Component configuration

by Aleksey Dvoykin

Introduction

The Pip.Services Toolkit offers a simple but very flexible mechanism for component configuration. Configurations can be loaded from various sources - configuration files, command line parameters, environment variables, configuration services, etc. Once loaded, they are passed to the specific component, which configures itself accordingly. In this recipe, we’ll be taking a look at this mechanism’s capabilities and how it can be utilized.

Configuration

The configurable interface

A component can be made configurable by adding the IConfigurable interface and implementing its configure method. This method will be called by the container right after container creation, with the loaded configuration being passed as a parameter.

interface IConfigurable {
	configure(config: ConfigParams): void;
}

Not available

import (
	"context"

	conf "github.com/pip-services4/pip-services4-go/pip-services4-components-go/config"
)

type IConfigurable interface {
	Configure(ctx context.Context, config *conf.ConfigParams)
}

Not available

See: IConfigurable

from abc import ABC
from pip_services4_components.config import ConfigParams

class IConfigurable(ABC):

    def configure(self, config: ConfigParams):
        raise NotImplementedError('Method from interface definition')

Not available

The ConfigParams object

The only parameter that is passed to the configure method is ConfigParams object. Simply put - this is a map that allows us to get a configuration parameter value by its corresponding key. Although various programming languages have unique syntax for initializing maps and objects, ConfigParams support initialization that is independent of the language being used.

import { ConfigParams } from "pip-services4-components-node"

let config = ConfigParams.fromTuples(
    "param1", 123,
    "param2", "2020-01-01T11:00:00.0Z"
);

Not available

import (
	cconfig "github.com/pip-services4/pip-services4-go/pip-services4-components-go/config"
)

config := cconfig.NewConfigParamsFromTuples(
	"param1", 123,
	"param2", "2020-01-01T11:00:00.0Z",
)

Not available

See: ConfigParams

config = ConfigParams.from_tuples(
  "param1", 123,
  "param2", "2020-01-01T11:00:00.0Z"
)

Not available

ConfigParams also provide some additional functionality. All keys and values are stored as strings, but ConfigParams supports performing data type conversion when extracting values. Another option available is the opportunity to set default values.

let param1 = config.getAsInteger("param1");
let param2 = config.getAsDateTimeWithDefault("param2", new Date());

Not available

param1 := config.GetAsInteger("param1")
param2 := config.GetAsDateTimeWithDefault("param2", time.Now())

Not available

from datetime import datetime

param1 = config.get_as_integer("param1")
param2 = config.get_as_datetime_with_default("param2", datetime.now())

Not available

The parameter kets can have a complex structure, grouped by sections using dot notation. ConfigParams can be used to work with entire sections as well.

let configWithSections = ConfigParams.fromTuples(
    "param1", 123,
    "options.param1", "ABC",
    "options.param2", "XYZ"
);
let options = configWithSections.getSection("options");

Not available

configWithSections := cconfig.NewConfigParamsFromTuples(
	"param1", 123,
	"options.param1", "ABC",
	"options.param2", "XYZ",
)
options := configWithSections.GetSection("options")

Not available

config_with_sections = ConfigParams.from_tuples(
  "param1", 123,
  "options.param1", "ABC",
  "options.param2", "XYZ"
)
options = config_with_sections.get_section("options")

Not available

Setting a default configuration

Another helpful option is the ability to set a default configuration.


let defaultConfig = ConfigParams.fromTuples(
  	"param1", 1,
  	"param2", "Default Value"
);
config = config.setDefaults(defaultConfig);

Not available

defaultConfig := cconfig.NewConfigParamsFromTuples(
	"param1", 1,
	"param2", "Default Value",
)
config = config.SetDefaults(defaultConfig)

Not available

default_config = ConfigParams.from_tuples(
  "param1", 1,
  "param2", "Default Value"
)
config = config.set_defaults(default_config)

Not available

Serializing the ConfigParams object

Lastly, ConfigParams objects can be serialized/deserialized to/from JSON, YAML, or a plain string.

let anotherConfig = ConfigParams.fromString("param1=123;param2=ABC");

Not available

anotherConfig := cconfig.NewConfigParamsFromString("param1=123;param2=ABC")

Not available

See: Commons module’s

another_config = ConfigParams.from_string("param1=123;param2=ABC")

Not available

To read more about what functionality is available through ConfigParams, be sure to check out the Commons documentation.

