Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Replace transpose by transpose-matrix in the examples. #8

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Replace transpose by transpose-matrix in the examples. #8

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
85 changes: 42 additions & 43 deletions README.org
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -259,8 +259,7 @@ LICENSE: BSD sans advertising (Mark H)
1. instantiating matrices and vectors
2. inversion using BLAS/LAPACK

<2012-10-25 Thu> except that transpose and the last example (geqrf)
are broken.
<2012-10-25 Thu> except that the last example (geqrf) is broken.

In Common lisp, it is useful, sometimes overkill, to properly
define the precision and form of the numbers (i.e. single, double, rational,
Expand Down Expand Up @@ -472,7 +471,7 @@ The following sets up data for linear least squares estimation:
(6d0)
(7d0)
(8d0))))
(transpose *y2*) ;; ERROR: not in lisp-matrix-user?
(transpose-matrix *y2*)



Expand Down Expand Up @@ -514,9 +513,9 @@ The following sets up data for linear least squares estimation:
;; Cholesky
(let ((myrand (rand 4 4)))
(princ myrand)
(princ (matrix-like-symmetric-p (m* (transpose myrand) myrand)))
(princ (m* (m* (transpose myrand) myrand)
(minv-cholesky (m* (transpose myrand) myrand))))))
(princ (matrix-like-symmetric-p (m* (transpose-matrix myrand) myrand)))
(princ (m* (m* (transpose-matrix myrand) myrand)
(minv-cholesky (m* (transpose-matrix myrand) myrand))))))


(progn
Expand Down Expand Up @@ -635,18 +634,18 @@ lm-demo.lisp : things that might work but should

(mref *m2b* 0 2) ;; => 3
*m2b*
(transpose *m2b*)
(transpose-matrix *m2b*)

;; simple subsetting is simple
(m= (row *m2b* 0)
(col (transpose *m2b*) 0)) ; => nil, orientation
(col (transpose-matrix *m2b*) 0)) ; => nil, orientation
(v= (row *m2b* 0)
(col (transpose *m2b*) 0)) ; => T, no orientation worries
(col (transpose-matrix *m2b*) 0)) ; => T, no orientation worries

(m= (col *m2b* 0)
(row (transpose *m2b*) 0))
(row (transpose-matrix *m2b*) 0))
(v= (col *m2b* 0)
(row (transpose *m2b*) 0))
(row (transpose-matrix *m2b*) 0))


(defvar *m3*
Expand All @@ -663,13 +662,13 @@ lm-demo.lisp : things that might work but should


(= (mref *m3* 0 1)
(mref (transpose *m3*) 1 0))
(mref (transpose-matrix *m3*) 1 0))

(= (mref *m3* 2 2)
(mref (transpose *m3*) 2 2))
(mref (transpose-matrix *m3*) 2 2))

*m3*
(transpose *m3*)
(transpose-matrix *m3*)

;;; Now we play with striding and slicing subsets. These work well
;;; for simple subsetting which can be done by counting/enumeration
Expand Down Expand Up @@ -697,8 +696,8 @@ lm-demo.lisp : things that might work but should
(m= (col *m4* 1)
(make-matrix 2 1 :initial-contents '((2d0) (12d0))))

(v= (row *m4* 0) (col (transpose *m4*) 0))
(v= (col *m4* 0) (row (transpose *m4*) 0))
(v= (row *m4* 0) (col (transpose-matrix *m4*) 0))
(v= (col *m4* 0) (row (transpose-matrix *m4*) 0))

*m4*
(row *m4* 0)
Expand Down Expand Up @@ -953,48 +952,48 @@ m01b
(col m01 3)

m01
(transpose m01)
(row (transpose m01) 0)
(row (transpose m01) 1) ; wrong: grab bad column, AND by 1 (pushed up)
(row (transpose m01) 2) ; ditto, wrong by 2
(row (transpose m01) 3) ; etc...wrong by 3
(transpose-matrix m01)
(row (transpose-matrix m01) 0)
(row (transpose-matrix m01) 1) ; wrong: grab bad column, AND by 1 (pushed up)
(row (transpose-matrix m01) 2) ; ditto, wrong by 2
(row (transpose-matrix m01) 3) ; etc...wrong by 3

