#atAllPut: (Matrix instance protocol)
#atColumn:put:repeat:
#initializeRows:columns:
#rows:columns: (Matrix class protocol)
#rows:columns:element: (Matrix class protocol)

Added convenience PMMatrix methods with tests

…ization cleaning on matrices + abstract class PMDataTransformer. 5 unit tests include.

…g. Results different from what you obtain from Python ...

…on classes (Jacobi transformation of the covariance matrix and SVD) and one abstract classes for all PCA.

Still 2 unit tests not green.

Close #79

Close #80

…d more tests for argmax.

…nh on a matrix

… completely the API and old version got some problems with encapsulation ...

…#127)

* [issue-29] Removed test for a deprecated message.

* [issue-29] Replaced a call to a deprecated message, corrected a style violation.

* [issue-29] Removed deprecated messages.

…ested quickfix (#121)

* Added Basic Scalar subtraction test

* added test for subtract PMVector from a Number

* QuickFix of the scalar subtraction with PMVector Issue

This happened when one of the input features was constant.
- Set scale to 1 for the feature which is constant
- Added two tests for this case

New methods `step` and `postStep` have been added.
`postStep` will contain all viz and debug methods.
A new `stateVariables` contains required variables for
the steps.

…ests (#132)

* [issue-105] Removed an obselete test.

* [issue-105] Improved the name of the class. The class comment says it is a generator.

* [issue-105] Improved the name of the class.

* [issue-105] Extracted duplicate code to an intention-revealing method.

* [issue-105] Extracted duplicate code to a helper method.

* [issue-105] Inlined the method as it's only used in one place.

* [issue-105] Improved the name of the method.

* [issue-105] Corrected code style.

* [issue-105] Improved the name of the method.

* [issue-105] Removed obselete code.

* [issue-105] Corrected the category.

* [issue-105] Improved the names of the methods.

* [issue-105] Improved the names of the method.

* [issue-105] Removed obselete test.

* [issue-105] Brought back a test deleted accidentally.

* [issue-105] Made similar code even more similar.

* [issue-105] Moved the code to a better place, an intention-revealing method.

* [issue-105] Made the code consistent with that of the other tests. This reads a little better now.

* [issue-105] Made similar code look the same, improved the name of a test method by using domain language.

- Renamed variable `computeErrorEvery` to `updateErrorEvery`
- Added `updateError` variable which calculates error only when true
- Added grid vizualization example
- Added linked-rings vizualization and generator

* Added Basic Scalar subtraction test

* added test for subtract PMVector from a Number

* QuickFix of the scalar subtraction with PMVector Issue

* Added methods for intuitive conversion between vectors and matrices

* Tests for matrix <-> vector interconversions

* Fixed implementation of reshape

* [issue-105] Removed unnecessary dot.

* [issue-105] The code is doing the same thing differently, so I have made it look exactly the same in order that we can extract the duplication elsewhere later.

* [issue-105] Extracted code to an intention-revealing method.

* [issue-105] Improved the name of the test method and extracted generation of a sample to an intention revealing variable.

* [issue-105] Extracted the generation of a sample to an intention-revealing method.

* [issue-105] In the case of the Poisson generator I understand that lambda is the mean/expected value (cf. wikipedia).

* [issue-105] Introduced the expectedValue message for consistency, in view of removing the mean message completely.

* [issue-105] Corrected the message name to conform to the Pharo style.

* [issue-105] Extracted duplicate code to a test helper class.

* [issue-105] Removed obselete comment.

* [issue-105] Made the assertion the same as used in the exponential generator test in hopes of moving it to a super class.

* [issue-105] Corrected and improved the name of the class and the method therein.

* [issue-105] Improved code formatting, replaced implicitly duplicate code with the random sample generator class.

* [issue-105] Improved the names of the temporary variables.

* [issue-105] Improved the name of the assertion.

* [issue-105] Removed an obselete comment.

* [issue-105] Improved the name of the assertion.

* [issue-105] Made the name of the test clearer.

* [issue-105] Inlined the temp variable as it does not enhance understandability.

- Also added relevant tests for the operators

* [issue-105] Moved the incrementing to the bottom so we can extract the tempering to a method in the next refactor.

* [issue-105] Extracted the tempering section to an intention-revealing method.

* [issue-105] Corrected the categorisation to match the super class's.

* [issue-105] Corrected the categorisation to match the super class's.

* [issue-105] Corrected the initialisation.

* [issue-106] Fake it til we make it - just test the random package for now.

* [issue-106] Corrected structure.

* [issue-106] Added another package to track.

* [issue-106] Added another two packages.

* [issue-106] Added the benchmark packages.

* [issue-106] Added the Chromosome, Clustering, Complex and Core packages.

* [issue-106] Added the Distribution packages.

* [issue-106] Added the FFT and FunctionFit packages.

* [issue-106] The FFT and FunctionFit packages seem broken so we replace
it with the K-packages.

* [issue-106] Added the Math-Matrix (and any Math-M*) package.

* [issue-106] Added all Math-N* packages.

* [issue-106] Added the Math-O* packages.

* [issue-106] The Math-ODE packages seems to cause the CI job to run out of
memory, so let's proceed to the Math-P* packages.

* [issue-106] Added the Math-Q* packages.

* [issue-106] Added the coverage status to the README.

* [issue-106] Replaced hard-coded package name with something that
keeps overage open to additional R-packages.

* [issue-106] Added the TNSE package.

* [issue-106] The link to the badge should point to the development branch.

* [issue-106] We can at least cover the FunctionFit package.

* [issue-106] Improved message names.

* [issue-106] The happy path should not be the guard clause; normally the subclass should respond to the message.

* [issue-106] Improved the name of the block variable.

* [issue-106] Moved related code closer together. initRest goes on to populate parameters, arguments and results.

* [issue-106] The asString message send is redundant.

* [issue-106] Improved the name of the temp variable.

* [issue-106] Improved the names of the block variables.

* [issue-106] The asString is redundant as the tests remain green.

* [issue-106] Extracted code to an intention-revealing method.

* [issue-106] Simplified the calculation of the occurence of a key.

* [issue-106] Inlined the temporary variable as it is now redundant.

* [issue-106] Extracted an code to an intention-revealing local temp and improved ormatting.

* [issue-106] Inlined the variable as it over wrote the earlier assignment.

* [issue-106] The initialize call looks like a special case, so transformed it into a guard clause.

* [issue-106] Improved code formatting.

* [issue-106] The check is redundant now because we have got passed the guard clause.

* [issue-106] The nil check is unnecessary now given the ifNone: call. The call to join: is for efficiency (O(n) vs O(n^2) for concatenation). This was recommended by Pharo.

* [issue-106] Reverted the change as we are only concatenating two strings which is not super inefficient.

* [issue-106] Removed obselete code.

* [issue-106] Cover all packages starting with Math-O.

* [issue-106] Back to hard-coding the package name (Travis ran out of memory).

* Clean tests of AutomaticDifferenciation

* Use double dispatch for +

* Double dispatch for * in DualNumber

* Use double dispatch for *,+ and /.
Now all tests are green again

* Add a test about derivatives

* PMDualNumber is now a subclass of Number and not Magnitude.

* Add conjugated and real method (to support dual numbers with complex numbers).
Add support in PMComplex classes also.
Add green tests.

* Use absSquared instead of abs2

* Remove constant comments

* Fix formatting

* Reformat correctly PMHyperDualNumber