How to add capturing group value to shed replacement

[WolfgangDrescher]

I am using this filter to hide all tokens of a voice:

shed -s 5 -e "s/^(.+)$/$1yy/D s/^.+$//L s/^.+$//I"

Demo: https://verovio.humdrum.org/?file=https://raw.githubusercontent.com/WolfgangDrescher/lassus-geistliche-psalmen/master/kern/02-quare-fremuerunt-gentes.krn&filter=shed%20-s%205%20-e%20%22s/^(.+)$/$1yy/D%20s/^.+$//L%20s/^.+$//I%22

Beside from this not being perfect because it eliminates the instrument informations, I could not find a way to add the token value to the replacement via capturing groups. $1yy does return only yy for all data tokens. $0yy does return -zshyy for all data tokens. So my shed filter if actually returning invalid humdrum. Is there a way to use capturing groups within the replacement in shed?

Also see this PR: rism-digital/verovio#3347

[craigsapp]

To not lose instrument information, you can use s/^[^I].+%//I for the last substitution:

https://verovio.humdrum.org/?file=https://raw.githubusercontent.com/WolfgangDrescher/lassus-geistliche-psalmen/master/kern/02-quare-fremuerunt-gentes.krn&filter=shed%20-s%205%20-e%20%22s/%5E(.+)$/$1yy/D%20s/%5E.+$//L%20s/%5E.+$//I%22

I could not find a way to add the token value to the replacement via capturing groups. $1yy does return only yy for all data tokens.

I don't think so. Demo:

https://verovio.humdrum.org?t=KiprZXJuCjRjCjRkCjRlCjRmCjRnCj0KKi0K&filter=shed%20-e%20%22s%2F(.*)%2F%241yy%2F%22

$0yy does return -zshyy for all data tokens. So my shed filter if actually returning invalid humdrum. Is there a way to use capturing groups within the replacement in shed?

I do not understand. $1 returns the contents of the first capture group, up to $9 for the last capture group. $0 works differently: it gives the entire matched string (between the first two slashes in the s/// operator. This includes all capture groups and the stuff outside of the capture groups.

For the first substitution s/^(.+)$/$1yy/D, the D is the default behavior, so it is optional.

s/^(.+)$/$1yy/D can also be written as s/.*/$0yy/. Data tokens can never be empty, so (.+) and (.*) will be equivalent. Without the capture group you can use $0 instead. With the the capture group s/(.*)/$1yy. Using anchors as ins/^(.*)$/$1yy is redundant since .* is greedy and will always start at the beginning of the string to the end of the string. Using ^ might have problems with interpretations and comments (since there is automatically assumed to be * or ! as the start of the token (that would need testing to double check).

---

Another way for preserving instrument names would be to prefix all interpretations with ZZZ and then remove ZZZ for Instrument-related interpretations:

https://verovio.humdrum.org?t=KiprZXJuCipJIlRlbm9yZQoqSSdUCipNNC80CjFjIDFlIDFnCj0KMWQKPQoqLQohISBIaWRlIGJvdGggY2hvcmRzIGFuZCBzaW5nbGUgbm90ZXMvcmVzdHM6CiEhIWZpbHRlcjogc2hlZCAtZSAicy8gL3l5IC9nOyBzLyguKikvJDF5eS8iCiEhIFByZWZpeCBhbGwgaW50ZXJwcmV0YXRpb25zIHdpdGggWlpaOgohISFmaWx0ZXI6IHNoZWQgLWUgInMvKC4qKS9aWlokMS9JIgohISBSZW1vdmUgWlpaIGZyb20gaW50ZXJwcmV0YXRpb25zIHRoYXQgc3RhcnQgd2l0aCBJOgohISFmaWx0ZXI6IHNoZWQgLWUgInMvWlpaSS9JL0kiCg==

You can then restore other interpretations that you find you want to preserve as well as instrument names.

[WolfgangDrescher]

I don't think so. Demo:

Yes in verovio it seems to run correctly. But when I run it in the CLI I get this:

cat 02-quare-fremuerunt-gentes.krn | shed -s 5 -e "s/^(.+)$/$1yy/D"

