Skip to main content

Subclass of numpy.matrix behaving as matrices in matlab.

Project description


This is my module about matrix operation. It imitates matlab grammar. If you love matlab as well as python, then this is your choice. It will be a good experience to operating matrices matlab-like.

Sugar: one can use single index to refer to the elements in a matrix. (see following examples)



  • mymat

  • matshow

  • matdemo

  • test_mat

  • linalg

  • denoise


Current Version:
  1. fix some bugs.

  2. Demonstrate Gauss elimination with tkinter.

  3. Define LinearEquation class

>>> import mymat.matdemo
>>> mymat.matdemo.main()
Main Feature(>0.1.x):
  1. MyMat, PyMat is now the subclass of MatBase

  2. improve some essential methods, fix some bugs

  3. the index can be in reserved order, such as A[3:1:-1,1]

  4. see numerical experiment in mat_demo (improved)

  5. add more methods (introduced below) and A[is,js]=[] now is legal

  6. fix some bugs, make the codes more robust

finally, another improvement is that when create a matrix, we use following codes to set dtype (may temporarily)

if data.dtype != np.complex128 and data.dtype != object:

kwargs.setdefault(‘dtype’, np.float64)

Main Feature(0.0.x):
  1. introduce operator | and & to concatenate matrices

  2. in setitem, the index is allowed to be out of range as matlab with the help of update method (see below)

  3. correct the codes of delete, improve the codes of many method

  4. add poly/expm (Tylor approximation) method to calculate p(A) and e^A

  5. add totex method, transforming a matrix to its tex-form

  6. the default dtype of MyMat is float64(complex128 when it is complex), but the integer matrix is int32. so, don’t forget to convert the dtype if neccessary. But this is temporary.


basic grammar


>>> import mymat
>>> A = mymat.MyMat([]) # use import mymat.pymat to import PyMat

operators (Python left, Matlab right):

A*B := A*B  (B*A := B*A)
A/B := A/B == A*B.I (B/A := B/A == B*A.I)
A ** B := A .* B  (B ** A := B .* A)
A//B := A./B  (B//A := B./A)
A<<B := A.^B  (B<<A := B.^A)
A^B := A^B
A|B := [A,B]   A&B :=[A;B]

We use matlab-type index, instead of python-type index, for example:

>>> A=TestMat(5)
[1, 2, 3, 4, 5;
 6, 7, 8, 9, 10;
 11, 12, 13, 14, 15;
 16, 17, 18, 19, 20;
 21, 22, 23, 24, 25]: M(5 X 5)
>>> A[[3,4,7,10]]    # with single index as in matlab
[11, 16, 7, 22]: M(1 X 4)
>>> A[[2,3],1:4]
[6, 7, 8, 9;
11, 12, 13, 14]: M(2 X 4)
>>> A[[1,3],[2,4]]   # use A.get(([1,3],[2,4])) to get matrix([2, 14])
[2, 4;
12, 14]: M(2 X 2)

>>> A[3:1:-1,:]     # reversing order
[11, 12, 13, 14, 15;
6, 7, 8, 9, 10;
1, 2, 3, 4, 5]: M(3 X 5)

Use delete method to delete some rows or columns, as in matlab:

>>> A=H(7)
>>> B=A.delete([1,3],slice(3))   #  <=> B=A.copy(); B[[1,3],[1,2,3]]=[]
>>> B.shape
(5, 4)

Linear equation:

>>> le = LinearEquation(A, b)
>>> print(le.totex())    # print tex of a linear equation

Demonstration and Visualization

demonstration and numerical experiment:

>>> import mymat
>>> import mymat.matdemo        # see Gauss elimination
>>> A=mymat.MyMat('1,1,1,6;0,4,-1,5;2,-2,1,1')  # or    A=mymat.MyMat('1&1&1&6\\0&4&-1&5\\2&-2&1&1') just copying the latex codes
>>> mymat.matdemo.guassDemo(A)  # show the process of getting the echelon form of A
>>> mymat.matdemo.denoiseDemo([n:noised signal(row vector)]) # see a denoising experiment

draw a matrix:

>>> import mymat.matshow    # draw a matrix on axes(require matplotlib)
>>> ms = mymat.matshow.MatrixShow(A);

Methods and Functions

other methods:

__call__: A(ind) == A[ind]
delete(ind1=row, ind2=col): delete row-rows and col-columns
proj(ind1=row, ind2=col): =0 out of A[row, col], for example A.proj(ind1=COLON, ind2=[2,3])=[2,3] where A=[1,2,3,4]
repmat((ind1, ind2)|ind): repeat matrix as in Matlab (like tile)
just: cut matrix to a certain size, and supplement zeros if the size is too large.
cat: as concatenate
equal: (A == B == C).all()
apply: A.apply(lambda x:x+1) == A+1
plus: == A + nI
robinson: A.robinson(j, x) == A[j<-x], namely A[:,j]=x used in Cramer rule
echelon: get the echelon form (include the corresponding column indexes)
tril, triu, diag are similar to matlab
row(col)_transform1/2/3: elementary row (column) transforms (Gauss tranforms)
comat: get the co-matrix (similar with delete)
cofactor: A.cofactor(i,j)=Aij get the cofactor based on comat
rho: the spectral radius
totex: to tex form of matrix
tolineq: to tex form of linear equations wrt augment matrix
argmin, argmax

class methods:

MyMat.zeros, MyMat.ones, MyMat.random, MyMat.randint, MyMat.eye

functions and variables:

ind2ind: the most essential function
times: translate single index to double index
compind: get complementary index (called in proj)


FM: Fourier matrix
FIM: Fourier inverse matrix
FUM: Fourier unitary matrix
Ho: Horsehold matrix
Ref: reflection matrix
H: Hilbert matrix
Elm1,Elm2,Elm3: elementary matrices (3 types)

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mymat-1.1.0.tar.gz (19.9 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page