{ "__type": "IngestedDoc", "__tag": 4010, "_content": {}, "_ordered_sections": [], "item_file": null, "item_line": null, "item_type": null, "aliases": [], "example_section_data": { "__type": "Section", "__tag": 4015, "children": [], "title": [], "level": 0, "target": null }, "see_also": [], "signature": null, "references": null, "qa": "tutorial:io", "arbitrary": [ { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Admonition", "__tag": 4056, "kind": "seealso", "base_type": "note", "children": [ { "__type": "AdmonitionTitle", "__tag": 4055, "children": [ { "__type": "Text", "__tag": 4046, "value": "seealso `NumPy IO routines `__" } ] } ] } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "File IO (" }, { "__type": "InlineRole", "__tag": 4003, "value": "scipy.io", "domain": null, "role": "mod", "inventory": null }, { "__type": "Text", "__tag": 4046, "value": ")" } ], "level": 0, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Code", "__tag": 4050, "value": ".. autosummary:: \n loadmat\n savemat\n whosmat", "execution_status": null } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "MATLAB files" } ], "level": 1, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "We'll start by importing " }, { "__type": "CrossRef", "__tag": 4002, "value": "scipy.io", "reference": { "__type": "RefInfo", "__tag": 4000, "module": "scipy", "version": "*", "kind": "api", "path": "scipy.io" }, "kind": "module" }, { "__type": "Text", "__tag": 4046, "value": " and calling it " }, { "__type": "InlineCode", "__tag": 4051, "value": "sio" }, { "__type": "Text", "__tag": 4046, "value": " for convenience:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> import scipy.io as sio", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "If you are using IPython, try tab-completing on " }, { "__type": "InlineCode", "__tag": 4051, "value": "sio" }, { "__type": "Text", "__tag": 4046, "value": ". Among the many options, you will find " } ] }, { "__type": "Code", "__tag": 4050, "value": "sio.loadmat\nsio.savemat\nsio.whosmat", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "These are the high-level functions you will most likely use when working with MATLAB files. You'll also find " } ] }, { "__type": "Code", "__tag": 4050, "value": "sio.matlab", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "This is the package from which " }, { "__type": "InlineCode", "__tag": 4051, "value": "loadmat" }, { "__type": "Text", "__tag": 4046, "value": ", " }, { "__type": "InlineCode", "__tag": 4051, "value": "savemat" }, { "__type": "Text", "__tag": 4046, "value": ", and " }, { "__type": "InlineCode", "__tag": 4051, "value": "whosmat" }, { "__type": "Text", "__tag": 4046, "value": " are imported. Within " }, { "__type": "InlineCode", "__tag": 4051, "value": "sio.matlab" }, { "__type": "Text", "__tag": 4046, "value": ", you will find the " }, { "__type": "InlineCode", "__tag": 4051, "value": "mio" }, { "__type": "Text", "__tag": 4046, "value": " module This module contains the machinery that " }, { "__type": "InlineCode", "__tag": 4051, "value": "loadmat" }, { "__type": "Text", "__tag": 4046, "value": " and " }, { "__type": "InlineCode", "__tag": 4051, "value": "savemat" }, { "__type": "Text", "__tag": 4046, "value": " use. From time to time you may find yourself re-using this machinery." } ] } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "The basic functions" } ], "level": 2, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "You may have a " }, { "__type": "InlineCode", "__tag": 4051, "value": ".mat" }, { "__type": "Text", "__tag": 4046, "value": " file that you want to read into SciPy. Or, you want to pass some variables from SciPy / NumPy into MATLAB." } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "To save us using a MATLAB license, let's start in " }, { "__type": "InlineRole", "__tag": 4003, "value": "Octave_", "domain": null, "role": null, "inventory": null }, { "__type": "Text", "__tag": 4046, "value": ". Octave has MATLAB-compatible save and load functions. Start Octave (" }, { "__type": "InlineCode", "__tag": 4051, "value": "octave" }, { "__type": "Text", "__tag": 4046, "value": " at the command line for me):" } ] }, { "__type": "Code", "__tag": 4050, "value": "octave:1> a = 1:12\na =\n\n 1 2 3 4 5 6 7 8 9 10 11 12\n\noctave:2> a = reshape(a, [1 3 4])\na =\n\nans(:,:,1) =\n\n 1 2 3\n\nans(:,:,2) =\n\n 4 5 6\n\nans(:,:,3) =\n\n 7 8 9\n\nans(:,:,4) =\n\n 10 11 12\n\noctave:3> save -6 octave_a.mat a % MATLAB 6 compatible\noctave:4> ls octave_a.mat\noctave_a.mat", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Now, to Python:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> mat_contents = sio.loadmat('octave_a.mat')\n>>> mat_contents\n{'__header__': b'MATLAB 5.0 MAT-file, written\n by Octave 3.2.3, 2010-05-30 02:13:40 UTC',\n '__version__': '1.0',\n '__globals__': [],\n 'a': array([[[ 1., 4., 7., 10.],\n [ 2., 5., 8., 11.],\n [ 3., 6., 9., 12.]]])}\n>>> oct_a = mat_contents['a']\n>>> oct_a\narray([[[ 1., 4., 7., 10.],\n [ 2., 5., 8., 11.],\n [ 3., 6., 9., 12.]]])