Problems with the function plot_overtime

[Ivanad29]

Dear Gilestro we have recently biuld up ethoscopes based on your development and we have been analyzing data with the python in collab scripts and suddenly the function plot_overtime stop to run it seems objets of the type behavpy do not support this function anymore. Any ideas?

AttributeError Traceback (most recent call last)
in <cell line: 0>()
1 # Crear el gráfico
----> 2 fig = df_48hr.xmv('sex', 'F').plot_overtime(variable='asleep', wrapped=False, facet_col='exp_group', facet_arg = [3, 2, 1], facet_labels = ['CONTROL Elav WHITE', 'CONTROL Trip WHITE', 'EXPERIMENTAL ELAV Trip'], avg_window=180, title='FEMALES')
3
4 # Actualizar diseño
5 fig.update_layout(autosize=False, width=1500, height=500)

/usr/local/lib/python3.11/dist-packages/pandas/core/generic.py in getattr(self, name)
6297 ):
6298 return self[name]
-> 6299 return object.getattribute(self, name)
6300
6301 https://github.com/Final

AttributeError: 'behavpy' object has no attribute 'plot_overtime'

I tried changing the plot function but the thing is the ones from seaborn or plotly crashes the arguments of sex, and groups since the columns ar not exactly part of the main dataframe

Thanks for your time and help,
Ivana

[Lblackhurst29]

Hello, i'm assuming that with seaborn it is the newest update 2.0?

Is data old behapy dataframes? As they will need to be re-initalised with the new structure as the base classes have been changed.

Example:

load old dataset

df = pd.read_pickle('path/to/pickle/file.pkl')

re-intialise with the new class

df = etho.behavpy(df, df.meta, canvas = 'seaborn')

Hopefully this sorts it. If so i'll update the notebooks and READme to make it more clear.

[Ivanad29]

I mention seaborn just as an example, What surprises me is that sunddenly the one week to the other plot_overtime was not supported anymore by behavpy object! So you are telling me this can cannot be solved and I need to import differently from now on the data(pickles) which mean I should change ALL the code of the differents graphs? because the rest of it it does still works its only the plot_overtime function

[Lblackhurst29]

Well that does seem different if it's just plot_overtime.

Could you please tell me the version so I can test it?

[Ivanad29]

Excuse my ignorance, you mean the version of what exaclty?

[Lblackhurst29]

it's the version of the package, each update I push out ups increases the version.

I recently did a big update in which I changed up a lot of the backend and added in the option to plot with seaborn or plotly.

To see the version in jupyter notebooks:

"""
import ethoscopy as etho
print(etho.version)
"""

The latest should be 2.0.

If it's not the latest then you will need to update the package with whatever package dependency manager you use, the basic one being pip.

[Ivanad29]

I have run import ethoscopy as etho
print(etho.version) and then restarted the kernel and imported everything again and still this happends:

AttributeError Traceback (most recent call last)
in <cell line: 0>()
----> 1 print(etho.version)

AttributeError: module 'ethoscopy' has no attribute 'version'

[Lblackhurst29]

This is odd.

Sorry, i've only just noticed you stated you are using google collab notebooks, which I don't often use. I have tried running the tutorial notebook in collab and it seems to work.

Maybe try it for yourself, the notebook is here: https://github.com/gilestrolab/ethoscopy/blob/main/tutorial_notebook/1_Overview_tutorial.ipynb . You can download it and upload it to collab, but run the !pip install ethoscopy first.

However, it should be producing the version correctly in collab.

I'm sorry I can't be much more help but it seems to be an issue with your install and not the code. From what I can tell. I might be wrong as i'm not experienced in all of this.

[Ivanad29]

well I run etho.version intsead of print(etho.version)
and the result was
2.0
so everything seems to be ok, for the exception of this plot_overtime fucntion :S
Thanks anyway for your time

[Lblackhurst29]

If you want and are happy to, you could share the collab notebook with me and the data, that was I can better identify the problem? As it is a very frustrating one.

