Matricesm

[0.9a23](https://pypi.org/project/MatricesM/)

Important updates summarized:
-> Add eigenvector calculation
-> Add eigenvalue decomposition
-> Add singular value decomposition
-> Enable better handling for dataframes
-> Parameter and attribute name changes
-> Lots of bug fixes

New functions and classes
-> Add **Group** object

New methods, properties and attributes
_Matrix_ object:
-> Add **uniques** method
-> Add **groupBy** method
-> Add **transform** method
-> Add **oneHotEncode** method
-> Add **name_update** method
-> Add **index_update** method
-> Add **swap** method
-> Add **__name __** property
-> Add **eigenvectors** property
-> Add **isDefective** property
-> Add **diagmat** property
-> Add **eigenvecmat** property
-> Add **EIGENDEC** property
-> Add **SVD** property
-> Add **setdiag** method
-> Add **EIGENVEC_ITERS**

_null_ object:
-> Add magic methods to null object
-> Add **__name __**

New parameters to old methods and functions
_Matrix_ object's methods:
-> Add _inf_limit_ to **prod** and **sum** to set upper bound for values
-> Add _fillnull_ option to **add** and **concat** methods

_roundto_ function:
-> Add _printing_decimal_ option to **roundForm** for passing value to _decimal_ parameter of returned Matrix

Updated or removed parameter names and attributes:
-> **listed** parameter is now **data**
-> Change **EIGEN_ITERS** to **QR_ITERS**
-> Remove **COL_LIMIT** option
-> Remove _iterations_ parameter from **shuffle** method
-> Add Matrix support to **apply** and **transform** method's _conditions_ parameter

Rest of the updates
-> Name correction of function **override** to **overwrite**
-> Change **ind** into a class wrapped as a property
-> dataframes have **null** as default fill
-> Update default values of some options
-> Enabled concatenating to empty matrices
-> Update methods to work with null object
-> Update docstrings
-> Update sample amount to get better accuracy for dtype prediction when a file is read

Issues:

->solved: 67 , 78


[Comparison](https://github.com/MathStuff/MatricesM/compare/0.9a22...0.9a23)

[0.9a22](https://pypi.org/project/MatricesM/)

General updates:
New functions and classes
->Using 'typing' module for annotations, hints for parameters
->Using 'shutil' module for fitting the matrix to the window
->Added 'read_file' function to matrix.py to read files
->Added 'roundto' function to round lists and complex numbers easily
->Added new errors : NotSubList ,OutOfRangeList
->Added more value validating functions
->Added base for 'date' class
->Added base for 'join' method

New methods, properties and attributes
->Added 'min' and 'max' methods
->Added 'ranked' method to return ranks of the values in the columns
->Added 'index' and 'indexname' parameters to Matrix for row indices
->Added 'd0' and 'd1' properties to return row and column sizes
->Added 'NOTES' attribute to print out notes when matrix is printed
->Added 'rename' to rename columns
->Added 'namereset' to reset column names
->Added 'indexreset' to reset row indices, removed 'indexSet'
->Added 'info' property to return column info about invalid and unique values

New parameters to old methods
->1-dimensional row indexing is enabled by using 'ind' property
->Added 'key' option to 'sortBy' to pass an optional funciton to use while sorting
->Added option to return a Matrix in statistical methods via 'get' parameter
->Added option 'onlyrow' to 'find' method to return only the row indices
->Added 'returnmat' option to most of the methods to return the matrix after function call
->Added option to choose correlation coefficient in 'corr' method. 'pearson' and 'spearman' works but 'kendall' currently returns None

Printing related updates
->Enabled better transpose method for dataframes
->Better representation of dataframes
->Better representation for 'describe'

Rest of the updates
->Added rounding to 'is...' properties with 'PRECISION' attribute for better results
->Better docstrings
->Updated exampleMatrices.py, removed 'expected output' text
->Customized __ getattr__ and __ setattr__
->Updated __ init__ to be more organized
->Re-wrote __ setitem__ to work and look better


Updated or removed parameter names and attributes:

->'directory' and 'header' parameters are removed
->'concat' method's 'concat_as' is now 'axis'
->'uptri' and 'lowtri' are now 'U' and 'L'
->'indexSet', 'highest' and 'lowest' methods are removed
->'asDict' parameters on statistical methods are now 'get'


