summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/langsci-avm/langsci-avm.dtx
blob: 1a54c0aa2081cfea3e925519e4d8ff6f88e55c97 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
% \iffalse meta-comment
% !TeX program = xelatex
% !TeX encoding = utf-8
%<*internal>
\begingroup
%</internal>
%<*install>
\input{l3docstrip.tex}
\keepsilent
\askforoverwritefalse
\preamble
------------------------------------------------------------------------------

langsci-avm is aimed at typesetting beautiful feature structures, 
also known as attribute-value matrices (AVMs), for use in linguistics.

The package provides a minimal and easy to read syntax. 
It depends only on the array package and can be placed almost everywhere, 
in particular in footnotes or graphs and tree structures. The package is meant
as an update to, and serves the same purpose as, Christopher Manning’s 
avm package, but shares no code base with that package.

langsci-avm was developed at Language Science Press to help in the
production of scientific texts in linguistics.

Copyright (C) 2020 by Felix Kopecky, Language Science Press

This work consists of the file  langsci-avm.dtx 
          and the derived files langsci-avm.ins,
                                langsci-avm.pdf.
                                
It may be distributed and/or modified under the conditions of the
LaTeX Project Public License (LPPL), either version 1.3c of this
license or (at your option) any later version. The latest version
of this license is at <http://www.latex-project.org/lppl.txt>.

This work is ``maintained'' (per LPPL maintenance status) by
Felix Kopecky <felix.kopecky@langsci-press.org>.

The development version can be found at

   https://github.com/langsci/langsci-avm

for those people who are interested. Pull requests are welcome.