Example

Below is an example of a configurable component:

export class DataController implements IConfigurable {
    private _max_page_size: number = 5;
    public constructor() { }

    public configure(config: ConfigParams): void {
     this._max_page_size = config.getAsIntegerWithDefault('max_page_size', this._max_page_size);
    }

    public getData(ctx: Context, filter: FilterParams, paging: PagingParams): Promise<DataPage<MyData>> {
     paging.take = Math.min(paging.take, this._max_page_size);  
          // Get data using max page size constraint.
    }
}

Not available

import (
	"context"
	"math"

	cconfig "github.com/pip-services4/pip-services4-go/pip-services4-components-go/config"
	cquery "github.com/pip-services4/pip-services4-go/pip-services4-data-go/query"
)

type DataService struct {
	_maxPageSize int
}

func NewDataService() *DataService {
	return &DataService{
		_maxPageSize: 5,
	}
}

func (c *DataService) Configure(ctx context.Context, config *cconfig.ConfigParams) {
	c._maxPageSize = config.GetAsIntegerWithDefault("max_page_size", c._maxPageSize)
}

func (c *DataService) GetData(ctx context.Context, correlationId string, filter *cquery.FilterParams, paging *cquery.PagingParams) (page *cquery.DataPage, err error) {
	paging.Take = int64(math.Min(float64(paging.Take), float64(c._maxPageSize)))
	// Get data using max page size constraint.
	return page, err
}

Not available

See: Commons module’s

from pip_services4_data.query import FilterParams, PagingParams, DataPage

class DataController(IConfigurable):
   __max_page_size: int = 5
   def configure(self, config: ConfigParams):
       self.__max_page_size = config.get_as_integer_with_default('max_page_size', self.__max_page_size)
   
   def get_data(self, correlation_id: str, filter: FilterParams, paging: PagingParams) -> DataPage: 
       paging.take = min(paging.take, self.__max_page_size)   # Get data using max page size constraint.

Not available

Manual configuration

Manual configuration can be done in the following manner:

let component = new DataController();
let config = ConfigParams.fromTuples("max_page_size", 100);
component.configure(config);

Not available

component := configurationexample.NewDataService()
config := cconfig.NewConfigParamsFromTuples("max_page_size", 100)
component.Configure(context.Background(), config)

Not available

component = DataController()
config = ConfigParams.from_tuples("max_page_size", 100)
component.configure(config)

Not available

However, a component’s configuration is usually stored in the microservice’s configuration file. The configure method will receive the parameters for the specific component (in the example below - everything between the component’s descriptor and the next descriptor or end of the file). To get more info on microservice configuration, read our Component Container recipe.

...
# Controller
- descriptor: "beacons:controller:default:default:1.0"
  max_page_size: 10
...

Resolvers

The NameResolver and OptionsResolver classes are helper classes that simplify the use of configurations.

NameResolver

NameResolver is a helper class that allows extraction of the component’s name from the configuration parameters. The name can be stored as a “id”/“name” parameter or inside a component’s descriptor.

Below is a simple example of how it can be used:

import { ConfigParams, NameResolver } from "pip-services4-components-node"

let config = ConfigParams.fromTuples(
	"descriptor", "myservice:connector:aws:connector1:1.0",
	"param1", "ABC",
	"param2", 123
);

let name = NameResolver.resolve(config); // Result: connector1

Not available

config := cconfig.NewConfigParamsFromTuples(
	"descriptor", "myservice:connector:aws:connector1:1.0",
	"param1", "ABC",
	"param2", 123,
)
name := cconfig.NameResolver.Resolve(config) // Result: connector1

Not available

See: NameResolver, OptionsResolver

from pip_services4_components.config import NameResolver

config = ConfigParams.from_tuples(
	"descriptor", "myservice:connector:aws:connector1:1.0",
	"param1", "ABC",
	"param2", 123
)
name = NameResolver.resolve(config) # Result: connector1

Not available

OptionsResolver

OptionsResolver is a helper class that extracts parameters from the “options” configuration section.

let config = ConfigParams.fromTuples(
	...
	"options.param1", "ABC",
	"options.param2", 123
);
let options = OptionsResolver.resolve(config); // Result: param1=ABC;param2=123

Not available

config := cconfig.NewConfigParamsFromTuples(
	...
	"options.param1", "ABC",
	"options.param2", 123,
)
options := cconfig.OptionsResolver.Resolve(config) // Result: param1=ABC;param2=123

Not available

from pip_services4_components.config import OptionsResolver

config = ConfigParams.from_tuples(
    #...
	"options.param1", "ABC",
	"options.param2", 123
)
options = OptionsResolver.resolve(config)   # Result: param1=ABC;param2=123

Not available

Configuration readers

Configuration parameters can be stored in microservice configurations, configuration files, or in configuration services. To help with configuration extraction, the Pip.Services Toolkit offers two special ConfigReader components. The interface for these components is defined in the Components module.

interface IConfigReader {
    readConfig(context: IContext, parameters: ConfigParams): Promise<ConfigParams>;
    addChangeListener(listener: INotifiable): void;
    removeChangeListener(listener: INotifiable): void;
}

