singer_sdk.RESTStream

class singer_sdk.RESTStream

Abstract base class for REST API streams.

__init__(tap: TapBaseClass, name: str | None = None, schema: dict[str, Any] | Schema | None = None, path: str | None = None) None

Initialize the REST stream.

Parameters
  • tap – Singer Tap this stream belongs to.

  • schema – JSON schema for records in this stream.

  • name – Name of this stream.

  • path – URL path for this entity stream.

property authenticator: APIAuthenticatorBase | None

Return or set the authenticator for managing HTTP auth headers.

If an authenticator is not specified, REST-based taps will simply pass http_headers as defined in the stream class.

Returns

Authenticator instance that will be used to authenticate all outgoing requests.

backoff_handler(details: Details) None

Adds additional behaviour prior to retry.

By default will log out backoff details, developers can override to extend or change this behaviour.

Parameters

details – backoff invocation details https://github.com/litl/backoff#event-handlers

backoff_max_tries() int

The number of attempts before giving up when retrying requests.

Returns

Number of max retries.

backoff_runtime(*, value: Callable[[Any], int]) Generator[int, None, None]

Optional backoff wait generator that can replace the default backoff.expo.

It is based on parsing the thrown exception of the decorated method, making it possible for response values to be in scope.

Parameters

value – a callable which takes as input the decorated function’s thrown exception and determines how long to wait.

Yields

The thrown exception

backoff_wait_generator() Generator[float, None, None]

The wait generator used by the backoff decorator on request failure.

See for options: https://github.com/litl/backoff/blob/master/backoff/_wait_gen.py

And see for examples: Code Samples

Returns

The wait generator

build_prepared_request(*args: Any, **kwargs: Any) PreparedRequest

Build a generic but authenticated request.

Uses the authenticator instance to mutate the request with authentication.

Parameters
Returns

A requests.PreparedRequest object.

calculate_sync_cost(request: requests.PreparedRequest, response: requests.Response, context: dict | None) dict[str, int]

Calculate the cost of the last API call made.

This method can optionally be implemented in streams to calculate the costs (in arbitrary units to be defined by the tap developer) associated with a single API/network call. The request and response objects are available in the callback, as well as the context.

The method returns a dict where the keys are arbitrary cost dimensions, and the values the cost along each dimension for this one call. For instance: { “rest”: 0, “graphql”: 42 } for a call to github’s graphql API. All keys should be present in the dict.

This method can be overridden by tap streams. By default it won’t do anything.

Parameters
  • request – the API Request object that was just called.

  • response – the requests.Response object

  • context – the context passed to the call

Returns

A dict of accumulated costs whose keys are the “cost domains”.

extra_retry_statuses: list[int] = [429]

Response code reference for rate limit retries

get_new_paginator() BaseAPIPaginator

Get a fresh paginator for this API endpoint.

Returns

A paginator instance.

get_records(context: dict | None) Iterable[dict[str, Any]]

Return a generator of record-type dictionary objects.

Each record emitted should be a dictionary of property names to their values.

Parameters

context – Stream partition or context dictionary.

Yields

One item per (possibly processed) record in the API.

get_url(context: dict | None) str

Get stream entity URL.

Developers override this method to perform dynamic URL generation.

Parameters

context – Stream partition or context dictionary.

Returns

A URL, optionally targeted to a specific partition or context.

get_url_params(context: dict | None, next_page_token: _TToken | None) dict[str, Any]

Return a dictionary of values to be used in URL parameterization.

If paging is supported, developers may override with specific paging logic.

Parameters
  • context – Stream partition or context dictionary.

  • next_page_token – Token, page number or any request argument to request the next page of data.

Returns

Dictionary of URL query parameters to use in the request.

property http_headers: dict

Return headers dict to be used for HTTP requests.

If an authenticator is also specified, the authenticator’s headers will be combined with http_headers when making HTTP requests.

Returns

Dictionary of HTTP headers to use as a base for every request.

next_page_token_jsonpath: str | None = None

Optional JSONPath expression to extract a pagination token from the API response. Example: “$.next_page”

parse_response(response: Response) Iterable[dict]

Parse the response and return an iterator of result records.

Parameters

response – A raw requests.Response object.

Yields

One item for every item found in the response.

prepare_request(context: dict | None, next_page_token: _TToken | None) requests.PreparedRequest

Prepare a request object for this stream.

If partitioning is supported, the context object will contain the partition definitions. Pagination information can be parsed from next_page_token if next_page_token is not None.

Parameters
  • context – Stream partition or context dictionary.

  • next_page_token – Token, page number or any request argument to request the next page of data.

Returns

Build a request with the stream’s URL, path, query parameters, HTTP headers and authenticator.

prepare_request_payload(context: dict | None, next_page_token: _TToken | None) dict | None

Prepare the data payload for the REST API request.

By default, no payload will be sent (return None).

Developers may override this method if the API requires a custom payload along with the request. (This is generally not required for APIs which use the HTTP ‘GET’ method.)

Parameters
  • context – Stream partition or context dictionary.

  • next_page_token – Token, page number or any request argument to request the next page of data.

Returns

Dictionary with the body to use for the request.

records_jsonpath: str = '$[*]'

JSONPath expression to extract records from the API response.

request_decorator(func: Callable) Callable

Instantiate a decorator for handling request failures.

Uses a wait generator defined in backoff_wait_generator to determine backoff behaviour. Try limit is defined in backoff_max_tries, and will trigger the event defined in backoff_handler before retrying. Developers may override one or all of these methods to provide custom backoff or retry handling.

Parameters

func – Function to decorate.

Returns

A decorated method.

request_records(context: dict | None) Iterable[dict]

Request records from REST endpoint(s), returning response records.

If pagination is detected, pages will be recursed automatically.

Parameters

context – Stream partition or context dictionary.

Yields

An item for every record in the response.

property requests_session: Session

Get requests session.

Returns

The requests.Session object for HTTP requests.

response_error_message(response: Response) str

Build error message for invalid http statuses.

WARNING - Override this method when the URL path may contain secrets or PII

Parameters

response – A requests.Response object.

Returns

The error message

Return type

str

property timeout: int

Return the request timeout limit in seconds.

The default timeout is 300 seconds, or as defined by DEFAULT_REQUEST_TIMEOUT.

Returns

The request timeout limit as number of seconds.

update_sync_costs(request: requests.PreparedRequest, response: requests.Response, context: dict | None) dict[str, int]

Update internal calculation of Sync costs.

Parameters
  • request – the Request object that was just called.

  • response – the requests.Response object

  • context – the context passed to the call

Returns

A dict of costs (for the single request) whose keys are the “cost domains”. See calculate_sync_cost for details.

abstract property url_base: str

Return the base url, e.g. https://api.mysite.com/v3/.

validate_response(response: Response) None

Validate HTTP response.

Checks for error status codes and wether they are fatal or retriable.

In case an error is deemed transient and can be safely retried, then this method should raise an singer_sdk.exceptions.RetriableAPIError. By default this applies to 5xx error codes, along with values set in: extra_retry_statuses

In case an error is unrecoverable raises a singer_sdk.exceptions.FatalAPIError. By default, this applies to 4xx errors, excluding values found in: extra_retry_statuses

Tap developers are encouraged to override this method if their APIs use HTTP status codes in non-conventional ways, or if they communicate errors differently (e.g. in the response body).

../_images/200.png
Parameters

response – A requests.Response object.

Raises