(row (transpose m01) 0)
(transpose (row (transpose m01) 0))
(row (transpose-matrix m01) 0)
(transpose-matrix (row (transpose-matrix m01) 0))

m01
(transpose m01)
(col (transpose m01) 0)
(col (transpose m01) 1) ; last rather than first
(col (transpose m01) 2) ;
(col (transpose m01) 3) ; ditto above
(transpose-matrix m01)
(col (transpose-matrix m01) 0)
(col (transpose-matrix m01) 1) ; last rather than first
(col (transpose-matrix m01) 2) ;
(col (transpose-matrix m01) 3) ; ditto above


(v= (row m01 0)
(col (transpose m01) 0)) ;; works
(col (transpose-matrix m01) 0)) ;; works

(m= (row m01 0)
(col (transpose m01) 0)) ;; fails, since dims unequal
(col (transpose-matrix m01) 0)) ;; fails, since dims unequal

m01
(transpose m01)
(transpose-matrix m01)
;; given the above...
;; FIXME: Big Barf!
(v= (row m01 1)
(col (transpose m01) 1) ) ;; fails badly. Real badly.
(col (transpose-matrix m01) 1) ) ;; fails badly. Real badly.

(v= (col m01 1)
(row (transpose m01) 1) ) ;; fails, but closer...
(row (transpose-matrix m01) 1) ) ;; fails, but closer...

(col m01 1)
(col (transpose m01) 1) ;; this is the problem, indexing issue...
(col (transpose-matrix m01) 1) ;; this is the problem, indexing issue...


;; and the same problem.
m3
(transpose m3)
(v= (col m3 1) (row (transpose m3) 1))
(v= (row m3 1) (col (transpose m3) 1))
(transpose-matrix m3)
(v= (col m3 1) (row (transpose-matrix m3) 1))
(v= (row m3 1) (col (transpose-matrix m3) 1))

;; Striding and Slicing issues:
;; Strides provide matrix sections; slicing provides vector'd sections.
Expand All @@ -1014,22 +1013,22 @@ m01b
(slice m01 :offset 5 :stride 2 :nelts 3 :type :row)


(slice (transpose m01) :offset 5 :stride 2 :nelts 3 :type :row)
(slice (transpose-matrix m01) :offset 5 :stride 2 :nelts 3 :type :row)
(slice m01
:offset 5
:stride 2
:nelts 3
:type :row)
(slice (transpose m01) :offset 5 :stride 2 :nelts 3 :type :row)
(slice (transpose-matrix m01) :offset 5 :stride 2 :nelts 3 :type :row)

;; slicing isn't affected by transposition -- doesn't affect the
;; counting. Would have suggested that column-major or row-major.
;; Should this be the case? (need to migrate to unit-tests).

(v= (slice m01 :offset 5 :stride 2 :nelts 3 :type :row)
(slice (transpose m01) :offset 5 :stride 2 :nelts 3 :type :row))
(slice (transpose-matrix m01) :offset 5 :stride 2 :nelts 3 :type :row))
(v= (slice m01 :offset 5 :stride 2 :nelts 3 :type :row)
(slice (transpose m01) :offset 5 :stride 2 :nelts 3 :type :column))
(slice (transpose-matrix m01) :offset 5 :stride 2 :nelts 3 :type :column))
;; and note the above -- vector equality doesn't depend on orientation...

(slice m01 :offset 1 :stride 2 :nelts 3 :type :column)
Expand Down Expand Up @@ -1073,9 +1072,9 @@ m01b

;; FIXME: the following has no applicable method -- only for
;; doubles, not integers.
(m* m2 (transpose m2))
(m* m2 (transpose-matrix m2))
;; but we can multiple doubles, but...
(m* m01 (transpose m01))
(m* m01 (transpose-matrix m01))



Expand Down