Re: More Regexp help

[Freeman Gilmore]

Arron, I am replying to this knowing that some code is wrong.

On Sun, Jun 28, 2020 at 1:05 AM Aaron Hill <lilypond@hillvisions.com> wrote:
>
> On 2020-06-27 6:45 pm, Freeman Gilmore wrote:
> > I have a string of accidentals and I want the sum of their deviation.
> >
> > If I have the string, “a bx3 cd -jk -DRx2 rtyu -HJK 31 17x4 -7 -41x3”
> > with
> > all possible examples.
> >
> > I want to convert it to,  “ a (* b 3) -jk (* -DR 2) rtyu -HJK n31  (*
> > n17
> > 4) -n7 (* -n41 3)”
> >
> > Then,  (a (* b 3) -jk (* -DR 2) rtyu -HJK n31  (* n17 4) -n7 (* -n41
> > 3)) .
> >    ( + will be applied to this)
> >
> > 1] Where  “a” “cd” “-jk” “rtyu” “-HJK” are unchanged.    (variables)
> >
> > 2] Numbers “31” “17x4” “-7” “-41x3” have “n” prefixed to them like,
> > “n31”
> > “n17x4” “-n7” “-n41x3”. (numbers converted to variables)
> >
> > 3] And “bx3” “-DRx2” “17x4” “-41x3” become “(* b 3)” “(* -DR 2)” “(*
> > n17
> > 4)” “(* -n41 3)”.  (number of times variables are used)
> >
> > This is way beyond my capability, and I would appreciate help with
> > this.
> > For my simple mind it would teach me more if this were done in say 3
> > steps
> > rather than mapped.
>
> -----------------------------------------------------
> Broken into individual steps, as you have requested:
>
> ;;;;
> (use-modules (ice-9 regex))
>
> (define string "a bx3 cd -jk -DRx2 rtyu -HJK 31 17x4 -7 -41x3")
> (format #t "\n[1] ~s" string)
>
> (set! string
I did not know to use 'set!' this way,    I keep redefining the string
as I went.
>
>    (regexp-substitute/global #f
>      "(^|\\s)(-?)([0-9]+)"
>      string
>      'pre 1 2 "n" 3 'post))
> (format #t "\n[2] ~s" string)
I have been using 'write', but the numbers sure would have helped.
>
> (set! string
>    (regexp-substitute/global #f
>      "(^|\\s)(\\S+)(x([0-9]+))"
>      string
>      'pre 1 "(* " 2 " " 4 ")" 'post))
> (set! string (string-append "(" string ")"))
> (format #t "\n[3] ~s" string)
>
> (define expr (with-input-from-string string read))
> (format #t "\n[4] ~s" expr)
> ;;;;
> ====
> [1] "a bx3 cd -jk -DRx2 rtyu -HJK 31 17x4 -7 -41x3"
> [2] "a bx3 cd -jk -DRx2 rtyu -HJK n31 n17x4 -n7 -n41x3"
> [3] "(a (* b 3) cd -jk (* -DR 2) rtyu -HJK n31 (* n17 4) -n7 (* -n41
> 3))"
> [4] (a (* b 3) cd -jk (* -DR 2) rtyu -HJK n31 (* n17 4) -n7 (* -n41 3))
> ====
>
>
> ---------------------------------------------------------------
> Processed in a single logical operation, as I would recommend:
I agree and will use it.   You did a mapping for the list of strings
but I was  not able to use it.  I was able to add the part I needed to
the individual step.   It is miner but I do not know how to add it to
the map.   What I have is a mess but it did allow me to make changes
as I went.  Would have helped if I had known about 'set!' and the
[numbering] of the steps.

I now have the list acdental, glyph strings and the sum of their pitch
deviationI
>
> ;;;;
> (use-modules (ice-9 regex))
>
> (define string "a bx3 cd -jk -DRx2 rtyu -HJK 31 17x4 -7 -41x3")
> (format #t "\n[1] ~s" string)
>
> (define expr
>    (map (lambda (m)
>           (let* ((prefix (match:substring m 1))
>                  (digits (match:substring m 3))
>                  (letters (match:substring m 4))
>                  (count (match:substring m 6))
>                  (name (or letters (string-append "n" digits)))
>                  (symbol (string->symbol (string-append prefix name))))
>             (if count
>               (list '* symbol (string->number count))
>               symbol)))
>         (list-matches
>           "(-?)(([0-9]+)|([A-Za-z]+))(x([0-9]+))?"
>           string)))
> (format #t "\n[2] ~s" expr)
> ;;;;
> ====
> [1] "a bx3 cd -jk -DRx2 rtyu -HJK 31 17x4 -7 -41x3"
> [2] (a (* b 3) cd -jk (* -DR 2) rtyu -HJK n31 (* n17 4) -n7 (* -n41 3))
> ====
>
>
> NOTE: In both cases above, I stopped short of evaluating the expression,
> as there would be a whole host of unbound variables.
I know about this and all would be defined, about 15 variables for my use.
>  Assuming a valid
> environment, the summation could be computed as easily as:
> (primitive-eval (cons '+ expr))
Thanks for adding this, what I have tried would not work.
>
>
> -- Aaron Hill
>
Thank you to all for helping with this. ƒg