\n>>> oct_a.shape\n(1, 3, 4)", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Now let's try the other way round:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> import numpy as np\n>>> vect = np.arange(10)\n>>> vect.shape\n(10,)\n>>> sio.savemat('np_vector.mat', {'vect':vect})", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Then back to Octave:" } ] }, { "__type": "Code", "__tag": 4050, "value": "octave:8> load np_vector.mat\noctave:9> vect\nvect =\n\n 0 1 2 3 4 5 6 7 8 9\n\noctave:10> size(vect)\nans =\n\n 1 10", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "If you want to inspect the contents of a MATLAB file without reading the data into memory, use the " }, { "__type": "InlineCode", "__tag": 4051, "value": "whosmat" }, { "__type": "Text", "__tag": 4046, "value": " command:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> sio.whosmat('octave_a.mat')\n[('a', (1, 3, 4), 'double')]", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "InlineCode", "__tag": 4051, "value": "whosmat" }, { "__type": "Text", "__tag": 4046, "value": " returns a list of tuples, one for each array (or other object) in the file. Each tuple contains the name, shape and data type of the array." } ] } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "How do I start?" } ], "level": 2, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "MATLAB structs are a little bit like Python dicts, except the field names must be strings. Any MATLAB object can be a value of a field. As for all objects in MATLAB, structs are, in fact, arrays of structs, where a single struct is an array of shape (1, 1)." } ] }, { "__type": "Code", "__tag": 4050, "value": "octave:11> my_struct = struct('field1', 1, 'field2', 2)\nmy_struct =\n{\n field1 = 1\n field2 = 2\n}\n\noctave:12> save -6 octave_struct.mat my_struct", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "We can load this in Python:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> mat_contents = sio.loadmat('octave_struct.mat')\n>>> mat_contents\n {'__header__': b'MATLAB 5.0 MAT-file, written by Octave 3.2.3, 2010-05-30 02:00:26 UTC',\n '__version__': '1.0',\n '__globals__': [],\n 'my_struct': array([[(array([[1.]]), array([[2.]]))]], dtype=[('field1', 'O'), ('field2', 'O')])\n }", "execution_status": null }, { "__type": "Code", "__tag": 4050, "value": ">>> oct_struct = mat_contents['my_struct']\n>>> oct_struct.shape\n(1, 1)\n>>> val = oct_struct[0, 0]\n>>> val\nnp.void((array([[1.]]), array([[2.]])), dtype=[('field1', 'O'), ('field2', 'O')])\n>>> val['field1']\narray([[ 1.]])\n>>> val['field2']\narray([[ 2.]])\n>>> val.dtype\ndtype([('field1', 'O'), ('field2', 'O')])", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "In the SciPy versions from 0.12.0, MATLAB structs come back as NumPy structured arrays, with fields named for the struct fields. You can see the field names in the " }, { "__type": "InlineCode", "__tag": 4051, "value": "dtype" }, { "__type": "Text", "__tag": 4046, "value": " output above. Note also:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> val = oct_struct[0, 0]", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "and:" } ] }, { "__type": "Code", "__tag": 4050, "value": "octave:13> size(my_struct)\nans =\n\n 1 1", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "So, in MATLAB, the struct array must be at least 2-D, and we replicate that when we read into SciPy. If you want all length 1 dimensions squeezed out, try this:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> mat_contents = sio.loadmat('octave_struct.mat', squeeze_me=True)\n>>> oct_struct = mat_contents['my_struct']\n>>> oct_struct.shape\n()", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Sometimes, it's more convenient to load the MATLAB structs as Python objects rather than NumPy structured arrays - it can make the access syntax in Python a bit more similar to that in MATLAB. In order to do this, use the " }, { "__type": "InlineCode", "__tag": 4051, "value": "struct_as_record=False" }, { "__type": "Text", "__tag": 4046, "value": " parameter setting to " }, { "__type": "InlineCode", "__tag": 4051, "value": "loadmat" }, { "__type": "Text", "__tag": 4046, "value": "." } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> mat_contents = sio.loadmat('octave_struct.mat', struct_as_record=False)\n>>> oct_struct = mat_contents['my_struct']\n>>> oct_struct[0,0].field1\narray([[ 1.]])", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "InlineCode", "__tag": 4051, "value": "struct_as_record=False" }, { "__type": "Text", "__tag": 4046, "value": " works nicely with " }, { "__type": "InlineCode", "__tag": 4051, "value": "squeeze_me" }, { "__type": "Text", "__tag": 4046, "value": ":" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> mat_contents = sio.loadmat('octave_struct.mat', struct_as_record=False, squeeze_me=True)\n>>> oct_struct = mat_contents['my_struct']\n>>> oct_struct.shape # but no - it's a scalar\nTraceback (most recent call last):\n File \"\", line 1, in \nAttributeError: 'mat_struct' object has no attribute 'shape'\n>>> type(oct_struct)\n\n>>> oct_struct.field1\n1.0", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Saving struct arrays can be done in various ways. One simple method is to use dicts:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> a_dict = {'field1': 0.5, 'field2': 'a string'}\n>>> sio.savemat('saved_struct.mat', {'a_dict': a_dict})", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "loaded as:" } ] }, { "__type": "Code", "__tag": 4050, "value": "octave:21> load saved_struct\noctave:22> a_dict\na_dict =\n\n scalar structure containing the fields:\n\n field2 = a string\n field1 = 0.50000", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "You can also save structs back again to MATLAB (or Octave in our case) like this:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> dt = [('f1', 'f8'), ('f2', 'S10')]\n>>> arr = np.zeros((2,), dtype=dt)\n>>> arr\narray([(0.0, ''), (0.0, '')],\n dtype=[('f1', '>> arr[0]['f1'] = 0.5\n>>> arr[0]['f2'] = 'python'\n>>> arr[1]['f1'] = 99\n>>> arr[1]['f2'] = 'not perl'\n>>> sio.savemat('np_struct_arr.mat', {'arr': arr})", "execution_status": null } ] } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "MATLAB structs" } ], "level": 2, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Cell arrays in MATLAB are rather like Python lists, in the sense that the elements in the arrays can contain any type of MATLAB object. In fact, they are most similar to NumPy object arrays, and that is how we load them into NumPy." } ] }, { "__type": "Code", "__tag": 4050, "value": "octave:14> my_cells = {1, [2, 3]}\nmy_cells =\n{\n [1,1] = 1\n [1,2] =\n\n 2 3\n\n}\n\noctave:15> save -6 octave_cells.mat my_cells", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Back to Python:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> mat_contents = sio.loadmat('octave_cells.mat')\n>>> oct_cells = mat_contents['my_cells']\n>>> print(oct_cells.dtype)\nobject\n>>> val = oct_cells[0,0]\n>>> val\narray([[ 1.]])\n>>> print(val.dtype)\nfloat64", "execution_status": null } ] }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Saving to a MATLAB cell array just involves making a NumPy object array:" } ] }, { "__type": "Blockquote", "__tag": 4059, "children": [ { "__type": "Code", "__tag": 4050, "value": ">>> obj_arr = np.zeros((2,), dtype=object)\n>>> obj_arr[0] = 1\n>>> obj_arr[1] = 'a string'\n>>> obj_arr\narray([1, 'a string'], dtype=object)\n>>> sio.savemat('np_cells.mat', {'obj_arr': obj_arr})", "execution_status": null } ] }, { "__type": "Code", "__tag": 4050, "value": "octave:16> load np_cells.mat\noctave:17> obj_arr\nobj_arr =\n{\n [1,1] = 1\n [2,1] = a string\n}", "execution_status": null } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "MATLAB cell arrays" } ], "level": 2, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Code", "__tag": 4050, "value": ".. autosummary:: \n readsav", "execution_status": null } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "IDL files" } ], "level": 1, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Code", "__tag": 4050, "value": ".. autosummary:: \n mminfo\n mmread\n mmwrite", "execution_status": null } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "Matrix Market files" } ], "level": 1, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Code", "__tag": 4050, "value": ".. autosummary:: \n read\n write", "execution_status": null } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "Wav sound files (" }, { "__type": "InlineRole", "__tag": 4003, "value": "scipy.io.wavfile", "domain": null, "role": "mod", "inventory": null }, { "__type": "Text", "__tag": 4046, "value": ")" } ], "level": 1, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Code", "__tag": 4050, "value": ".. autosummary:: \n loadarff", "execution_status": null } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "Arff files (" }, { "__type": "InlineRole", "__tag": 4003, "value": "scipy.io.arff", "domain": null, "role": "mod", "inventory": null }, { "__type": "Text", "__tag": 4046, "value": ")" } ], "level": 1, "target": null }, { "__type": "Section", "__tag": 4015, "children": [ { "__type": "Code", "__tag": 4050, "value": ".. autosummary:: \n netcdf_file", "execution_status": null }, { "__type": "Paragraph", "__tag": 4045, "children": [ { "__type": "Text", "__tag": 4046, "value": "Allows reading of NetCDF files (version of " }, { "__type": "Link", "__tag": 4049, "children": [ { "__type": "Text", "__tag": 4046, "value": "pupynere" } ], "url": "https://pypi.org/project/pupynere/", "title": "" }, { "__type": "Text", "__tag": 4046, "value": " package)" } ] }, { "__type": "Target", "__tag": 4061, "label": "pupynere" }, { "__type": "Target", "__tag": 4061, "label": "octave" }, { "__type": "Target", "__tag": 4061, "label": "matlab" } ], "title": [ { "__type": "Text", "__tag": 4046, "value": "Netcdf" } ], "level": 1, "target": null } ], "local_refs": [] }