Skip to main content
The Figranium Python SDK raises one exception class for every failure: FigraniumError. It covers HTTP error responses, network failures, timeouts, and request aborts. You can branch on error.status for HTTP semantics, on error.code for machine-readable categories, and on error.__cause__ for the original transport exception.

FigraniumError shape

FigraniumError extends the standard Exception class, so .message and the traceback are available as usual.

Basic handling

Fields

The HTTP status returned by the Figranium server, if a response arrived. 0 when the request never reached a status, such as a network failure, DNS error, or cancelled request.
Server-provided machine-readable error code, taken from the response body’s error field. Also used for SDK-generated conditions:
Free-form value from the response body’s details or detail field. Typically an object describing which validation failed, which record was missing, or which field was invalid.
Value of the x-request-id response header, when the server or a proxy sets one. Include this in bug reports to make server-side logs easier to correlate.
The original httpx.Response object, when the error came from an HTTP response. Available for advanced consumers who need the raw headers or body.

Branching on status

Transport failures

When the request never gets a response (server unreachable, DNS failure, TLS error), status is 0 and code is NETWORK_ERROR. The __cause__ property carries the underlying httpx.HTTPError.

Timeouts and cancellation

When a request exceeds the configured timeout, the SDK raises FigraniumError with code="REQUEST_ABORTED". The __cause__ is the httpx.TimeoutException.
Always check isinstance(error, FigraniumError) before reading SDK-specific fields. Unknown errors should be rethrown so they bubble up to your global error handler.