Not available

import (
	"context"

	cconfig "github.com/pip-services4/pip-services4-go/pip-services4-components-go/config"
)

type IConfigReader interface {
	ReadConfig(ctx context.Context, correlationId string, parameters *cconfig.ConfigParams) (*cconfig.ConfigParams, error)
}

Not available

See: Components

from typing import Optional

class IConfigReader(ABC):

    def read_config_(self, correlation_id: Optional[str], parameters: ConfigParams) -> ConfigParams:
        raise NotImplementedError('Method from interface definition')

Not available

MemoryConfigReader

The MemoryConfigReader is a ConfigReader that stores configuration data in memory.

let config = ConfigParams.fromTuples(
	"connection.host", "localhost",
	"connection.port", "8080"
);

let configReader = new MemoryConfigReader();
configReader.configure(config);

let parameters = ConfigParams.fromValue(process.env);
let result = await configReader.readConfig(ctx, parameters); 
// Result: connection.host=localhost;connection.port=8080

Not available

import (
	"context"
	"os"

	cconfig "github.com/pip-services4/pip-services4-go/pip-services4-components-go/config"
	creader "github.com/pip-services4/pip-services4-go/pip-services4-config-go/config"
)


config := cconfig.NewConfigParamsFromTuples(
	"connection.host", "localhost",
	"connection.port", "8080",
)

configReader := creader.NewEmptyMemoryConfigReader()
configReader.Configure(context.Background(), config)
parameters := cconfig.NewConfigParamsFromValue(os.Args)
configReader.ReadConfig(context.Background(), parameters) // Result: connection.host=localhost;connection.port=8080

Not available

See: MemoryConfigReader, ConfigReader

import sys

from pip_services4_config.config import MemoryConfigReader

config = ConfigParams.from_tuples(
	"connection.host", "localhost",
	"connection.port", "8080"
)
config_reader = MemoryConfigReader()
config_reader.configure(config)
parameters = ConfigParams.from_value(sys.argv)
config_reader.read_config_("123", parameters) # Result: connection.host=localhost;connection.port=8080

Not available

JsonConfigReader

The JsonConfigReader is a ConfigReader that can read configurations from a JSON file.

{ "key1": "{{KEY1_VALUE}}", "key2": "{{KEY2_VALUE}}" }

let configReader = new JsonConfigReader("config.json");
let parameters = ConfigParams.fromTuples("KEY1_VALUE", 123, "KEY2_VALUE", "ABC");
let result = await configReader.readConfig(ctx, parameters); // Result: key1=123;key2=ABC

Not available

import (
	"context"

	cconfig "github.com/pip-services4/pip-services4-go/pip-services4-components-go/config"
	creader "github.com/pip-services4/pip-services4-go/pip-services4-config-go/config"
)

configReader := creader.NewJsonConfigReader("config.json")
parameters := cconfig.NewConfigParamsFromTuples("KEY1_VALUE", 123, "KEY2_VALUE", "ABC")
configReader.ReadConfig(context.Background(), parameters) // Result: key1=123;key2=ABC

Not available

See: JsonConfigReader

from pip_services4_config.config import JsonConfigReader

configReader = JsonConfigReader("config.json")
parameters = ConfigParams.from_tuples("KEY1_VALUE", 123, "KEY2_VALUE", "ABC")
configReader.read_config_("123", parameters)    # Result: key1=123;key2=ABC

Not available

YamlConfigReader

The YamlConfigReader is a ConfigReader that can read configurations from a YAML file.

key1: "1234"
key2: "ABCD"

let configReader = new YamlConfigReader("config.yml");
let parameters = ConfigParams.fromTuples("KEY1_VALUE", 123, "KEY2_VALUE", "ABC");
let result = await configReader.readConfig(ctx, parameters); // Result: key1=1234;key2=ABCD

Not available

import (
	"context"

	cconfig "github.com/pip-services4/pip-services4-go/pip-services4-components-go/config"
	creader "github.com/pip-services4/pip-services4-go/pip-services4-config-go/config"
)

configReader := creader.NewYamlConfigReader("config.yml")
parameters := cconfig.NewConfigParamsFromTuples("KEY1_VALUE", 123, "KEY2_VALUE", "ABC")
configReader.ReadConfig(context.Background(), parameters) // Result: key1=123;key2=ABC

Not available

See: YamlConfigReader

from pip_services4_config.config import YamlConfigReader

configReader = YamlConfigReader("config.yml")
parameters2 = ConfigParams.from_tuples("KEY1_VALUE", 123, "KEY2_VALUE", "ABC")
configReader.read_config_("123", parameters2)    # Result: key1=1234;key2=ABCD

Not available