Classes & Methods

Optimizer

Handles the connection to the Global Optimizer.

Methods:

optimizer.activate(api_token: str, user: str, pw: str)

Parameters:

• api_token – Glaium API token for the organization
• user – Username (manager or owner) for the organization
• pw – Password for user login

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response: {}

optimizer.model(name: str, operation: str)

Parameters:

• name – Name for the model. The current available models are:
  • mobile_gaming
  • adtech
  • says (coming soon)
  • ecommerce (coming soon)
• operation – Operation type:
  • get – Load a specific Model.
  • set – Update or save the current systems set in a new or existent Model.
  • del – Delete a specific Model.

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – json containing all systems (agents) from the model: {agent_name_1:agent_id_1,…agent_name_N:agent_id_N}

optimizer.goal(name: str, description: str, objective: dict, constraints: dict, resources: dict)

Parameters:

• name – Goal name
• description – Goal description
• objective – dictionary containing {’system’:system_id, ’target’:’metric_name’, ’comparison’: comparison operator (>, ≥, =, ≤, <) , ’outcome’: value expected, ’by’:date}
• constraints – dictionary containing {system_id_1:conditionals,…,system_id_N:conditionals}
• resources – dictionary containing {system_id_1:{resource_1:availability, … resource_N:availability}, … system_id_N:{resource_1:availability, … resource_N:availability}}

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – goal_id

System (Agent)

Defines, orchestrate and optimize systems (autonomous agents).

Methods:

system.set(name: str, inputs: dict, output: dict, frequency: str, duration: int, constraints: str, resources: dict)

Parameters:

• name – Name for the system (agent)
• inputs – dictionary containing {system_id_1:units needed,…,system_id_N:units needed}
  • units needed – An specific amount (float) or np.inf for all that is available
• output – dictionary containing {’name’: output_name, ‘unity’: type of unity, ’cumulative’:yes/no for accumulation between cycles, ‘returns’; yes/no for returning units from other systems}
• frequency – when to start each cycle: ‘every hour’, ‘daily’, ‘every other day’, ‘every 1st day of month’…
• duration – max duration in seconds to use as watchdog timer.
• constraints – conditionals formula
• resources – dictionary containing {system_id_1:{resource_1:usage, … resource_N:usage}, … system_id_N:{resource_1:usage, … resource_N:usage}}

Returns: 

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – system_id (ok) or error message

system.optimization(system_id)

Parameters:

• system_id – system_id

Returns: 

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – Dictionary containing: {
    • ’next_cycle’ : cycle number,
    • ‘next_when’: datetime to start next cycle,
    • ‘output’ : units expected for output,
    • ‘inputs’: {‘input_1_name’ : units suggested for input 1, … ’input_N_name’ : units suggested for input N}}

system.start(system_id,cycle)

Parameters:

• system_id – Id of the system (agent) starting its process
• cycle (optional) – number of cycle starting

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – None (ok) or error message

system.stop(system_id, status, stock, cycle)

Parameters:

• system_id – Id of the system (agent) starting its process
• status – json containing {status_type:status_msg}
  • status_type: ok, halt (motive of halting process), hands-up (requires human intervention for <status_msg>) and error (type of error)
• stock – current volume of units available in the output
• cycle (optional) – number of cycle starting

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – None (ok) or error message

Data

Manages data extraction, transformation, and loading (ETL) via Glaium Cube process.

Methods:

data.integration(integration: dict, operation: str)

Parameters:

• integration – Dictionary containing: {name:integrator,… name_N:integrator}
  • name – Name of the Integration
  • integrator – Type of the Integration. See https://glaium.io/integrators/ for list available. Optional for pause, resume and delete operations.
• operation – Operation type
  • add – Add one or more new Integration(s)
  • pause – Pause one or more new Integration(s)
  • resume – Resume one or more new Integration(s)
  • delete – Delete one or more existent Integration(s)

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – None (ok) or error message

data.pull(metrics: lst, dimensions: lst, conditional: str, period: str)

Parameters:

• metrics – Numeric primitives
• dimensions – Categorical primitives
• conditional – formula
• period – initial and end datetimes or temporal expressions (Ex: “last 10 days”, “end of year”, “next 2 months”, “yesterday”, …)

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – pd.DataFrame with data or error message

data.enhance(primitive: str, how: str)

Parameters:

• primitive – Name of the Primitive to normalize
• how – Dictionary containing: {’enhancement_function’:’params’,…} with the following available functions:
  • normalize: list of terms to detect. Example: [term1,term2,…termN]
  • replace: dictionary containing: {search_value:replace_value,…}
  • conditional: dictionary containing: {’if_condition’:value,…’else’:default_value}

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – pd.DataFrame (ok) or error message

data.push(data, integrations, when, frequency)

Parameters:

• data – pd.DataFrame
• integrations – list of Integrations to push data to
• when – datetime to push data to the list of integrations.
• frequency – temporal expression. Ex: ‘every hour’, ‘daily’, ‘every other day’, ‘every 1st day of month’…

Returns:

• Dictionary containing: {error_code:response}
  • error code: 0 (ok) or not 0 (error)
  • response – None (ok) or error message