[Ivanad29]

Hi! yes of course. teh Collab notebook is this one: (https://colab.research.google.com/drive/1sTrgp21CKw4N_uFzCp9ZN7epK10bdQbu?usp=sharing)

and the data is this pickle:
(https://www.dropbox.com/scl/fi/d90ijq0chqsvbaa1qddb6/analisis_10-03-2025.pkl?rlkey=yf59rsel0zxnlzxsm6hk19wu3&st=wmdisump&dl=0)

[Lblackhurst29]

Hello,

So I had a look through and the issue is what I thought it was at the beginning, any old data saved in the behavpy form is incompatible with the latest update.

So any data you have previously saved will need to be re-initialised with the new update like so:

df = pd.read_pickle('analisis_10-03-2025.pkl') # replace with the above
df_new = etho.behavpy(df, df.meta, canvas = 'seaborn')

This would only be for old data and you could overwrite the old file with the newly initialised behavpy table.

You could fix the package version to the previous one when using a Python environment:

# 1.3.11 is the last update of the package that used the previous structure
!pip install ethoscopy==1.3.11

However, this won't work in google collab (at least when I tried) because this downgrades Pandas and Numpy to versions incompatible with collab. This would work on a local environment or if you hosted a jupyter notebook server yourself.

Sorry about this change, but I had to update it all to keep the software up to date with the newer versions of Pandas and Numpy which collab also rely on.

[Ivanad29]

Oh these are good news, knowing how to deal with it. Don't be sorry, all the contrary many many thanks for your time and help!! I'll try this and let you know how it goes. Have a nice day, Ivana El lun, 7 de abr. de 2025 9:53 a. m., L.Blackhurst ***@***.***> escribió:

…

Hello, So I had a look through and the issue is what I thought it was at the beginning, any old data saved in the behavpy form is incompatible with the latest update. So any data you have previously saved will need to be re-initialised with the new update like so: df = pd.read_pickle('analisis_10-03-2025.pkl') # replace with the above df_new = etho.behavpy(df, df.meta, canvas = 'seaborn') This would only be for old data and you could overwrite the old file with the newly initialised behavpy table. You could fix the package version to the previous one when using a Python environment: # 1.3.11 is the last update of the package that used the previous structure !pip install ethoscopy==1.3.11 However, this won't work in google collab (at least when I tried) because this downgrades Pandas and Numpy to versions incompatible with collab. This would work on a local environment or if you hosted a jupyter notebook server yourself. Sorry about this change, but I had to update it all to keep the software up to date with the newer versions of Pandas and Numpy which collab also rely on. — Reply to this email directly, view it on GitHub <#5 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWKEOSOEEXW5U7LXK6YBDR32YJYNXAVCNFSM6AAAAAB2G27FLKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOOBTGIZTSOJUGU> . You are receiving this because you authored the thread.Message ID: ***@***.***> [image: Lblackhurst29]*Lblackhurst29* left a comment (gilestrolab/ethoscopy#5) <#5 (comment)> Hello, So I had a look through and the issue is what I thought it was at the beginning, any old data saved in the behavpy form is incompatible with the latest update. So any data you have previously saved will need to be re-initialised with the new update like so: df = pd.read_pickle('analisis_10-03-2025.pkl') # replace with the above df_new = etho.behavpy(df, df.meta, canvas = 'seaborn') This would only be for old data and you could overwrite the old file with the newly initialised behavpy table. You could fix the package version to the previous one when using a Python environment: # 1.3.11 is the last update of the package that used the previous structure !pip install ethoscopy==1.3.11 However, this won't work in google collab (at least when I tried) because this downgrades Pandas and Numpy to versions incompatible with collab. This would work on a local environment or if you hosted a jupyter notebook server yourself. Sorry about this change, but I had to update it all to keep the software up to date with the newer versions of Pandas and Numpy which collab also rely on. — Reply to this email directly, view it on GitHub <#5 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWKEOSOEEXW5U7LXK6YBDR32YJYNXAVCNFSM6AAAAAB2G27FLKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOOBTGIZTSOJUGU> . You are receiving this because you authored the thread.Message ID: ***@***.***>