Prophet 1.0.1 vs fbProphet
opened this issue · comments
Would we be able to upgrade to the newer version of Prophet, that is now just called prophet? It seems like it would be an easy fix overall that I would be happy to take on.
Hi @KarateKat95 👍
This is a very good idea. Can you please create the changes in your forked version and submit a pull request with all the necessary changes? You have to test it first that your changes work!
Thanks
AutoViML
I have completed that. Thanks!
But doesn't work yet. I opened new pull request.
But doesn't work yet. I opened new pull request.
@turkalpmd See #82, where I already made this pull request a few months ago. It just needs to be merged by @AutoViML.
But doesn't work yet. I opened new pull request.
@turkalpmd See #82, where I already made this pull request a few months ago. It just needs to be merged by @AutoViML.
Yes, I want to trying this today but it didnt work properly, after that. To be honest, I created a new pull request without checking the issue section. Then I saw you did something similar. I can delete it of course, but maybe a new notification will go away to @AutoViML . It is very easy changing. Thank you.
Hi @KarateKat95 and @turkalpmd
Thank you so much for filing the pull requests. I am very concerned about changing from FBProphet to Prophet without testing it - can you guys please test it with a few datasets before I approve it? 👍
I am sure it will work fine but just want to be sure.
Thanks
AutoVimal
Hi @KarateKat95 and @turkalpmd Thank you so much for filing the pull requests. I am very concerned about changing from FBProphet to Prophet without testing it - can you guys please test it with a few datasets before I approve it? +1
I am sure it will work fine but just want to be sure. Thanks AutoVimal
I will testing this. But I didn't change any architecture, I changed just library name.
Hi @KarateKat95 and @turkalpmd Thank you so much for filing the pull requests. I am very concerned about changing from FBProphet to Prophet without testing it - can you guys please test it with a few datasets before I approve it? +1
I am sure it will work fine but just want to be sure. Thanks AutoVimalI will testing this. But I didn't change any architecture, I changed just library name.
Thanks @turkalpmd; I will run a few tests with some datasets on my end as well this weekend.
Any updates? Is the upgrade and change from one version to another smooth? Can you please update the code everywhere there is an import statement of FB prophet to Prophet?
I'd like you to test it on a few datasets after installing and importing auto-ts with the new version of prophet in your computers.
You can use pypi-test.org by uploading and testing your new versions of auto-ts. Hope that helps,
Thank you in advance 👍
AutoVimal
I was on-call this week. I can be doing some updates tomorrow. I don't know if I haven't tested the codes on GitHub before, but I'll learn. I'll fix all other import commands today. Thank you for your patience.
Kind regards.
Any updates? Is the upgrade and change from one version to another smooth? Can you please update the code everywhere there is an import statement of FB prophet to Prophet?
I'd like you to test it on a few datasets after installing and importing auto-ts with the new version of prophet in your computers.
You can use pypi-test.org by uploading and testing your new versions of auto-ts. Hope that helps,
Thank you in advance +1 AutoVimal
$ pip install git+git://github.com/AutoViML/Auto_TS
Collecting git+git://github.com/AutoViML/Auto_TS
Cloning git://github.com/AutoViML/Auto_TS to /tmp/pip-req-build-0xdvyn41
Running command git clone --filter=blob:none --quiet git://github.com/AutoViML/Auto_TS /tmp/pip-req-build-0xdvyn41
fatal: unable to connect to github.com:
github.com[0: 140.82.121.4]: errno=Connection timed out
error: subprocess-exited-with-error
× git clone --filter=blob:none --quiet git://github.com/AutoViML/Auto_TS /tmp/pip-req-build-0xdvyn41 did not run successfully.
│ exit code: 128
╰─> See above for output.
note: This error originates from a subprocess, and is likely not a problem with pip.
error: subprocess-exited-with-error
× git clone --filter=blob:none --quiet git://github.com/AutoViML/Auto_TS /tmp/pip-req-build-0xdvyn41 did not run successfully.
│ exit code: 128
╰─> See above for output.
note: This error originates from a subprocess, and is likely not a problem with pip.
fbprophet contain just test/test_auto_ts.py line 18 and auto_ts/build_prophet.py line 19-22 32 and 545
this changing is very simple. unfortunately Git command and pip command doesnt work.
Hi @turkalpmd :
There is a simple way to install auto-ts without dependencies 👍
pip install auto-ts --ignore-installed --no-deps
Then you can install FB prophet independently using the following command:
conda install -c conda-forge fbprophet
Hope this helps,
AutoVimal
Hi @turkalpmd : There is a simple way to install auto-ts without dependencies 👍
pip install auto-ts --ignore-installed --no-depsThen you can install FB prophet independently using the following command:
conda install -c conda-forge fbprophetHope this helps, AutoVimal
I will try this weekend
Hi @turkalpmd @KarateKat94
I have updated Auto_TS with the latest version of prophet 1.0 - please double check if this works and let me know
Thank you for all you pull requests and commitment!
AutoVimal
I will try in next 24 hour :D. Thank you.
Yes It is works perfect. But I found another bug from univariable notebook, and you must change pypi release, pypi installed old .68 version;
---------------------------------------------------------------------------
UnboundLocalError Traceback (most recent call last)
/tmp/ipykernel_7312/3941123415.py in <module>
4 model_type=['best'],dask_xgboost_flag=False,
5 verbose=2)
----> 6 model.fit(traindata, ts_column,target)
~/.local/lib/python3.9/site-packages/auto_ts/__init__.py in fit(self, traindata, ts_column, target, sep, cv)
573
574 self.ml_dict[name]['model'] = model
--> 575 self.ml_dict[name]['forecast'] = forecast_df_folds
576 self.ml_dict[name][self.score_type] = score_val
577 self.ml_dict[name]['model_build'] = model_build
UnboundLocalError: local variable 'forecast_df_folds' referenced before assignment