matlabtopython-compat 0.1.1

MATLAB runtime for Python: cell arrays, structs, sprintf, column-major flatten. The package that makes ported MATLAB code actually run.

matlabtopython-compat

The Python runtime for ported MATLAB code.

When you translate MATLAB to Python, some idioms don't map cleanly to numpy or scipy alone:

• Cell arrays with grow-on-write semantics
• Structs that support both s.name and s.('name') access
• sprintf cycling format specifiers through vector arguments
• X(:) column-major flatten (numpy defaults to row-major)
• [sorted_vals, idx] = sort(X) returning a pair

This package fills those gaps so ported code runs without hand-editing every MATLAB-ism into a numpy equivalent. Pair it with the online converter at mtopython.com for a zero-edit path, or use it directly if you're porting MATLAB by hand.

Install

pip install matlabtopython-compat

Optional .mat file I/O adds a scipy dependency:

pip install "matlabtopython-compat[mat-io]"

What's in it

Helper	MATLAB equivalent	Why it's here
CellArray	c = {a, b, c}	Lists work for most cells but miss c{1} = ... grow-on-write semantics.
Struct	s.name, s.(name)	Combines attribute + dict access in one object.
sprintf	sprintf('%d\n', [1 2 3])	Vector arguments cycle through format specifiers — Python's % doesn't.
fprintf	fprintf(fid, fmt, args)	Same, with file-handle form.
tic / toc	tic; ... toc	MATLAB-compatible timing output.
flatten_fortran	X(:)	Column-major flatten (numpy defaults to row-major).
sort_with_index	[s, i] = sort(X)	Returns both sorted values and indices.

Example

from matlabtopython_compat import CellArray, Struct, sprintf, tic, toc

# Cells with grow-on-write
c = CellArray()
c.cell_set(1, 'Alice')
c.cell_set(2, 42)
print(c.cell_get(1))  # 'Alice'

# Struct with both access styles
s = Struct(name='Alice', age=30)
print(s.name, s['age'])

# Vector-argument sprintf
print(sprintf('%d\n', [1, 2, 3]))  # '1\n2\n3\n'

# MATLAB-style timing
t0 = tic()
# ... code ...
toc()  # "Elapsed time is X seconds."

License

MIT. No runtime calls to external services; everything is pure Python.