Please report any bugs or feature requests to 

   https://github.com/langsci/langsci-avm/issues

------------------------------------------------------------------------------
\endpreamble
\declarepreamble\minimalpreamble
\endpreamble
\postamble
\endpostamble
\usedir{tex/latex/langsci-avm}
\generate{
  \file{\jobname.sty}{\from{\jobname.dtx}{package}}
}
%</install>
%<install>\endbatchfile
%<*internal>
\usepreamble\minimalpreamble
\usepostamble\defaultpostamble
\usedir{source/latex/langsci-avm}
\generate{
  \file{\jobname.ins}{\from{\jobname.dtx}{install}}
}
\endgroup
%</internal>
%<*driver>
\documentclass{l3doc}
\usepackage[linguistics]{forest}
\usepackage{langsci-avm,enumitem,url,gb4e}
\setlist[description]{font=\normalfont}
\begin{document}
  \DocInput{\jobname.dtx}
\end{document}
%</driver>
% \fi
% \title{\pkg{langsci-avm}}
% \author{Felix Kopecky\thanks{\protect\url{mailto:felix.kopecky@langsci-press.org}. 
% 		Please submit bug reports and feature requests to
% 		\protect\url{https://github.com/langsci/langsci-avm/issues}.
% 	}}
% \date{Version 0.2 -- 29th June 2020}
% \maketitle
% \frenchspacing
% \begin{documentation}
% \section{Introduction}
% \pkg{langsci-avm} is a \LaTeX3 package aimed at typesetting beautiful feature structures, also known as \emph{attribute-value matrices}, for use in linguistics. The package provides a minimal and easy to read syntax. It depends only on the \pkg{array} package and can be placed almost everywhere, in particular in footnotes or graphs and tree structures. The package is meant as an update to, and serves the same purpose as, Christopher Manning's \pkg{avm} package, but shares no code base with that package. When you come from \pkg{avm}, please see Section~\ref{sec:switching} for a quick conversion guide.
%
% To start using \pkg{langsci-avm}, place |\usepackage{langsci-avm}| in your preamble.\\\\
% \emph{This document is structured as follows:} Section~\ref{sec:basiccommands} describes the input syntax for AVMs and their parts. Ways to customise your AVM's layout follow in Section~\ref{sec:options}, and selected usage cases are presented in Section~\ref{sec:applications}. There's also an administrative and \TeX nical appendix at the end of this document, in case you are interested.
% 
% \subsection{Example}
% \begin{verbatim}
% \avm{
%   [ ctxt & [ max-qud \\
%              sal-utt & \{ [ cat \\ 
%                             cont <ind & i> ] \}
%            ]
%   ]
% }
% \end{verbatim}
% {\centering\avm{
%     [ ctxt & [max-qud \\
%                 sal-utt & \{ [  cat\\ cont <ind & i> ]  \}\\
%                 ]
%     ]
% }\medskip\\}
% \subsection{Acknowledgements}
% Thanks to Phelype Oleinik for help on recursion and expansion with \LaTeX3. Thanks to Ahmet Bilal Özdemir and Stefan Müller for their contributions in planning and testing this package.
% 
% \section{Structuring AVMs}
% \label{sec:basiccommands}
% 
% \begin{function}{\avm}
%   \begin{syntax}
%     \cs{avm} \oarg{options} \marg{structure}
%   \end{syntax}
%   The heart of this package and its root document comand is |\avm|. In the scope of the command, delimiter characters are processed to open and close (sub-)structures, as described in Section~\ref{sec:commands}. Special elements are described in Section~\ref{sec:specialcommands}. For a description of the layout \meta{options}, see Section~\ref{sec:options}.
% 
%   A \meta{structure} is basically the content of a stylised \cmd{tabular}: The columns are separated by \cmd{&} and a new line is entered with \cmd{\\}.
% \end{function}
%  
% \subsection{Entering (sub-)structures within \texttt{\textbackslash avm}}
% \label{sec:commands}
% 
% \begin{function}{[...],<...>,(...),\{...\}}
% \begin{syntax}
%     \cmd{[} \meta{structure} \cmd{]}
%     \cmd{<} \meta{structure} \cmd{>}
%     \cmd{(} \meta{structure} \cmd{)}
%     \cmd{\{} \meta{structure} \cmd{\}}
%   \end{syntax}
%   Within the scope of \cs{avm}, these delimiters create (sub-)structures that are enclosed by the respective delimiter.  Due to the special meaning that curly braces have in \LaTeX, these are the only ones that need to be run with an escape token (|\|). It is currently possible to mix delimiters, e.g. with |<|\meta{structure}|)|, but this may change in future versions.
%   
%   \pkg{langsci-avm} expects your (sub-)structures to have \emph{at most two columns}, so that for every line in each (sub-)structure, there should be no more than one \cmd{&}. It is recommended to have at least some lines with a \cmd{&} in your \meta{structure}. Currently, display issues may appear in some structures if none are given.\bigskip\\
%   \noindent\begin{minipage}[c]{.5\linewidth}
% \begin{verbatim}
% \avm{
%   [ < ( \{ ... \} ) > ]
% }
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.5\linewidth}
% \avm{
%   [ < ( \{ ... \} ) > ]
% }
% \end{minipage}\bigskip\\
% \noindent\begin{minipage}[c]{.5\linewidth}
% \begin{verbatim}
% \avm{
%   [ \{ ... \} \\
%   < ( ... ), ( ... ) > ]
% }
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.5\linewidth}
% \avm{
%   [ \{ ... \} \\
%   < ( ... ), ( ... ) > ]
% }
% \end{minipage}
% \end{function}
% 
% \begin{function}{!...!}
% \begin{syntax}
%     \cmd{!} \meta{text} \cmd{!}
%   \end{syntax}
%   Escapes the |avm| mode so that all delimiters can be used as usual characters. If you need |!| as a regular character, see Section~\ref{sec:options} for how to change the |switch|.
% \end{function}
% 
% \subsection{Commands for tags, types, unusal lines, and relations}\label{sec:specialcommands}
% \begin{function}[updated=2020-04-29]{\tag,\0,\1,...,\9}
%   \begin{syntax}
%    \cs{tag} \marg{identifier}
%    \cmd{\0, \1, \2, \3, \4, \5, \6, \7, \8, \9}
%   \end{syntax}
%   \cs{tag} puts its \marg{identifier} in a box, more precisely an \cs{fbox}. Within the box, the |tags| font is applied. \cs{0}, \cs{1}, ..., \cs{9} are shortcuts to \cs{tag} and place the respective number in the box. For example, \cs{4} is equivalent to |\tag{4}|. The shortcuts do not take any arguments.
%   
% If you want to use this command outside an AVM, you can obtain, for example, \avm{\4}, by using |\avm{\4}|, or the equivalent  |{\fboxsep.25ex\fbox{\footnotesize 4}}|.\bigskip\\   
% \noindent\begin{minipage}{.6\linewidth}
% \begin{verbatim}
% \avm{[ attr1 & \4\\
%        attr2 & \4[attr3 & val3\\
%                   attr4 & val4] ]}
%  \end{verbatim}
%  \end{minipage}\begin{minipage}{.4\linewidth}
%  \avm{[ attr1 & \4\\
% 	attr2 & \4[attr3 & val3\\
% 	attr4 & val4] ]}
%   \end{minipage}
% \end{function}
% 
% \begin{function}[updated=2020-03-30]{\type,\type*} 
%   \begin{syntax}
%     \cs{type}\meta{*} \marg{type}
%   \end{syntax}
