi Matthew, 

I think we have accumulated enough useful additions/changes to release a new version. I’ll try to take a look in the weekend if there is anything left that could potentially be wrapped up fairly quickly to make it into a new version and we can see what Matt thinks about it. 

Just in case you’re not aware of it, you can use pip to install from GitHub and specify a commit to use. For example, if you’d want get the fix for the issue you are referring to, you could do: 

$ pip install git+git://github.com/lmfit/lmfit-py.git@f0707cccec057e0e922b105f6ca809ff7e59b524

Hope that helps,

Renee

--
You received this message because you are subscribed to the Google Groups "lmfit-py" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lmfit-py+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lmfit-py/e47b90a6-48a4-4193-b8e8-195f03629cafn%40googlegroups.com.

[Matt Newville]

Hi Renee, Matthew, 

I think releasing 1.0.3 is a fine idea.  I am not aware of any outstanding issues that would need to be resolved.  Tagging the current master version as 1.0.3 seems perfectly reasonable to me.  Is there anything you would like to see, Renee?   Does anyone have any objections or suggestions to pushing out 1.0.3 in the next week or so?

To view this discussion on the web visit https://groups.google.com/d/msgid/lmfit-py/1705B0CF-A76D-45AE-8926-E182DDEF62FC%40gmail.com.

--Matt

[Matthew Giammar]

Hey Renne, Matt,

I did not know you could install dependencies directly from github! Adding that to requirements.txt did the trick. Thanks again for the excellent support!

Best,

Matthew

[Peter Stahlecker]

Will you inform when the new version has been released, so I can update?

Thanks!

I am ‚new‘ to Python, so I do not know how all of this works)

To view this discussion on the web visit https://groups.google.com/d/msgid/lmfit-py/dfbfd2e2-f683-4c9f-88ac-3b5fdb93e8e2n%40googlegroups.com.

--

Best regards,

Peter Stahlecker

[Renee Otten]

Hi Matt, 

On Sep 17, 2021, at 6:41 PM, Matt Newville <newv...@cars.uchicago.edu> wrote:

I think releasing 1.0.3 is a fine idea.  I am not aware of any outstanding issues that would need to be resolved.  Tagging the current master version as 1.0.3 seems perfectly reasonable to me.  Is there anything you would like to see, Renee?   Does anyone have any objections or suggestions to pushing out 1.0.3 in the next week or so?

I was thinking about the ones listed below; I am not sure though how quickly we could come to a solution for them and would be fine with cutting a release soon (they are certainly not blockers). I do have a few, small documentation changes though which I can probably push today/tomorrow.

1. emcee and multiprocessing: https://groups.google.com/g/lmfit-py/c/ai0Tgrp6j_w/m/zqhZ-0LyAgAJ

I did some investigative work back then and even though we are passing initial the correct Pool to “emcee” it gets lost at some point. I don’t remember the details anymore and couldn’t figure out the issue, but it would be nice to resolve. On the other hand, if the plan of moving “emcee” out of the minimizers into a “post-fitting” methods like the ConfidenceInterval is still alive we’d need to go through the code anyway and could hopefully fix this at the same time.

2. discussion about the DoniachModel: https://github.com/lmfit/lmfit-py/discussions/709 

I am not sure whether there was a conclusion about what to do with this. 

3. Warning message about non-Parameter keywords to a Model function: https://groups.google.com/g/lmfit-py/c/blUL0pUiLZA/m/b_Qp_1upAwAJ 

[Matt Newville]

Hi Renee, 

Thanks for compiling that.

On Mon, Sep 20, 2021 at 10:51 AM Renee Otten <otten...@gmail.com> wrote:

Hi Matt, 

On Sep 17, 2021, at 6:41 PM, Matt Newville <newv...@cars.uchicago.edu> wrote:

I think releasing 1.0.3 is a fine idea.  I am not aware of any outstanding issues that would need to be resolved.  Tagging the current master version as 1.0.3 seems perfectly reasonable to me.  Is there anything you would like to see, Renee?   Does anyone have any objections or suggestions to pushing out 1.0.3 in the next week or so?

I was thinking about the ones listed below; I am not sure though how quickly we could come to a solution for them and would be fine with cutting a release soon (they are certainly not blockers). I do have a few, small documentation changes though which I can probably push today/tomorrow.

1. emcee and multiprocessing: https://groups.google.com/g/lmfit-py/c/ai0Tgrp6j_w/m/zqhZ-0LyAgAJ

I did some investigative work back then and even though we are passing initial the correct Pool to “emcee” it gets lost at some point. I don’t remember the details anymore and couldn’t figure out the issue, but it would be nice to resolve. On the other hand, if the plan of moving “emcee” out of the minimizers into a “post-fitting” methods like the ConfidenceInterval is still alive we’d need to go through the code anyway and could hopefully fix this at the same time.

I think this would be a fair amount of work and would vote "postpone".  

2. discussion about the DoniachModel: https://github.com/lmfit/lmfit-py/discussions/709 

I am not sure whether there was a conclusion about what to do with this. 

Right, I think there sort of was not a complete resolution on what to do here (well, except that "XPS-specific version of DoniachModel" would be fine") and no actual PR.    In principle, I'd be willing to help, but I'm not actually an XPS person, so maybe wait until we have a PR?

  

3. Warning message about non-Parameter keywords to a Model function: https://groups.google.com/g/lmfit-py/c/blUL0pUiLZA/m/b_Qp_1upAwAJ 

Yeah, this one sort of confused me, and probably should be fixed.  I have not looked at this in a while, but I can look into this again.  I don't know if it is an easy fix or not.  

I might add a #4 (or maybe it is "#3.5"), that having a decorator or some other check to ensure independent variables have values passed in or default values might help fix this, and is sort of related to Github Issue #747.

I think that might be doable, but I have not tried it and I doubt very much I'll do that this week.

Honestly, I would lean toward "release 1.0.3 sooner rather than later", and not expect changes from PRs that do not yet exist,  leaving these as To-Do for 1.0.4.   I'd like to think I could be persuaded to work on some of these, but I'm not seeing that as happening this week.

--Matt 

[Renee Otten]

Sounds good to me. I’ll try to do a final PR with some minor, unrelated changes tonight or tomorrow and then you can tag a version.

Renee

On Sep 20, 2021, at 5:53 PM, Matt Newville <newv...@cars.uchicago.edu> wrote:



--
You received this message because you are subscribed to the Google Groups "lmfit-py" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lmfit-py+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/lmfit-py/CA%2B7ESbrKiTV%3DGPf7eQRmut%2BdF%2BVmLuLoiyhpF-ri3pV%2BV3xz_w%40mail.gmail.com.