DAY 30 OF 100]FRIDAY ,30 JAN 2015

1Python

decorators unpacking args ,passing objs 

DAY 28 OF 100] WEDNESDAY ,28 JAN 2015

1Python

counter , benchmark decorator functions/examples

DAY 27 OF 100] TUESDAY ,27 JAN 2015

1Python

functools.wraps debug help for decorated function

DAY 26 OF 100] MONDAY ,26 JAN 2015

1Linear Algebra 

figured out the algorithm from first principles.It is imp to expand the equation A = LDLt into it's partitioned matrix form and then equate the RHS LHS terms in each partition.Now it only needs a rank1 update and a vector scaling.

implemented the LDLt algo using laff routines.looks like it takes less number of flops since sqrt(alpha11) call is not needed ..

tested for 3x3 and 4x4 matrices ..

will do the triangular solves (which may require 3 passes .. for L,D and Lt ) later..

implemented routines for
Lz=b
Dy=z
Ltx=y

tested for 3x3 and 4x4 matrices
Ltx=y solve does not transpose L explicitly

DAY 25 OF 100] SUNDAY ,25 JAN 2015

1Linear Algebra

did LDLt factorization by hand for 3x3 and 4x4 matrices .Was confused by L*D  mult..

will try to figure out algo as per Geijn's notations

DAY 24 OF 100] SATURDAY ,24 JAN 2015

1.Python

decorator maker with kwargs

DAY 23 OF 100] FRIDAY ,23 JAN 2015

1.Linear Algebra

cholesky fact with SPD -

     modified code to avoid transposing the column a21 to calculate rank1 update

Back sunbstitution implemented. No transposing of the LTM needed to do LLt = A

The algo now uses l21(which would be aa12t in the transpose) instead of  transposing L to get Lt .

All 3 routines are now tested

cholesky = takes in A which is an  SPD and converts it into an LTM with some old elements(of A) in upper triang part

solve_lzb = takes in the LTM (actually updated A )produced by cholesky , and a vector b.Updates b with the elements of z

solve_uxb = takes in the  LTM produced by cholesky, and the updated b (which now is actually z)

Originally,the L has to be transposed to get Lt and then solve Lt*x = b. Here we work with the required elements in LTM and ignore the old elements from A.The vector b is updated to contain elements of x.

This exercise helped to think about traversal thru matrix.. 

satisfying day :)

DAY 22 OF 100] THURSDAY ,22 JAN 2015

1.Linear Algebra

cholesky fact with SPD - implementation done

Ltm_lzb solve implementation done 

need to test on more datasets

will do the impl of uxb solve tomorrow

DAY 21OF 100] WEDNESDAY ,21 JAN 2015

1.Linear Algebra

cholesky fact with SPD example

successfully did back subst for Lz = b and Ltx = z 

need to work out impl..

DAY 20 OF 100] TUESDAY ,20 JAN 2015

1.Linear Algebra

cholesky fact with SPD example

further fwd/back subst failed!

find out why

DAY 19 OF 100] MONDAY ,19 JAN 2015

Python

SO - deco maker with arg passing

DAY 18 OF 100] SUNDAY ,18 JAN 2015

Python

SO article   -- decorator maker

DAY 17 OF 100] SATURDAY ,17 JAN 2015

Tkinter

ch10/simple dialog - modified version 

DAY 16 OF 100] FRIDAY ,16 JAN 2015

1.Linear Algebra

SPD - choelsky  factorization impl
need more work here.. 
tried LU fact/and solves 
similar on SPD failed when cholesky was attempted

2.Python

SO article    -- manual and by syntax deco with arbit/args

3.Tkinter

ch10/simple dialog - modified version  transient and grab_set

DAY 15 OF 100] THURSDAY ,15 JAN 2015

1.Linear Algebra

SPD - choelsky  factorization

2.Python

SO article    -passing args to method s

3.Tkinter

ch10/simple dialog - problems ,

DAY 14 OF 100] WEDNESDAY ,14 JAN 2015

1.Linear Algebra

8.3 (Almost) Never, Ever Invert a Matrix

8.3.1 Solving Ax = b

2.Python

SO article on decorators
upto accumulation and arg passing

Tkinter
ch10 / improving simple dialog

DAY 13 OF 100] TUESDAY ,13 JAN 2015

1.Linear Algebra

Challenge 8.2.5.5   -- final work / impl

8.3 (Almost) Never, Ever Invert a Matrix

8.3.1 Solving Ax = b

DAY 12 OF 100] MONDAY ,12 JAN 2015

1.Linear Algebra

Challenge 8.2.5.5   -- figured out the updates ,need more work ,will finish tomorrow