%   Will output the \meta{type} in the |types| font (roman italics by default). The starred variant \cs{type*} will span the complete (sub-)structure and \emph{can only be placed in the first column} of this structure. After the starred \cs{type*}, a |\\| is recommended, but can usually be omitted.\bigskip\\   
%   \noindent\begin{minipage}{.6\linewidth}
% \begin{verbatim}
% \avm{[ \type*{A type spanning a line}
%        attr & [\type{type}] ]}
% \end{verbatim}
%   \end{minipage}\begin{minipage}{.4\linewidth}
%   	\avm{[
%   		\type*{A type spanning a line}
%   		attr & [\type{type}]
%   		]}
%   \end{minipage}
% \end{function}
% 
% \begin{function}{\punk}
% \begin{syntax}
%   \cs{punk} \marg{attribute}\marg{type}
% \end{syntax}
% Some \meta{attributes} think that the layout of the other attributes in their community leaves no space for them to express their individuality. They desire a life outside the confines of the alignment defined by the others, while still remaining a member of the matrix.
%    
% Technically, this is a line with no snapping to the column layout, but with spacing between the \meta{attribute} and \meta{type}. After \cs{punk}, a |\\| is recommended, but can be omitted in ``normal'' cases.\bigskip\\   
% \noindent\begin{minipage}{.6\linewidth}
% \begin{verbatim}
% \avm{[ attr1 & val1\\
%       \punk{a quite long attr2}{val2} ]}
%        attr3 & val3\\
%        attr4 & val4
%       ]}
% \end{verbatim}
% \end{minipage}\begin{minipage}{.4\linewidth}
%    \avm{[
%    	attr1 & val1\\
%    	\punk{a quite long attr2}{val2}
%       attr3 & val3\\
%       attr4 & val4
%    	]}
%    \end{minipage}
% \end{function}
% 
% \begin{function}[updated=2020-03-16]{\+}
%   In the scope of \cs{avm}, \cs{+} comes out as ``$\oplus$''. ``+'' can be obtained normally. \textit{In the earlier Version 0.1.0-beta, \cmd{+} produced ``$\oplus$''.}
% \end{function}
% 
% \begin{function}[added=2020-03-17]{\-}
%   In the scope of \cs{avm}, \cs{-} comes out as ``$\ominus$''. To use the ``optional hyphenation'' meaning of \cs{-}, please write |!\-!|.
% \end{function}
% 
% \begin{function}[added=2020-03-17]{\shuffle}
%   In the scope of \cs{avm}, \cs{shuffle} is a shortcut for ``$\bigcirc$'' to mark the shuffle relation.
% \end{function}
% 
% \section{AVM layout}\label{sec:options}
% 
% \subsection{Defining styles}
% 
% You can customise many aspects of how an AVM is printed, including the fonts or spacing between delimiters and content. You can apply them locally via the \oarg{options} of \cs{avm} or by using \cs{avmsetup}. And you can also define your own styles and use them via the \oarg{style = } option in \cs{avm}. 
% 
% \begin{function}{\avmsetup}
%   \begin{syntax}
%     \cs{avmsetup} \marg{options}
%   \end{syntax}
%   \marg{options} is a comma-separated list of |key = value| settings. See the list below for all user-configurable options. The \marg{options} are the same as in \cs{avm}\oarg{options}. When inserted in \cs{avm}\oarg{options}, they apply locally, and globally if given to \cs{avmsetup}. Local settings always override global ones, and you can have any feasible number of \cs{avmsetup}s in your document.
% 
% \end{function}
% 
% \begin{function}[added=2020-05-11]{\avmdefinestyle}
%   \begin{syntax}
%     \cs{avmdefinestyle} \marg{name} \marg{settings}
%   \end{syntax}
% Instead of applying settings globally or per AVM, you can also define styles and assign them to AVMs, as in |\avm[style=|\meta{name}|]{...}|. The \meta{settings} are a comma-separated list of |key = value| settings, and should be a subset of the settings from \cs{avmsetup}. For example, the following |plain| style highlights neither attributes, values, nor types:
% 
% \begin{verbatim}
% \avmdefinestyle{plain}{attributes=\normalfont,
%                        values=\normalfont,
%                        types=\normalfont}
% \end{verbatim}
% The style is applied with |\avm[style=plain]{...}|.
% \end{function}
% 
% \bigskip
% 
% \noindent Now to the list of settings you can actually apply:
% 
%   \begin{description}[style=nextline]
%   \item[|style =| \meta{name} \hfill (initially empty)] In addition to any style that you possibly define yourself, a style |narrow| is pre-defined in the package (see Section~\ref{sec:spacing}).
% 
%   \item[|stretch =| \meta{factor} \hfill (initially |0.9|)] Define \cs{arraystretch}, i.e. a factor in the determination of line height.
%   
%   \item[|columnsep =| \meta{length} \hfill (initially |0.5ex|)] Define the \cs{tabcolsep}, i.e. horizontal space between columns. The first and second column will have |0\columnsep| to the left and right, respectively. Between the two the distance is |2\columnsep|. Using relative units (like |ex| or |em|) may be a good idea so that |columnsep| scales well with changes in font size.
%   
%   \item[|delimfactor =| \meta{factor} \hfill (initially |1000|)] Sets \cs{delimiterfactor}. The calculation for the minimum height of a delimiter is $y \cdot f /1000$, where $y$ is the height of the content and $f$ the value of  \cmd{delimfactor}. The default |1000| ensure that the delimiters' height is at least that of the structure.
%   
%   \item[|delimfall =| \meta{length} \hfill (initially |0pt|)] Controls \cs{delimitershortfall}, i.e. the maximum height that the delimiters can be shorter than the enclosed structure. The default |0pt| ensure that the delimiters are not shorter than the contents.
%   
%   \item[|extraskip| = \meta{length} \hfill (initially \cs{smallskipamount})] If a substructure is immediately followed by a \cmd{\\}, an extra amount of vertical skip is added so that the content of the next line, possibly another delimiter, does not clash with the delimiter in that line. This automatic skip insertion can be circumvented with placing a \cs{relax} before the linebreak, i.e. |\relax\\|.
%   
%   \item[|attributes =| \meta{font settings} \hfill (initially |\textbackslash scshape|)] The font for attributes, i.e. the first column of each structure.
%   
%   \item[|values =| \meta{font settings} \hfill (initially |\textbackslash itshape|)] The font for values, i.e. the second column of each structure.
% 
%   \item[|apptovalues=| \meta{code} \hfill (initially \cs{/})] The \meta{code} is applied after the second column (``append to''). This is useful if |values| is set to |\itshape|, since \cs{itshape} does not automatically insert italic correction.
%   
%   \item[|types =| \meta{font settings} \hfill (initially |\textbackslash itshape|)] The font used in \cs{type} and \cs{type*}.
%   \item[|tags =| \meta{format settings} \hfill (initially |\textbackslash footnotesize|)] The font (size) used in \cs{tag} and the shortcuts \cs{1}...\cs{9}.
