Development setup

Prerequisites

This following packages must be installed

• python
• poetry
• git

Configuration

• poetry configuration, add environment variable POETRY_VIRTUALENVS_IN_PROJECT=true
• vscode configuration, add environment variable PYTHON_VENV_LOC
  • on windows: PYTHON_VENV_LOC=.venv\\bin\\python.exe
  • on linux: PYTHON_VENV_LOC=.venv/bin/python
• git configuration

git config --global user.name 'your name'
git config --global user.email 'your email'

Initialization

• First setup poetry install
• Then poetry shell

Build and publish with poetry

Build

Manuel steps to generate and publish the package to TestPyPI with poetry, documentation from packaging.python

Build the package, generate distribution archives

poetry build

Publish to Test PyPI

Add Test PyPI as an alternate package repository

poetry config repositories.testpypi https://test.pypi.org/legacy/

Upload/publish package/distribution archive to TestPyPI (a separate instance of the Python Package Index)

poetry publish -r testpypi

Installation with pip

pip install --index-url https://test.pypi.org/simple/ energy-study

or

pip3 install --index-url https://test.pypi.org/simple/ energy-study

Code of Conduct

History (changelog)

Swagger

curl -X POST -H "content-type:application/json" -d '{"swaggerUrl":"https://petstore.swagger.io/v2/swagger.json"}' https://generator.swagger.io/api/gen/clients/python

you can POST to https://generator.swagger.io/api/gen/clients/{language} with the following HTTP body

{
  "options": {
    "packageName": "energ_study"
  },
  "spec": {

  }
}

Swagger code generation

https://swagger.io/tools/swagger-codegen/ Online generator for api client

curl -X POST -H "content-type:application/json" -d '{"swaggerUrl":"https://petstore.swagger.io/v2/swagger.json"}' https://generator.swagger.io/api/gen/clients/ruby

https://generator.swagger.io/#/clients/generateClient

RTE API

https://data.rte-france.com/catalog/consumption

Il est conseillé de faire un appel par heure à ce service et de ne pas dépasser une période de 155 jours par appel.

Prediction J-1 Error

The residuals are the rescaled one-step prediction errors $$ \hat{W_t} = (X_t - \hat{X_t}) / \sqrt{r_{t-1}} $$

$$ r_{t-1} = E(X_t - \hat{X_t}) / \sigma^2 $$

$\sigma^2$ is the white noise variance of the fitted model

To check the appropriateness of the model we therefore examine the residual series $\hat{W_t}$, and check that it resembles a realization of a white noise sequence.

• Plot

• QQ-Plot (normal)

• QQ-Plot (t-distr)

• Histogram

• ACF/PACF

• ACF Abs vals/Squares

• Tests of randomness

• distribution

• descriptive statistics: mean, median, std, ...

• ACF and PACF

• stationarity ?

Open Data

Données de modèle de prévision d'ensemble arpege

donneespubliques.meteofrance.fr

ECMWF Data

{
    "url"   : "https://api.ecmwf.int/v1",
    "key"   : "XXX",
    "email" : "XXX"
}

https://confluence.ecmwf.int/display/WEBAPI/Access+ECMWF+Public+Datasets

See Public Datasets on https://apps.ecmwf.int/datasets/

pip install ecmwf-api-client

sudo apt -y install libgeos-dev