Issues:

->solved: 68 , 71 , 72 , 73 , 74

->invalid: 56


Bug and performance fixes:

->Eigenvalue calculation is fixed in most of the cases, EIGEN_ITERS may still need tweaks
->Fixed the case where feature names are not unique
->Fixed 'match' method returning repeating indices
->Fixed all known issues for __ repr__,__ getitem__, __ setitem__, __ delitem__
->Fixes on rounding methods
->Index error fixes on some methods
->Faster calculations via less 'matrix' attribute calls on all methods
->Overall better argument validation

[Comparison](https://github.com/MathStuff/MatricesM/compare/0.9a21...0.9a22)

[0.9a21](https://pypi.org/project/MatricesM/)

-Added 'match' method to use regex over dataframes

-Faster bitwise operations
-Fixes on __ setitem__, matfill
-Fixed 33

[Comparison](https://github.com/MathStuff/MatricesM/compare/0.9a20...0.9a21)

[0.9a20](https://pypi.org/project/MatricesM/)

-Importing the module is now can also be done by:
python
from MatricesM import *
OR
import MatricesM as mm

-dtypes are now passed as objects ( int,float,complex,dataframe ), not strings.
python
Matrix(100,dtype=int)

Matrix((2500,4),fill=gauss,ranged=(10,3),dtype=dataframe)


-Changed '_dir' to '__directory'
-Changed '_header' to '__header'
-Changed parameters to [*args and **kwargs](https://github.com/MathStuff/MatricesMbasic-syntax).

-Added more special distributions: betavariate,gammavariate,expovariate and lognormvariate
-Added 'PRECISION' attribute to be used in stdize,normalize and possibly more
-Added 'echelon' property to return the echelon form
-Added 'kwargs' property to return all attribute names and their values in a dictionary
-'cov' now returns the covariance matrix if no arguments passed
-'replace' method's 'column' parameter now also accepts lists
-'replace' method now has an option to return the matrix after evaluation
-Added 'asDict' to 'mode','median' and 'freq' methods

-Fixed eigenvalue calculation for most of the cases
-Added matrix multiplication written in Cython
-Updated some doc-strings
-Optimizations on some of the basic functions
-Fixes on creating randomly filled matrices, some of the setters and more
-Fixed/closed issues : 34, 36, 37, 46, 47, 48, 52, 64, 66,

[Comparison](https://github.com/MathStuff/MatricesM/compare/0.9a19...0.9a20)

[0.9a19](https://pypi.org/project/MatricesM/)

-Importing the module is now done by:
python
from MatricesM.matrix import *
OR
import MatricesM.matrix as mm

Distribution names can now be given as:
python
Matrix((1000,3), fill=gauss, ranged=(50,20)) Old : Matrix((1000,3), fill='gauss', ranged=(50,20))
Matrix((10000,2), fill=triangular, ranged=(0,100,65)) Old : Matrix((10000,2), fill='triangular', ranged=(0,100,65))

Distribution names can bu used now as strings
Matrix(10, fill="gauss", dtype="dataframe")

-Added 'count' method to count valid values for each column
-Added 'sum' and 'prod' methods to return sums and products of the elements in each column
-Added __delitem __ to delete rows or columns using 'del'


-Fixes on __repr __ and more
-Fixed/closed issues : 30 , 28 , 19 , 21 , 61

[Comparison](https://github.com/MathStuff/MatricesM/compare/0.9a18...0.9a19)

[0.9a18](https://pypi.org/project/MatricesM/)

-Better representation/printing when the matrix is called. Read [README.md](https://github.com/MathStuff/MatricesMprinting-options)
-Added much better handling for getting and setting specific parts of the matrix
-Added 'replace' method to replace values in the matrix. [Examples](https://github.com/MathStuff/MatricesMreplace-values-in-the-matrix)
-'col' method and all statistics methods accept column names now
-Enabled strings,lists,ranges in 'fill' parameter
-Added properties for LU and QR to get both matrices

-Fixed 54 , 62

-Added 'implicit' to reduce reloading and copying time
-Added rref calculation in Cython

[Comparison](https://github.com/MathStuff/MatricesM/compare/0.9a17...0.9a18)