%   \item[|switch =| \meta{token} \hfill (initially |!|)] Define the escape token. Change this if you need to use ``!'' as a text glyph.
%   \item[|customise = | \meta{settings} \hfill (initially empty)] An interface to input custom commands to be run at the beginning of every \cs{avm}.
%   \end{description}
% 
% \subsection{Defining input patterns}
% 
% \begin{function}[added=2020-06-29]{\avmdefinecommand}
%   \begin{syntax}
%     \cs{avmdefinecommand} \marg{name} \oarg{label} \marg{settings}
%   \end{syntax}
% Sub-structures often come in patterns. For example, AVMs often have a \textsc{phon} attribute, which is mapped to a list, the entries of which are in italics. \cs{avmdefinecommand} can account  for this and other input patterns. For example,
% 
% \begin{verbatim}
% \avmdefinecommand{custom}{...}
% \end{verbatim}
% will create a command \cs{custom} available only in the scope of \cs{avm} (this means that you can have a different meaning in the rest of your document). The \meta{settings} will then be applied to the scope in which \cs{custom} is called. If an optional \meta{label} is given, the label will be printed, in the current font, before the \meta{settings} are applied.
% 
% \cs{custom} generated in this way automatically advances to the value column after the \meta{label} is printed. This means that commands generated with \cs{avmdefinecommand} should be called in the attribute column of an existing structure. This behaviour can be circumvented with the starred variant \cs{name*}, which is automatically generated by \cs{avmdefinecommand} as well. However, it seems advisable to use the starred variants sparingly. 
% 
% Here's an example for the aforementioned \text{phon} pattern:
% 
% \begin{verbatim}
% \avmdefinecommand{phon}[phon]
%   {
%     attributes  = \itshape,
%     delimfactor = 900,
%     delimfall   = 10pt
%   }
% \end{verbatim}
% 
% This creates a command \cs{phon} (and the variant \cs{phon*}) within the scope of any \cs{avm}. It will print the label \texttt{phon} in the current font and then apply three settings locally: italics for the attribute (first) column, and two settings for very narrow delimiter fitting. 
% 
% This results in: (The font of this documentation has little support for IPA.)\avmdefinecommand{phon}[phon] {attributes  = \itshape,delimfactor = 900, delimfall   = 10pt}\\\\
% \noindent\begin{minipage}[c]{.5\linewidth}
% \begin{verbatim}
% \avm{
%   [\type*{word}
%    \phon <lin'gwistiks>\\
%    synsem & [ ... ]
%   ]
% }
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.5\linewidth}
% \avm{
%   [\type*{word}
%    \phon <lin'gwistiks>\\
%    synsem & [ ... ]
%   ]
% }
% \end{minipage}
% 
% Note that any other structure type would have worked instead of \avm{<>}. But \avm{<>} and any other markers for sub-structures are left unchanged by \cs{phon} and other custom commands. This is why the \emph{attribute} font is changed by \cs{phon}, although \emph{lin'gwistiks} is technically a value. Remember that \cmd{<} creates a new list sub-substructure, and the first content is printed in its attribute font.
% \end{function}
% 
% \section{Applications}\label{sec:applications}
% 
% \subsection{Spacing and size of delimiters}\label{sec:spacing}
% \pkg{langsci-avm} automatically detects if the end of a sub-structure is followed by a line break. This is useful to find cases in which two sub-structures are printed immediately below each other, and to add extra spacing (the |extraskip| from the options). This automatic detection can be suppressed with \cs{relax}. See below for the effect of that detection:\bigskip\\
% \noindent\begin{minipage}[c]{.5\textwidth}\centering
% \begin{verbatim}
% \avm{[ [attr1 & val1 \\ 
%         attr2 & val2 ] \\
%        [attr1 & val1 \\ 
%         attr2 & val2 ]
%      ]}
% \end{verbatim}
% 
% \avm{[ [attr1 & val1 \\ 
%         attr2 & val2 ] \\
%        [attr1 & val1 \\ 
%         attr2 & val2 ]
%      ]}
% \end{minipage}\begin{minipage}[c]{.5\textwidth}\centering
% \begin{verbatim}
% \avm{[ [attr1 & val1 \\ 
%         attr2 & val2 ] \relax\\
%        [attr1 & val1 \\ 
%         attr2 & val2 ]
%      ]}
% \end{verbatim}
% 
% \avm{[ [attr1 & val1 \\ 
%         attr2 & val2 ] \relax\\
%        [attr1 & val1 \\ 
%         attr2 & val2 ]
%      ]}
% \end{minipage}\\\\
% 
% If many delimiters are nested, this occasionally results in larger delimiter sizes. There is a pre-defined |narrow| style that resets |delimfall| (to |5pt|) and |delimfactor| (to |997|), which are  the values recommended in the \textit{\TeX book}. This results in a more compact appearance:\bigskip\\ 
% \noindent\begin{minipage}[t]{.5\textwidth}\centering
% \begin{verbatim}
% \avm{[ attr \{<\1>\}]}
% \end{verbatim}
% 
% \avm{[ attr <\{\1\}>]}
% \end{minipage}\begin{minipage}[t]{.5\textwidth}\centering
% \begin{verbatim}
% \avm[style=narrow]{[ attr \{<\1>\}]}
% \end{verbatim}
% 
% \avm[style=narrow]{[ attr <\{\1\}>]}
% \end{minipage}  
% 
% \subsection{Disjunctions and other relations}
% Sometimes AMVs are placed beside other content to express disjunctions or other relations. In \pkg{langsci-avm} this is done naturally:\bigskip\\
% \noindent\begin{minipage}[c]{.5\textwidth}
% \begin{verbatim}
% \avm{ [attr1 & val1\\
%        attr2 & val2\\
%        attr3 & val3] } $\lor$ 
% \avm{ [attr1' & val1'\\
%        attr2' & val2'\\
%        attr3' & val3'\\] }
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.5\textwidth}
%  \avm{ [attr1 & val1\\
%          attr2 & val2\\
%          attr3 & val3] } $\lor$ 
%  \avm{ [attr1' & val1'\\
%         attr2' & val2'\\
%         attr3' & val3'\\] }
% \end{minipage}\bigskip\\
% \noindent\begin{minipage}[c]{.5\textwidth}
% \begin{verbatim}
% \textit{sign} $\to$ 
% \avm{ [ attribute1 & value1\\
%         attribute2 & value2\\
%         attribute3 & value3 ] }
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.5\textwidth}
%   \textit{sign} $\to$ 
% \avm{ [ attribute1 & value1\\
% 	attribute2 & value2\\
% 	attribute3 & value3 ] }
% \end{minipage}
% 
% \pagebreak\subsection{Use as a vector}
% It's possible to use \pkg{langsci-avm} for feature vectors rather than matrices, as may be useful in generative grammar.\medskip\\
% \noindent\begin{minipage}[c]{.8\textwidth}
% \begin{verbatim}
% \avm[attributes=\normalfont]{[v1\\v2\\v3]}$\varphi$
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.2\textwidth}
% \avm[attributes=\normalfont]{[v1\\v2\\v3]}$\varphi$
% \end{minipage}
% 
% \subsection{Combinations with \pkg{gb4e}, \pkg{expex}, and \pkg{linguex}}
% 
% This package works fine with \pkg{gb4e} and its fork \pkg{langsci-gb4e}. To align the example number at the top of your structure, please use \cs{attop} from \pkg{gb4e}:\bigskip\\
