search

maurobio / cornpy

Cornell Ecology Programs in Python
GNU General Public License v3.0
0 stars 0 forks source link

Having trouble with cornpy install #1

Open Griette opened 3 years ago

Griette commented 3 years ago

Using pip install cornpy (but in a conda environment), even when pip is installed in that environment, still leads to the error ModuleNotFoundError: No module named 'cornpy'

I'm not great with Python - but I've always been able to fix this error. Any suggestions? Is it with the package itself, because there is no conda install, or something else on my end?

I just need to check how comparable TWINSPAN results are to other clustering methods...

maurobio commented 3 years ago

Dear Griette,

Hi. Thanks for your message. I am afraid I cannot be of much help, because I have never used "conda" and have tested the package installation only with a regular installation of the Python interpreter (version 2.7) using pip. To makes things worse, I do not use Python anymore. I have long moved to R, which offers a much more powerful and updated data analytic platform, including everything one needs to the analysis of ecological data, including an excelent package for running TWINSPAN ( https://github.com/zdealveindy/twinspanR).

As you may know, TWINSPAN is a divisive method, therefore it should not be directly compared with the "other" clustering methods (that is, the hierarchical agglomerative ones). However, if I may offer my humble opinion, in my personal experience with the analysis of vegetation data from the central Amazon rainforest, TWINSPAN performs better than the SAHN methods.

With warmest regards,

Em sex., 1 de out. de 2021 às 16:45, Griette van der Heide < @.***> escreveu:

Using pip install cornpy (but in a conda environment), even when pip is installed in that environment, still leads to the error ModuleNotFoundError: No module named 'cornpy'

I'm not great with Python - but I've always been able to fix this error. Any suggestions? Is it with the package itself, because there is no conda install, or something else on my end?

I just need to check how comparable TWINSPAN results are to other clustering methods...

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/maurobio/cornpy/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAPMACPUNHR6YGDOPQURDA3UEYFXHANCNFSM5FFMDCMA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Mauro J. Cavalcanti E-mail: @.*** Web: http://sites.google.com/site/maurobio "Life is complex. It consists of real and imaginary parts."

Griette commented 3 years ago

Dear Mauro,

You actually helped a lot - I did not try Python 2.7, and I guess I didn't notice that anywhere in GitHub! I'll try that out.

Since the package date was 2019, I was not thinking about Python version 2. I have an environment for that though, because of arcpy.

Because of having to deal with arcpy I know Python better than I know R. Started out in R, but Python is so much cleaner to write and read. I find R cumbersome, which is my downfall, because most people use R and therefore so should I.

I basically work in Python, unless or until there is a reason to switch. I have done most of the clustering analyses in R a year ago, but nothing that worked really well came out of it. I knew about the wrapper for TWINSPAN in R, but didn't try it out.

Since I had to build a very long nested for loop, that has to run many many times, and clustering methods are available in Python, I switched to Python. There is just one part in the for loop where I am implementing a clustering algorithm. I chose spectral co-clustering in the past because the results coincide with what I consider to be the 'reality,' but now I am trying to back-trace if I should have done that, considering it is out of line with what vegetation ecologists are using.

But it may honestly not be that different from TWINSPAN...

All clustering methods so far (including SAHN), except for the co-clustering method, lead to plot and species clusters that make very little sense in terms of the ecology of the Chaco (which is much simpler than that of the Amazon!). I know many ecologists have walked away from TWINSPAN, many still argue that TWINSPAN gives better results without understanding why, and many others argue for SAHN. Since all the vegetation literature from the Humid Chaco is based on past analyses using TWINSPAN, I thought I should give it a try.

Thanks for all your help!

I will try to work more in R ;-)

Cheers,

maurobio commented 3 years ago

Dear Griette,

I am glad to help, and to hear from your insights.

The apparent 'simplicity' of Python is very deceptive indeed. After spending almost ten years working with that language, I become very disappointed with what turned out to be a lot of its shortcomings, especially when compared with R (which may seem at first to be cumbersome, but with time Python scripts can become no less cumbersome and, in fact, appear to be quite limited because of the minimalist approach of Python to everything!).

I cannot understand why some ecologists have walked away from TWINSPAN, as it seems to work quite well (but of couse keeping in mind that NO method is 'perfect'). I am definetely amongst those who consider it useful, after trying it in many datasets (from tropical vegetation data to coral reef fish communities).

I will be much pleased to receiving any papers you may have already published or will surely publish in the future, concerning any aspects of quantitative analysis of ecological data.

With warmest regards,

Em sex., 1 de out. de 2021 às 17:45, Griette van der Heide < @.***> escreveu:

Dear Mauro,

You actually helped a lot - I did not try Python 2.7, and I guess I didn't notice that anywhere in GitHub! I'll try that out.

Since the package date was 2019, I was not thinking about Python version

  1. I have an environment for that though, because of arcpy.

Because of having to deal with arcpy I know Python better than I know R. Started out in R, but Python is so much cleaner to write and read. I find R cumbersome, which is my downfall, because most people use R and therefore so should I.

I basically work in Python, unless or until there is a reason to switch. I have done most of the clustering analyses in R a year ago, but nothing that worked really well came out of it. I knew about the wrapper for TWINSPAN in R, but didn't try it out.

Since I had to build a very long nested for loop, that has to run many many times, and clustering methods are available in Python, I switched to Python. There is just one part in the for loop where I am implementing a clustering algorithm. I chose spectral co-clustering in the past because the results coincide with what I consider to be the 'reality,' but now I am trying to back-trace if I should have done that, considering it is out of line with what vegetation ecologists are using.

But it may honestly not be that different from TWINSPAN...

All clustering methods so far (including SAHN), except for the co-clustering method, lead to plot and species clusters that make very little sense in terms of the ecology of the Chaco (which is much simpler than that of the Amazon!). I know many ecologists have walked away from TWINSPAN, many still argue that TWINSPAN gives better results without understanding why, and many others argue for SAHN. Since all the vegetation literature from the Humid Chaco is based on past analyses using TWINSPAN, I thought I should give it a try.

Thanks for all your help!

I will try to work more in R ;-)

Cheers,

  • Griette

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/maurobio/cornpy/issues/1#issuecomment-932539370, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAPMACKOBOIILWHLKNHZWXTUEYMVLANCNFSM5FFMDCMA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Mauro J. Cavalcanti E-mail: @.*** Web: http://sites.google.com/site/maurobio "Life is complex. It consists of real and imaginary parts."