Comments (15)
I am not sure whether it's related, but trying to do pip install greykite
broke my Jupyter installation for me
from greykite.
Yes, I think most of these issues are related to prophet installation. We are working on a solution and it should be resolved in the next release which we hope to publish in a timely fashion. Please stay tuned for an update.
from greykite.
Hi Marco,
That is a reasonable comment. We wanted to avoid installation issues (when library is installed on its own). Some of which were coming from the requirements for Prophet which we might make optional or remove in future releases. We will also see which requirements are possible to be made less strict. Thanks for the input.
Reza
from greykite.
As an extended question, is there extended usage of FBProphet included? I only see a wrapper class to run Prophet models (yet not the most recent version), which can totally be done by directly calling the Prophet package independently, but didn't find any extended usage. If that's the case, including (FB)Prophet as part of the requirements seems redundant, as it causes major installation issues for this awesome package.
from greykite.
Hi everyone,
Thanks for the comments in this thread. We have just released greykite==0.2.0
on Pypi which
- removes the dependency on
fbprophet
(it is optional now, but you need to install it manually if you want to use the Greykite wrapper onfbprophet
) - removes dependencies that are not useful during runtime.
- changes pinned dependencies to ranged dependencies.
Hopefully this resolves the installation issues.
from greykite.
Please see greykite==0.2.0
. This issue is resolved. Thanks everyone!
from greykite.
Hi Reza,
We recently received a request to add a wrapper for GreyKite as a part of HCrystalBall. I plan to update the new release to prophet>=1.0
and also searching through your codebase, it seems, the migration might be not such a difficult move?
Could you let me know if both - the addition to HCrystalBall is welcomed and whether you'd directly consider prophet 1 over e.g. fbprophet 0.6 in case you would consider updating the dependencies? Would make my life much easier to know.
Or maybe leave the prophet as an optional dependency?
Michal
from greykite.
This is causing problems for me as well.
from greykite.
yeah I think as it currently stands the only way to use greykite is to make a new conda environment / virtualenv specially for it :)
from greykite.
Why is pystan in the requirements?
from greykite.
@CatChenal - I know PyStan is a prerequisite for Prophet: https://facebook.github.io/prophet/docs/installation.html
from greykite.
Will the next version move from fbprophet
to prophet
, following the prophet
library's name migration?
from greykite.
@aks43725 The prophet wrapper is an estimator wrapper, allowing it to be used in the Greykite pipeline (including cross-validation, trai/test split, etc.) We changed the fbprophet
to optional so you don't need it to install Greykite anymore. And if you have fbprophet==0.5, 0.6, 0.7
it will be compatible, but without some later added parameters in 0.6 or 0.7
.
from greykite.
@ahgraber Currently we unpinned fbprophet
and the library works with fbprophet==0.5 ,0.6, 0.7
except for a small fraction of later added parameters in fbprophet==0.6, 0.7
. We currently don't support prophet==1.0
because it has a different library name. In the long term we might want to support prophet==1.0
, but this may not happen in the near future.
from greykite.
@ahgraber Currently we unpinned
fbprophet
and the library works withfbprophet==0.5 ,0.6, 0.7
except for a small fraction of later added parameters infbprophet==0.6, 0.7
. We currently don't supportprophet==1.0
because it has a different library name. In the long term we might want to supportprophet==1.0
, but this may not happen in the near future.
Thanks for the update. I believe the limitation on prophet==1.0
will prevent greykite + prophet compatibility with python 3.9. Regardless, looking forward to this update!
from greykite.
Related Issues (20)
- run_forecast_config crash with regressors but not without HOT 1
- In get_auto_silverkite_model_template, `min_gap_in_seconds` fails if input is not sorted HOT 1
- Add support for HistGradientBoostingRegressor HOT 2
- Predicted values and metrics for N-step ahead forecast. HOT 5
- Implemening regressors HOT 3
- Importing Forecaster gives error for United States class (Holiday?) HOT 4
- Loosening dill version HOT 2
- GreyKite 0.3.0 Library Issue HOT 7
- Library import issues HOT 1
- ValueError: ``MULTISTAGE_EMPTY`` can not be used without overriding. HOT 2
- Request to release tag 0.5.1 HOT 4
- Unable to get greykite 0.5.0 HOT 4
- Unable to run codes.
- Regressors Already Forecastd, No Lag Needed. But, getting warning "RuntimeWarning: Input data has many null values. Missing 21.45% of one input." HOT 1
- Lower Python version requirement to allow 3.8.10 HOT 1
- `design_info` is needed to make predictions on new data HOT 4
- Dealing with lot of 0s (zeroes) in Greykite Multistage Forecasting HOT 1
- mutable default <class 'greykite.framework.templates.autogen.forecast_config.ModelComponentsParam'> for field SILVERKITE is not allowed: use default_factory HOT 1
- MLFLOW support for the SilverKite Algorithm HOT 1
- Greykite design question
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from greykite.