% \noindent\begin{minipage}[c]{.5\textwidth}
% \begin{verbatim}
% \begin{exe}
%     \ex\attop{
%     \avm{[ attr1 & val1\\
%            attr2 & val2\\
%            attr3 & val3]}
%     }
%   \end{exe}
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.5\textwidth}
%   \begin{exe}
% 	\ex\attop{
% 		\avm{[ attr1 & val1\\
% 			attr2 & val2\\
% 			attr3 & val3]}
% 	}
% \end{exe}
% \end{minipage}\bigskip\\
% 
% \noindent The same can be achieved with \pkg{expex} using \cs{envup} from \pkg{lingmacros} (see below) or using this \emph{experimental} syntax:
% 
% \begin{verbatim}
% \ex \vtop{\strut\vskip-\baselineskip{
%   \avm{[ attr1 & val1\\
%          attr2 & val2\\
%          attr3 & val3]} 
% }}
% \xe
% \end{verbatim}
% 
% \noindent Examples typed with \pkg{linguex} can be combined with \cs{evnup} from \pkg{lingmacros} to align AVMs (many thanks to Jamie Findlay for pointing this out):
% 
% \begin{verbatim}
% \ex. \envup{\avm{[ attr1 & val1\\
%          attr2 & val2\\
%          attr3 & val3]} 
%       }
% \end{verbatim}
% 
% \subsection{Combinations with \pkg{forest}}
% This package also works fine with \pkg{forest}. As per the \pkg{forest} documentation, it is recommended to protect any \cs{avm}-statements with |{}| in nodes:\bigskip\\
% \noindent\begin{minipage}[c]{.6\textwidth}
% \begin{verbatim}
% \begin{forest}
%  [A [B] [{\avm{[attr1 & val1\\
%                 attr2 & val2\\
%                 attr3 & val3]}} ] ]
% \end{forest}
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.4\textwidth}
% \begin{forest}
% 	[A [B] [{\avm{[attr1 & val1\\
% 			attr2 & val2\\
% 			attr3 & val3]}} ] ]
% \end{forest}
% \end{minipage}
% 
% It may happen that extensive AVMs protrude into the space reserved for other \pkg{forest} nodes or edges. In this case, the \pkg{forest} setting \cmd{for children = \{anchor=north\}} may be useful: (If you like, try this tree without that setting.)\bigskip\\
% \noindent\begin{minipage}[c]{.6\textwidth}
% \begin{verbatim}
% \begin{forest}
%   [A, for children = {anchor=north}
%      [B] [{\avm{[attr1 & val1\\
%      attr2 & a long value val2\\
%      attr3 & val3\\
%      attr4 & val4\\
%      attr5 & val5]}} ] 
%   ]
% \end{forest}
% \end{verbatim}
% \end{minipage}\begin{minipage}[c]{.4\textwidth}
% \begin{forest}
% 	[A, for children = {anchor=north} [B] [{\avm{[attr1 & val1\\
% 			attr2 & a long value val2\\
%           attr3 & val3\\
%           attr4 & val4\\
% 			attr5 & val5]}} ] ]
% \end{forest}
% \end{minipage}
% 
% \subsection{Switching from Christopher Manning's \pkg{avm} package}\label{sec:switching}
% Switching from \pkg{avm} to \pkg{langsci-avm} will require some, though hopefully minimal, changes to the code. In particular, \pkg{langsci-avm} doesn't distinguish between ``active'' and ``passive'' modes, there is now a single way of sorting (see \cs{type}, which replaces \cs{asort} and \cs{osort}), and tags are now produced without |@| (\cs{4} instead of \cmd{@4}, etc.). 
% 
% Paths can be printed with a normal \cmd{|}, and $\oplus$ and other relation symbols can be input more easily (see Section~\ref{sec:commands}), though the package will also work with \verb+$|$+ and \verb+$\oplus$+. 
% 
% \pkg{langsci-avm} is not yet able to draw lines in elements of AVMs. This feature is planned for Version 0.3.
% 
% 
% \subsection{Spanning both columns}
% You can use the \pkg{multicol} package to span both columns in a (sub-)structure. Please remember that every structure has two columns, so the only sensible usage is 
% 
% \begin{verbatim}
% \multicolumn{2}{l}{...}
% \end{verbatim}
% 
% \noindent but only in the first column of a (sub-)structure. For a special usage case, see \cs{type} and \cs{type*} (which do not depend on \pkg{multicol}).
% 
% \section{Caveats and planned features}\label{sec:caveats}
% \begin{enumerate}
% \item There are currently no error messages. If you do not receive the intended output, please make sure that your code fits the syntax described in this documentation. If your code is fine but the output is not, please submit a bug report or feature request at \url{https://github.com/langsci/langsci-avm/issues}.
% \item[] These features are planned for the future:
% \item A check whether the delimiters are balanced, i.e. whether all (sub-)structures are closed by a \verb+]+, \verb+}+, etc.
% \item Introduce the ability to draw (curved) lines between structures and elements.
% \item Improve the appearance of (very) large angle brackets so that they vertically span the complete structure they enclose, maybe using \pkg{scalerel}.
% \end{enumerate}
% 
% \section{Feedback and bug reports}
% Comments, usage reports, and feature requests are welcome! Please open an issue for any of these at \url{https://github.com/langsci/langsci-avm/issues}, or write to me at \url{mailto:felix.kopecky@langsci-press.org} if you feel the need for a feature not listed here, big or small.
% \end{documentation}
% 
% \begin{implementation}
% 
% \section{Implementation}
%    \begin{macrocode}
%<*package>
%    \end{macrocode}
%    \begin{macrocode}
%<@@=avm>
%    \end{macrocode}
%    \begin{macrocode}
\RequirePackage{xparse,array}
%    \end{macrocode}
%
%    \begin{macrocode}
\ProvidesExplPackage {langsci-avm} 
  {2020-06-29} {0.2}
  {AVMs and feature structures in LaTeX3}
%    \end{macrocode}
%    \begin{macro}{\avm}
% This document command initialises an AVM. The first, optional argumet is a key-value list of settings (see \cs{keys_define:nn} below) and the second is the AVM itself, given in the syntax described in this documentation.
% 
% \cs{avm} enters a group so that keys- and macro-assignemts remain local. It then initialises the commands and shortcuts and any user customisation, sets its mode to |true| and assigns the keys as given in the optional argument (if any). After the wrapper \cs{avm_wrap:n} is called, the group is closed.
%
%    \begin{macrocode}
\NewDocumentCommand{\avm}{ O{} +m }
  {
    \c_group_begin_token
    \__avm_initialise_document_commands:
    \__avm_initialise_custom_commands:
    \tl_use:N \l__avm_defined_commands_tl
    \bool_set_true:N \l__avm_mode_bool
    \keys_set:nn { avm } { #1 }
    \__avm_wrap:n { #2 }
    \c_group_end_token
  }
%    \end{macrocode}
%    \end{macro}
% 
%    \begin{macro}{\avmsetup}
% Forward the key-value settings given as the optional argument to \cs{avm} to the keys defined in |\keys_define:nn { avm }|. For the meaning of these keys and initial values, see Section~\ref{sec:basiccommands}.
% 
%    \begin{macrocode}
\NewDocumentCommand{\avmsetup}{ m } 
  { \keys_set:nn { avm } { #1 } }

