Skip to content

Archimedes Python SDK

archimedes is the name of the Python SDK for accessing Optimeering data.

Installation

You can install the SDK directly from PyPi:

pip install archimedes-python-client

Authentication

There are two options for authenticating towards the API when using the Python Client.

  • Using personal credentials
  • Using machine credentials

Personal Credentials:

To use personal credentials you first need to install the archimedes command line tool arcl.

pip install arcl

arcl auth login # then, follow the instructions to login with your Microsoft account

Once logged in, API login credentials will be stored in the ~/.archimedes directory.

Machine Credentials:

To setup Machine Credentials (ie. Service Account credentials) you first need to obtain an API Secret. Only Group owners have the ability to create API Secrets.

  1. Go to the User Portal - https://user-admin.fabapps.io/index/
  2. Click on the Group for your Company and then the View API Secrets button
  3. Click on the Add New button if you want to generate a new Secret
  4. Take note of the:
    1. Directory (tenant) ID
    2. Application (client) ID
    3. The Secret value

To use the generated credentials set the following environment variables before running invoking the Python Client: 

export AZURE_AD_TENANT_ID=<<replace with Dirctory(tenant) ID from above>>
export AZURE_AD_APP_ID=<<replace with Application (client) ID from above>>
export AZURE_AD_APP_CLIENT_CREDENTIAL=<<replace with Secret Value>>
export USE_APP_AUTHENTICATION=True

SDK Reference

rk_within_day.comparison_by_price

(*start: str | Timestamp | datetime*, *end: str | Timestamp | datetime*, *price_area: str*, *ref_price: int*, *ref_dt: str | Timestamp | datetime | None = None*, **, model_name: str= None*, *access_token: str | None = None*, ***kwargs*) → DataFrame

Get the probability of the RK price being less than or equal to the user specified reference price for each hour between start and end (inclusive).

Parameters:

  • start – The first hour to fetch prediction for (inclusive).
  • end – The last time to fetch prediction for (inclusive).
  • price_area – The name of the price area (eg - ‘NO2’, ‘SE4’).
  • ref_price – The reference price in € for price comparison.
  • ref_dt – The time of the latest RK data to include in the comparison. If None, the latest available reference date is used.
  • model_name - Name of prognosis model name to fetch data from.
  • access_token – Access token

rk_within_day.directions

(*start: str | Timestamp | datetime*, *end: str | Timestamp | datetime*, *price_area: str*, *ref_dt: str | Timestamp | datetime | None = None*, **, model_name: str= None*, *access_token: str | None = None*, ***kwargs*) → DataFrame

The predicted probability of any given hour is up(UP), down(DOWN) or no direction(NONE).

Parameters:

  • start – The first hour to fetch prediction for (inclusive).
  • end – The last time to fetch prediction for (inclusive).
  • price_area – The name of the price area (eg - ‘NO2’, ‘SE4’).
  • ref_dt – The time of the latest RK data to include in the forecast. If not provided, all the data for the given hours with all available ref_dt will be returned.
  • model_name - Name of prognosis model name to fetch data from.
  • access_token – Access token

rk_within_day.distributions

(*start: str | Timestamp | datetime*, *end: str | Timestamp | datetime*, *price_area: str*, *ref_dt: str | Timestamp | datetime | None = None*, **, model_name: str= None*, *access_token: str | None = None*, ***kwargs*) → DataFrame

The predicted imbalance price distribution. In the response, spread is defined as the difference between the RK price and the spot price (priceRK - pricespot).

Parameters:

  • start – The first hour to fetch prediction for (inclusive).
  • end – The last time to fetch prediction for (inclusive).
  • price_area – The name of the price area (eg - ‘NO2’, ‘SE4’).
  • ref_dt – The time of the latest RK data to include in the forecast. If not provided, all the data for the given hours with all available ref_dt will be returned.
  • model_name - Name of prognosis model name to fetch data from.
  • access_token – Access token

rk_within_day.large_down_fps

(*start: str | Timestamp | datetime*, *end: str | Timestamp | datetime*, *price_area: str*, *large_price_level: int*, *ref_dt: str | Timestamp | datetime | None = None*, *conditional: bool = True*, **, model_name: str= None*, *access_token: str | None = None*, ***kwargs*) → DataFrame

Indicator for whether a given hour will experience a large down RK price.

Parameters:

  • start – The first hour to fetch prediction for (inclusive).
  • end – The last time to fetch prediction for (inclusive).
  • price_area – The name of the price area (eg - ‘NO2’, ‘SE4’).
  • large_price_level – The abs(priceRK - priceSpot) spread in €/MWh above which price the RK-spot spread is considered large. Supported values are 20, 50 and 100.
  • ref_dt – The time of the latest RK data to include in the forecast. If not provided, all the data for the given hours with all available ref_dt will be returned.
  • conditional – Whether to fetch conditional data (one of ‘True’ and ‘False’). Defaults to True.
  • model_name - Name of prognosis model name to fetch data from.
  • access_token – Access token

rk_within_day.large_up_fps

(*start: str | Timestamp | datetime*, *end: str | Timestamp | datetime*, *price_area: str*, *large_price_level: int*, *ref_dt: str | Timestamp | datetime | None = None*, *conditional: bool = True*, **, model_name: str= None*, *access_token: str | None = None*, ***kwargs*) → DataFrame

Indicator for whether a given hour will experience a large up RK price.

Parameters:

  • start – The first hour to fetch prediction for (inclusive).
  • end – The last time to fetch prediction for (inclusive).
  • price_area – The name of the price area (eg - ‘NO2’, ‘SE4’).
  • large_price_level – The abs(priceRK - priceSpot) spread in €/MWh above which price the RK-spot spread is considered large. Supported values are 20, 50 and 100.
  • ref_dt – The time of the latest RK data to include in the forecast. If not provided, all the data for the given hours with all available ref_dt will be returned.
  • conditional – Whether to fetch conditional data (one of ‘True’ and ‘False’). Defaults to True.
  • model_name - Name of prognosis model name to fetch data from.
  • access_token – Access token

quantile_average

(start: str | Timestamp | datetime, end: str | Timestamp | datetime.datetime, price_area: str, ref_dt: str | Timestamp | datetime | None = None, *, model_name: str= None, access_token: str = None, **kwargs) -> DataFrame*

Gets the average of predicted prices over different quantile.

Parameters:

  • start - The first hour to fetch prediction for (inclusive).
  • end - The last time to fetch prediction for (inclusive).
  • price_area - The name of the price area (eg - 'NO2', 'SE4').
  • ref_dt - The time of the latest RK data to include in the comparison. If None, the latest available reference date is used.
  • model_name - Name of prognosis model name to fetch data from.
  • access_token - Access token
Powered by the Engineering Team