Skip to content

Fix custom float dtype f-string formatting truncating exponent notation #369

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
xhmelon wants to merge 2 commits into
base: main
Choose a base branch
from
+33 −0
Open

Fix custom float dtype f-string formatting truncating exponent notation #369

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f1ccc47
Fix custom float dtype (e.g. bfloat16) f-string formatting truncating…
xhmelon Mar 23, 2026
5b9d2a6
Remove redundant type conversion.
xhmelon Mar 23, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 33 additions & 0 deletions ml_dtypes/_src/custom_float.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -353,6 +353,38 @@ Py_hash_t PyCustomFloat_Hash(PyObject* self) {
return HashImpl(&_Py_HashDouble, self, static_cast<double>(x));
}

template <typename T>
PyObject* PyCustomFloat_Format(PyObject* self, PyObject* args) {
PyObject* format_spec = nullptr;
if (!PyArg_ParseTuple(args, "U", &format_spec)) {
return nullptr;
}

T x = reinterpret_cast<PyCustomFloat<T>*>(self)->value;
// Round to 6 significant digits to match PyCustomFloat_Str/Repr, which
// use std::ostringstream with its default precision of 6. This avoids
// exposing false precision from the float64 expansion.
char buf[14]; // max %.6g output: "-9.98378e+38" + '\0' = 14
Copy link
Copy Markdown
Collaborator

@hawkinsp hawkinsp Mar 25, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm finding this part strange. For any floating point type, numpy appears to do the following: https://github.com/numpy/numpy/blob/f105cf2d7c20c9829b431c2ae0cdb1f07efaccf2/numpy/_core/src/multiarray/scalartypes.c.src#L633

i.e., convert the type to a python float using nb_format and then call PyObject_Format on the result.

Is there a reason for us to deviate here?

Copy link
Copy Markdown
Collaborator

@hawkinsp hawkinsp Mar 30, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What do you think?

std::snprintf(buf, sizeof(buf), "%.6g", static_cast<double>(x));
double d = std::strtod(buf, nullptr);

PyObject* float_obj = PyFloat_FromDouble(d);
if (!float_obj) {
return nullptr;
}

PyObject* result = PyObject_Format(float_obj, format_spec);
Py_DECREF(float_obj);
return result;
}

template <typename T>
PyMethodDef CustomFloatType_methods[] = {
{"__format__", PyCustomFloat_Format<T>, METH_VARARGS,
"Format a custom float value."},
{nullptr, nullptr, 0, nullptr},
};

template <typename T>
PyType_Slot CustomFloatType<T>::type_slots[] = {
{Py_tp_new, reinterpret_cast<void*>(PyCustomFloat_New<T>)},
Expand All @@ -362,6 +394,7 @@ PyType_Slot CustomFloatType<T>::type_slots[] = {
{Py_tp_doc,
reinterpret_cast<void*>(const_cast<char*>(TypeDescriptor<T>::kTpDoc))},
{Py_tp_richcompare, reinterpret_cast<void*>(PyCustomFloat_RichCompare<T>)},
{Py_tp_methods, reinterpret_cast<void*>(CustomFloatType_methods<T>)},
{Py_nb_add, reinterpret_cast<void*>(PyCustomFloat_Add<T>)},
{Py_nb_subtract, reinterpret_cast<void*>(PyCustomFloat_Subtract<T>)},
{Py_nb_multiply, reinterpret_cast<void*>(PyCustomFloat_Multiply<T>)},
Expand Down
Loading