\keys_define:nn { avm }
  {
    stretch .code:n        = {\def\arraystretch{#1}},
    stretch .initial:n     = {0.9},
    columnsep .dim_set:N   = \tabcolsep,
    columnsep .initial:n   = {.5ex},
    delimfactor .int_set:N = \delimiterfactor,
    delimfactor .initial:n = {1000},
    delimfall .dim_set:N   = \delimitershortfall,
    delimfall .initial:n   = {0pt},
    attributes .code:n     = {\cs_set:Nn \__avm_font_attribute: {#1}},
    attributes .initial:n  = {\scshape},
    types .code:n          = {\cs_set:Nn \__avm_font_type: {#1}},
    types .initial:n       = {\itshape},
    values .code:n         = {\cs_set:Nn \__avm_font_value: {#1}},
    values .initial:n      = {\itshape},
    tags .code:n           = {\cs_set:Nn \__avm_font_tag: {#1}},
    tags .initial:n        = {\footnotesize},
    apptovalues .code:n    = {\cs_set:Nn \__avm_deinit_second_column: {#1}},
    apptovalues .initial:n = {\/},
    singleton .code:n      = {\cs_set:Nn \__avm_font_singleton: {#1}},
    singleton .initial:n   = {\normalfont},    
    switch .code:n         = {\tl_set:Nn \__avm_mode_switch_character {#1}},
    switch .initial:n      = { ! },
    extraskip .dim_set:N   = \l__avm_extra_skip_dim,
    extraskip .initial:n   = {\smallskipamount},
    customise .code:n      = {\cs_set:Nn \__avm_initialise_custom_commands: {#1}},
    customise .initial:n   = { },
    style .choice:,
    style / narrow .code:n = {\delimiterfactor=997\delimitershortfall5pt},
  }
%    \end{macrocode}
%    \end{macro}
  
