Pinned packages that need to be updated

[neg2sode]

Describe the bug
Calling ggplot in plotnine triggers dependency errors, due to incompatible versions of pinned packages.

Steps to reproduce

1. Creat a new Jupyter notebook.
2. Run the code below.
3. See “ModuleNotFoundError: No module named 'matplotlib.layout_engine’” raised from ggplot.py

`import numpy as np
import pandas as pd
import statsmodels.api as sm
import statsmodels.formula.api as smf
from itertools import combinations
import plotnine as p

import ssl
ssl._create_default_https_context = ssl._create_unverified_context
def read_data(file):
return pd.read_stata("https://github.com/scunning1975/mixtape/raw/master/" + file)

auto = pd.read_stata('https://github.com/scunning1975/mixtape/raw/master/auto.dta')
auto['length'] = auto['length'] - auto['length'].mean()

lm1 = sm.OLS.from_formula('price ~ length', data=auto).fit()
lm2 = sm.OLS.from_formula('price ~ length + weight + headroom + mpg', data=auto).fit()

coef_lm1 = lm1.params
coef_lm2 = lm2.params
resid_lm2 = lm2.resid

auto['y_single'] = coef_lm1[0] + coef_lm1[1]*auto['length']
auto['y_multi'] = coef_lm1[0] + coef_lm2[1]*auto['length']

p.ggplot(auto) +
p.geom_point(p.aes(x = 'length', y = 'price')) +
p.geom_smooth(p.aes(x = 'length', y = 'y_multi'), color = "blue") +
p.geom_smooth(p.aes(x = 'length', y = 'y_single'), color="red")`

Expected behavior
Expected to run normally, with graphs presented in the output.

Screenshots

Environment (please complete the following information):

• Device: iPad Air 5th generation
• iOS version: iPadOS 17.5.1
• App version 3.1.0

[alexstaravoitau]

Thanks for reporting @neg2sode—yep, I'm planning to update the pinned packages soon.

[neg2sode]

Good to hear! I look forward to see that in future updates.

[AndreFelixM]

Not sure why this has been closed when the update has not been done yet. I am also just missing an update to pandas to be able to use this program properly. Having the version that is over 1y old is blocking me to use it.

[alexstaravoitau]

@AndreFelixM Yep agreed, reopening for now.

[andre-f-m]

Hi Alex, is there any update on this? Do you have an ETC for the package updates? Thank you

[alexstaravoitau]

@andre-f-m There are still a couple of requested packages I want to add before starting on the upgrade (which will be a bit of a journey because of internal dependencies and cross-dependencies between packages). Don't have an ETA I'm afraid, but will update the issue when there are any developments.

[andre-f-m]

@alexstaravoitau , do you have any update on the overall upgrade of packages? I see multiple people requesting other packages for months and you mentioned that will happen in next overall packages upgrade, but we don’t see it happening. This open point has been open for 6 months and sadly we don’t have an ETA. Hard to know if we move on or just wait for another month, and without an ETA, the first option might be the way to go… Your product is awesome, but this update is holding me back to use it properly.

[alexstaravoitau]

@andre-f-m Apologies it’s taking a while—I want to address a few more package requests before embarking on upgrading the existing ones. These upgrades are always tricky, because it’s a single environment—so all pinned packages need to work together, and sometimes it means some packages need to stay at older versions… I do believe we should be able to upgrade most of them though. Meanwhile, perhaps you could try an older version of the package you’re trying to use that has pinned dependencies?

The next Juno release is coming in the next week or two, and it will add OpenCV, which has been a popular request for a while now. There are a couple of smaller/easier packages that I plan to add after that and look into upgrading the environment next.