8.2.7 Cost of Matrix Inversion review

8.3 (Almost) Never, Ever Invert a Matrix

8.3.1 Solving Ax = b

DAY 11 OF 100] SUNDAY ,11 JAN 2015

1.Linear Algebra

Algorithm: [A;B] := GJ INVERSE ALT(A;B)  review/write

challenge postponed for tomorrow

8.2.6 Pivoting

8.2.7 Cost of Matrix Inversion
need to review this tomorrow

DAY 10 OF 100] SATURDAY ,10 JAN 2015

1.Linear Algebra

Algorithm: [A;B] := GJ INVERSE ALT(A;B)
Homework 8.2.5.4

Challenge 8.2.5.5

8.2.6 Pivoting

DAY 9 OF 100] FRIDAY ,9 JAN 2015

1.Linear Algebra

review earlier 8.2.4 material / write

8.2.5 Computing inverse via Gauss-Jordan Elimination, Alternative

Homework 8.2.5.1
Homework 8.2.5.2
Homework 8.2.5.3

Algorithm: [A;B] := GJ INVERSE ALT(A;B)
Homework 8.2.5.4

2.Python
Tkinter
ch10 / improving simple dialog

DAY 8 OF 100] THURSDAY ,8 JAN 2015

1.Linear Algebra

Homework 8.2.4.3

Homework 8.2.4.4

Homework 8.2.4.5

Algorithm: [A;B] := GJ INVERSE PART1(A;B)

Algorithm: [A;B] := GJ INVERSE PART2(A;B)

walk thru these algos with a sample exercise.. compare outcomes in each step

2.Python

Tkinter 

ch9:Data entry 

ColorChooser

Chapter 10. Dialog Windows
simple dialog window as a class ,problems of this design

Grid layout

  DAY 7 OF 100]WEDNESDAY ,7 JAN 2015

1.Linear Algebra
8.2.4 inverse of A via Gauss-Jordan Elimination

DAY 6 OF 100]TUESDAY ,6 JAN 2015

1.Linear Algebra

8.2.3 Solving Ax = b via Gauss-Jordan Elimination: Multiple Right-Hand Sides

tested algo  by hand on sample set 

2.Python

Tkinter

data entry simple dialog

filedialog

DAY 5 OF 100] MONDAY ,5 JAN 2015

1.Linear Algebra

apply Gauss-jordan algo as given part1,2 to a sys of linear eqns

8.2.3 Solving Ax = b via Gauss-Jordan Elimination: Multiple Right-Hand Sides

Homework 8.2.3.1

Homework 8.2.3.3
Homework 8.2.3.4
Homework 8.2.3.5

2.Python

Tkinter
messagebox

DAY 4 OF 100] SUNDAY ,4 JAN 2015

1.Linear Algebra

Homework 8.2.2.3

Homework 8.2.2.4

Gauss-jordan algo outline

2.Python

Tkinter

status bars 
custom StatusBar class with set and clear methods
application with set clear buttons and statusbar

DAY 3 OF 100] SATURDAY ,3 JAN 2015

1.Linear Algebra

Unit 8 :

try gauss jordan on a lufact/gelim  failing case

2.Python

SO article on decorators

Tkinter 

popup menu

tear-off menu item explored

toolbars

4 pomodoros

DAY 2 OF 100] FRIDAY ,2 JAN 2015

1.Linear Algebra

Unit 8 :

review and write upto 8.2.1

 8.2.2 Solving Ax = b via Gauss-Jordan Elimination: Gauss Transforms

Homework 8.2.2.1

2.Python

SO article on python decorators.type what you learned

function as reference
manual decoration of function .. returns a  wrapper  function object
decorator syntax using @ equiv to the above

Tkinter

ch8 Application Windows

basic menu widget explored.

popup menu ? explore more tomorrow

6 pomodoros