%    \begin{macro}{\avmdefinestyle}
% Define a style to be used together with the |style| key.
%    \begin{macrocode}
\NewDocumentCommand{\avmdefinestyle}{ m m }
  {
    \keys_define:nn { avm }
      {
        style / #1 .code:n = { \keys_set:nn { avm } { #2 } }
      }
  }
%    \end{macrocode}
%    \end{macro}

%    \begin{macro}{\avmdefinecommand}
% A factory function that creates commands for the layout of sub-structures and saves them to \cs{l__avm_defined_commands_tl}. The first argument describes the command's name, the second any (optional) label. The manufactured definitions are activated in the AVM group so that they remain local.
%    \begin{macrocode}
\NewDocumentCommand{\avmdefinecommand}{ m O{} m }
  {
    \tl_put_right:Nn \l__avm_defined_commands_tl
      {
        \exp_args:Nc \DeclareDocumentCommand { #1 } { s } 
          { 
             #2 \IfBooleanF { ##1 } { & } \avmsetup{ #3 } 
          } 
      }
  }
%    \end{macrocode}
%    \end{macro}
  
%    \begin{variable}{\l__avm_mode_bool, \l__avm_parens_tracker, \l__avm_defined_commands_tl}
% We need an auxiliary variable to store the current mode. \cs{l__avm_parens_tracker} is a stack for a future check whether the delimiters given to \cs{avm} are balanced. \cs{l__avm_defined_commands_tl} is a token list that stores any commands provided by the user via \cs{avmdefinecommand}
%    \begin{macrocode}
\bool_new:N \l__avm_mode_bool
\seq_new:N \l__avm_parens_tracker
\tl_new:N \l__avm_defined_commands_tl
%    \end{macrocode}
%    \end{variable}

%    \begin{macro}[int]{\seq_set_split:NVn}
% In preparation for \cs{avm_wrap:n}, we need to split the user input at each occurrence of the escape character. Since the character is given in a variable, we need a variant of the sequence splitter that takes the \emph{evaluation} of the variable, rather than the variable itself, as its second argument.
%    \begin{macrocode}
\cs_generate_variant:Nn \seq_set_split:Nnn { NVn }
%    \end{macrocode}
%    \end{macro}

%    \begin{variable}{\l__avm_in_first_column}
% A boolean to check whether we are in the first column (value |true|) or in the second (value |false|).
%    \begin{macrocode}
\bool_new:N \l__avm_in_first_column
%    \end{macrocode}
%    \end{variable}

%    \begin{macro}[int]{\__avm_init_first_column:, \__avm_init_second_column:}
% These macros apply the settings for the columns in a (sub-)structure. They take care of font selection and report the currently active column back to the system. Knowing which column is active is important when closing the (sub-)structure. If the structure is closed without a second column present, we need to skip back |2\tabcolsep|.
%    \begin{macrocode}
\cs_new:Nn \__avm_init_first_column:
  {
    \bool_set_true:N \l__avm_in_first_column
    \normalfont\__avm_font_attribute:
  }

\cs_new:Nn \__avm_init_second_column:
  {
    \bool_set_false:N \l__avm_in_first_column
    \normalfont\__avm_font_value:
  }
%    \end{macrocode}
%    \end{macro}

%    \begin{macro}[int]{\__avm_kern_unused_columns:}
% A helper macro to fill the horizontal space if a row is ended prematurely, i.e. if no \cmd{&} is present.
%    \begin{macrocode}
\cs_new:Nn \__avm_kern_unused_columns:
  {
    \bool_if:NTF \l__avm_in_first_column
      { \span\hspace*{-2\tabcolsep} }
      {  }
  }
%    \end{macrocode}    
%    \end{macro}

%    \begin{macro}[int]{\__avm_extra_skip:}
% This function is used together with the delimiter replacements. It checks whether the delimiter is followed by a line break, in which case an extra skip is automatically inserted
%    \begin{macrocode}
\cs_new:Nn \__avm_extra_skip:
  {
    \peek_meaning_ignore_spaces:NTF \\ {\vspace*{\l__avm_extra_skip_dim}} {}
  }
%    \end{macrocode}
%    \end{macro}

