`python_util.h`¶

Defines

Functions

static inline PyLongObject *PyLong_New(std::size_t python_digits)¶: Creation of a new PyLongObject that can be returned to Python

static inline bool python_long_is_negative(const nanobind::int_ &py_long_int)¶: Test if Python long integer (nb::int_) is negative

static std::vector<apy_limb_t> limb_vec_from_py_long_vec(const std::size_t count, const PyLongObject *py_long)¶: Convert Python limbs (of length PYLONG_BITS_IN_DIGIT) to apy_type_t limbs (of length APY_LIMB_SIZE_BITS).

static inline std::vector<apy_limb_t> python_long_to_limb_vec(const nanobind::int_ &py_long_int, std::optional<std::size_t> n_min_limbs = std::nullopt)¶: Python arbitrary long integer object to apy_limb_t vector. If n_min_limbs is set, at least that many limbs will be available in the result.

template<class RANDOM_ACCESS_ITERATOR> static inline nanobind::int_ python_limb_vec_to_long(RANDOM_ACCESS_ITERATOR begin, RANDOM_ACCESS_ITERATOR end, bool vec_is_signed = false, std::optional<unsigned> bits_last_limb = std::nullopt)¶: Convert a limb vector (std::vector<apy_limb_t>) to a Python long integer object wrapped in a nanobind::int_.

template<typename ...PyTypes> static inline bool is_iterable_and_exclude(const nanobind::handle &obj)¶: Test whether a nb::object is a Python iterable, and simultaneously that the object is not any of the PyArgs types.

static inline std::vector<std::size_t> _python_iterable_extract_shape_recursive_descent(const nanobind::iterable &sequence, std::string_view err_prefix)¶: Retrieve the sequence of a python sequence object through recursive descent along the dimensions.

template<bool IS_COMPLEX_COLLAPSE = false> static inline std::vector<std::size_t> python_iterable_extract_shape(const nanobind::iterable &seq, std::string_view err_prefix)¶: Retrieve the shape of a possibly nested Python sequence of iterable object. When IS_COMPLEX_COLLAPSE is true, the last dimension is truncated if it is exactly equal to two.

template<typename ...PyTypes> static std::vector<nanobind::object> python_iterable_walk(const nanobind::iterable &py_seq, std::string_view err_prefix)¶: Walk a (possibly nested) Python iterable sequence, py_seq, of iterable objects and store handles to them them in a std::vector<nb::object>. The sequence is walked in a depth-first manner and all elements must match (nb::isinstance<PyTypes> || ...). If any object in the sequence py_seq does not match any PyTypes (or another Python sequence) a std::domain_error exception is raised.

template<typename ...PyTypes> static std::vector<nanobind::object> python_sequence_walk(const nanobind::sequence &py_seq, std::string_view err_prefix)¶: See: python_iterable_walk()

template<bool IS_COMPLEX_COLLAPSE = false> static inline std::vector<std::size_t> python_sequence_extract_shape(const nanobind::sequence &py_seq, std::string_view err_prefix)¶: See: python_iterable_extract_shape()

python_util.h¶