!!!COM: Lassus, Rudolphus
!!!CDT: 1563-1625
!!!OPR: Geistliche Psalmen mit dreyen Stimmen
!!!ONM: 2
!!!OTL@@LA: Quare fremuerunt gentes
!!!OTL-incipit@@DE: Warumb empören sich die Haiden
**kern	**text	**kern	**text	**kern	**text
*staff3	*staff3	*staff2	*staff2	*staff1	*staff1
*Ivox	*	*Ivox	*	*Ivox	*
*I"Bassus	*	*I"Tenor	*	*I"Cantus	*
*I'B	*	*I'T	*	*I'C	*
*clefC3	*	*clefC2	*	*clefG2	*
*k[]	*	*k[]	*	*k[]	*
*d:dor	*	*d:dor	*	*d:dor	*
*M2/1	*	*M2/1	*	*M2/1	*
*met(C|)	*	*met(C|)	*	*met(C|)	*
*MM180	*	*MM180	*	*MM180	*
=1	=1	=1	=1	=1	=1
*xywh-#128:294,61,2729,518	*	*xywh-#72:316,78,2764,509	*	*xywh-#12:288,51,2770,524	*
!	!	!LO:TX:a:i:t=c.f.	!	!	!
00r	.	1d	War-	yy	.
.	.	2d	-umb	yy	War-
.	.	2d	em-	.	.
=2	=2	=2	=2	=2	=2
.	.	2c	-pö-	yy	-umb
.	.	2d	-ren	yy	em-
.	.	2e	sich	yy	-pö-
.	.	2f	die	yy	-ren
=3	=3	=3	=3	=3	=3
0r	.	1e	Hai-	yy	sich
.	.	.	.	yy	die
.	.	2.d	den/	yy	Hai-
.	.	.	.	yy	den/
.	.	4e	.	.	.
!!...

[craigsapp]

To hide the clef: use *clefG2yy. If there is no clef in the data, the Humdrum-to-MEI converter adds one automatically based on the pitch range of the first 20 notes:

https://verovio.humdrum.org?t=KiprZXJuCipjbGVmQzMKNGMKNGQKNGUKNGYKNGcKPQoqLQohISBoaWRlIGFsbCBzaW5nbGUgbm90ZXM6CiEhIWZpbHRlcjogc2hlZCAtZSAicy8uKi8kMHl5LyIKISEgaGlkZSBhbGwgaW50ZXJwcmV0YXRpb25zOgohISFmaWx0ZXI6IHNoZWQgLWUgInMvLiovWlpaJDAvSSIKISEgcmVzdG9yZSBjbGVmIGFuZCBtYWtlIGl0IGludmlzaWJsZToKISEhZmlsdGVyOiBzaGVkIC1lICJzL1paWihjbGVmLiopLyQxeXkvSSIK

[WolfgangDrescher]

It's a problem with double quotes in the shell. -zsh headed me into the right direction. Running the command with single quotes instead will work correctly:

cat 02-quare-fremuerunt-gentes.krn | shed -s 5 -e 's/^(.+)$/$1yy/D'

!!!COM: Lassus, Rudolphus
!!!CDT: 1563-1625
!!!OPR: Geistliche Psalmen mit dreyen Stimmen
!!!ONM: 2
!!!OTL@@LA: Quare fremuerunt gentes
!!!OTL-incipit@@DE: Warumb empören sich die Haiden
**kern	**text	**kern	**text	**kern	**text
*staff3	*staff3	*staff2	*staff2	*staff1	*staff1
*Ivox	*	*Ivox	*	*Ivox	*
*I"Bassus	*	*I"Tenor	*	*I"Cantus	*
*I'B	*	*I'T	*	*I'C	*
*clefC3	*	*clefC2	*	*clefG2	*
*k[]	*	*k[]	*	*k[]	*
*d:dor	*	*d:dor	*	*d:dor	*
*M2/1	*	*M2/1	*	*M2/1	*
*met(C|)	*	*met(C|)	*	*met(C|)	*
*MM180	*	*MM180	*	*MM180	*
=1	=1	=1	=1	=1	=1
*xywh-#128:294,61,2729,518	*	*xywh-#72:316,78,2764,509	*	*xywh-#12:288,51,2770,524	*
!	!	!LO:TX:a:i:t=c.f.	!	!	!
00r	.	1d	War-	1ryy	.
.	.	2d	-umb	1ayy	War-
.	.	2d	em-	.	.
=2	=2	=2	=2	=2	=2
.	.	2c	-pö-	2ayy	-umb
.	.	2d	-ren	2ayy	em-
.	.	2e	sich	2gyy	-pö-
.	.	2f	die	2ayy	-ren
=3	=3	=3	=3	=3	=3
0r	.	1e	Hai-	2byy	sich
.	.	.	.	2ccyy	die
.	.	2.d	den/	2byy	Hai-
.	.	.	.	2ayy	den/
.	.	4e	.	.	.
!!...

[craigsapp]

Yes, that is right: The terminal shell is trying to interpret $1 as a variable, and so the single quotes around the string telling it not to do that.

$ set "DOLLAR ONE" "DOLLAR TWO" "DOLLAR THREE"

$ echo $1
DOLLAR ONE

$ echo $2
DOLLAR TWO

$ echo $3
DOLLAR THREE

$ echo $1 $2 $3
DOLLAR ONE DOLLAR TWO DOLLAR THREE

$ echo $0
-bash

$ echo "$1 $2 $3"
DOLLAR ONE DOLLAR TWO DOLLAR THREE

$ echo '$1 $2 $3'
$1 $2 $3

$ echo \$1
$1

$ echo \$1 $1
$1 DOLLAR ONE

VHV filters are not running in a shell, so it does not matter if " or ' are used, but I should switch in the documentation to using only ' so there is not confusion when running the shed filter as a command-line tool.