core

prairielearn.core ¶

Utilities for building questions and elements in PrairieLearn.

from prairielearn import ...
# or ...
from prairielearn.core import ...

from_json ¶

from_json(v: _JSONSerializedType | Any) -> Any

Converts a JSON serialized value (from to_json) back to its original type.

If v has the format {'_type': ..., '_value': ...} as would have been created using to_json(...), then it is replaced according to the following table:

JSON `_type` field	Python type
`complex`	`complex`
`np_scalar`	numpy scalar defined by `_concrete_type`
`ndarray`	non-complex `ndarray`
`complex_ndarray`	complex `ndarray`
`sympy`	`sympy.Expr`
`sympy_matrix`	`sympy.Matrix`
`dataframe`	`pandas.DataFrame`
`dataframe_v2`	`pandas.DataFrame`
`networkx_graph`	corresponding networkx graph
missing	input value v returned

If v encodes an ndarray and has the field '_dtype', this function recovers its dtype.

If v does not have the format {'_type': ..., '_value': ...}, then it is returned without change.

to_json ¶

to_json(
    v: Any | _JSONPythonType,
    *,
    df_encoding_version: Literal[1, 2] = 1,
    np_encoding_version: Literal[1, 2] = 1,
) -> Any | _JSONSerializedType

Convert a value to a JSON serializable format.

If v has a standard type that cannot be json serialized, it is replaced with a {'_type': ..., '_value': ...} pair that can be json serialized.

This is a complete table of the mappings:

Type	JSON `_type` field	notes
complex scalar	`complex`	including numpy
non-complex ndarray	`ndarray`	assumes each element can be json serialized
complex ndarray	`complex_ndarray`
`sympy.Expr`	`sympy`	any scalar sympy expression
`sympy.Matrix`	`sympy_matrix`
`pandas.DataFrame`	`dataframe`	`df_encoding_version=1`
`pandas.DataFrame`	`dataframe_v2`	`df_encoding_version=2`
networkx graph type	`networkx_graph`
numpy scalar	`np_scalar`	`np_encoding_version=2`
any	`v`	if v can be json serialized

Note

The 'dataframe_v2' encoding allows for missing and date time values whereas the 'dataframe' (default) does not. However, the 'dataframe' encoding allows for complex numbers while 'dataframe_v2' does not.

If np_encoding_version is set to 2, then numpy scalars serialize using '_type': 'np_scalar'.

If df_encoding_version is set to 2, then pandas DataFrames serialize using '_type': 'dataframe_v2'.

See from_json for details about the differences between encodings.

If v is an ndarray, this function preserves its dtype (by adding '_dtype' as a third field in the dictionary).

If v can be json serialized or does not have a standard type, then it is returned without change.