To update app, train+deploy or build-train-deploy?

[Noelia Osés Fernández <no...@vicomtech.org>]

Hi all,

I have created an app using ActionML's UR and Prediction IO trained with
some initial data.

Now, as users interact with the ecomm website, I will have more data to
train the app. Once a week I will take all the new data from the last week
and retrain the app.

Which is the correct way to keep the app update?

1- pio build; pio train; pio deploy

or

2- pio train; pio deploy

Thank you very much for your help!

I don't really know what "pio build" does so I don't know if I have to
re-do it every time.

Best regards,
Noelia

Re: To update app, train+deploy or build-train-deploy?

[Noelia Osés Fernández <no...@vicomtech.org>]

Thank you for the quick reply!!

On 4 December 2017 at 10:39, Александр Лактионов <lo...@gmail.com>
wrote:

> Hi Noelia,
> You don’t have to build project every time after train, just once.
> Train will save ES index with model and move alias specified in
> engine.json so deploy can work simultaneously.
> Dont forget to run `pio train` in the same dir where you do `pio deploy`,
> the same `pwd` is necessary to swap the same alias
>
> > 4 дек. 2017 г., в 12:36, Noelia Osés Fernández <no...@vicomtech.org>
> написал(а):
> >
> > Hi all,
> >
> > I have created an app using ActionML's UR and Prediction IO trained with
> some initial data.
> >
> > Now, as users interact with the ecomm website, I will have more data to
> train the app. Once a week I will take all the new data from the last week
> and retrain the app.
> >
> > Which is the correct way to keep the app update?
> >
> > 1- pio build; pio train; pio deploy
> >
> > or
> >
> > 2- pio train; pio deploy
> >
> > Thank you very much for your help!
> >
> > I don't really know what "pio build" does so I don't know if I have to
> re-do it every time.
> >
> > Best regards,
> > Noelia
> >
>
>


-- 
<http://www.vicomtech.org>

Noelia Osés Fernández, PhD
Senior Researcher |
Investigadora Senior

noses@vicomtech.org
+[34] 943 30 92 30
Data Intelligence for Energy and
Industrial Processes | Inteligencia
de Datos para Energía y Procesos
Industriales

<https://www.linkedin.com/company/vicomtech>
<https://www.youtube.com/user/VICOMTech>
<ht...@Vicomtech_IK4>

member of:  <http://www.graphicsmedia.net/>     <http://www.ik4.es>

Legal Notice - Privacy policy <http://www.vicomtech.org/en/proteccion-datos>

Re: To update app, train+deploy or build-train-deploy?

[Александр Лактионов <lo...@gmail.com>]

Hi Noelia,
You don’t have to build project every time after train, just once. 
Train will save ES index with model and move alias specified in engine.json so deploy can work simultaneously.
Dont forget to run `pio train` in the same dir where you do `pio deploy`, the same `pwd` is necessary to swap the same alias

> 4 дек. 2017 г., в 12:36, Noelia Osés Fernández <no...@vicomtech.org> написал(а):
> 
> Hi all,
> 
> I have created an app using ActionML's UR and Prediction IO trained with some initial data.
> 
> Now, as users interact with the ecomm website, I will have more data to train the app. Once a week I will take all the new data from the last week and retrain the app.
> 
> Which is the correct way to keep the app update?
> 
> 1- pio build; pio train; pio deploy
> 
> or
> 
> 2- pio train; pio deploy
> 
> Thank you very much for your help!
> 
> I don't really know what "pio build" does so I don't know if I have to re-do it every time.
> 
> Best regards,
> Noelia
>