%    \begin{macro}[int]{\__avm_module_begin:,\__avm_module_end:,etc.}
% The replacement instructions for \cs{__avm_parse:n}
%    \begin{macrocode}
\cs_new:Nn \__avm_module_begin: 
  { 
    \begin{tabular}{@{} 
                    >{\__avm_init_first_column:}l 
                    >{\__avm_init_second_column:}l 
                    <{\__avm_deinit_second_column:}
                    @{}}
  }
\cs_new:Nn \__avm_module_end:
  { 
    \__avm_kern_unused_columns:
    \end{tabular}
  }
\cs_new:Nn \__avm_replace_lbrace: 
  { 
    \__avm_parse_output:nw 
      {  \c_math_toggle_token\left\lbrace\__avm_module_begin:  } 
  }
\cs_new:Nn \__avm_replace_rbrace: 
  { 
    \__avm_parse_output:nw 
      { \__avm_module_end:\right\rbrace\c_math_toggle_token\__avm_extra_skip: }
  }
\cs_new:Nn \__avm_replace_lbrack: 
  { 
    \__avm_parse_output:nw 
      {  \c_math_toggle_token\left\lbrack\__avm_module_begin:  } 
  }     
\cs_new:Nn \__avm_replace_rbrack: 
  { 
    \__avm_parse_output:nw 
      {  \__avm_module_end:\right\rbrack\c_math_toggle_token\__avm_extra_skip:  } 
  }
\cs_new:Nn \__avm_replace_lparen:
  {    
    \__avm_parse_output:nw 
      {  \c_math_toggle_token\left(\__avm_module_begin:  }
  } 
\cs_new:Nn \__avm_replace_rparen: 
  { 
    \__avm_parse_output:nw 
      {  \__avm_module_end:\right)\c_math_toggle_token\__avm_extra_skip:  } 
  } 
\cs_new:Nn \__avm_replace_langle: 
  { 
    \__avm_parse_output:nw 
      {  \c_math_toggle_token\left<\__avm_module_begin:  } 
  } 
\cs_new:Nn \__avm_replace_rangle: 
  { 
    \__avm_parse_output:nw 
      { \__avm_module_end:\right>\c_math_toggle_token\__avm_extra_skip: } 
  }
\cs_new:Nn \__avm_replace_plus:
  { 
    \__avm_parse_output:nw { \ensuremath { \oplus \! } } 
  }
\cs_new:Nn \__avm_replace_minus:
  { 
    \__avm_parse_output:nw { \ensuremath { \ominus \! } } 
  }
\cs_new:Nn \__avm_replace_circle:
  { 
    \__avm_parse_output:nw { \ensuremath { \bigcirc \, } } 
  }
%    \end{macrocode}
%    \end{macro}

%    \begin{macro}[updated=2020-03-30]{\tag,\type,\punk}
%    \begin{macrocode}
\cs_new:Npn \__avm_controls_tag:n #1 
  { \fboxsep.25ex\fbox{\normalfont\__avm_font_tag: #1} } 
\cs_new:Npn \__avm_controls_type:n #1 
  { \c_group_begin_token\normalfont\__avm_font_type: #1\c_group_end_token }
\cs_new_protected:Npn \__avm_controls_type_starred:n #1 
  {
    \bool_set_false:N \l__avm_in_first_column
    \normalfont\__avm_font_type: #1 
    \__avm_deinit_second_column:\span\hspace*{-2\tabcolsep}
    \peek_meaning_ignore_spaces:NTF \\ {} {\\} 
  }
\cs_new_protected:Npn \__avm_controls_punk:nn #1 #2 
  {  
    \bool_set_false:N \l__avm_in_first_column
    \normalfont\c_group_begin_token\__avm_font_attribute:#1%
    \c_group_end_token\hspace{2\tabcolsep}%
    \c_group_begin_token\__avm_font_type: #2\c_group_end_token%
    \__avm_deinit_second_column:\span\hspace*{-2\tabcolsep}
    \peek_meaning_ignore_spaces:NTF \\ {} {\\}                                               
  }

\cs_new:Nn \__avm_initialise_document_commands: 
  {
    \cs_if_exist:NTF \tag 
      {  \RenewDocumentCommand{\tag}{m}{ \__avm_controls_tag:n {##1} } }
      {  \NewDocumentCommand{\tag}{m}{   \__avm_controls_tag:n {##1} } }
    \cs_if_exist:NTF \0
      {  \RenewDocumentCommand{\0}{}{    \__avm_controls_tag:n {0} } }
      {  \NewDocumentCommand{\0}{}{      \__avm_controls_tag:n {0} } }
    \cs_if_exist:NTF \1   
      {  \RenewDocumentCommand{\1}{}{    \__avm_controls_tag:n {1} } }
      {  \NewDocumentCommand{\1}{}{      \__avm_controls_tag:n {1} } }
    \cs_if_exist:NTF \2
      {  \RenewDocumentCommand{\2}{}{    \__avm_controls_tag:n {2} } }
      {  \NewDocumentCommand{\2}{}{      \__avm_controls_tag:n {2} } }
    \cs_if_exist:NTF \3
      {  \RenewDocumentCommand{\3}{}{    \__avm_controls_tag:n {3} } }
      {  \NewDocumentCommand{\3}{}{      \__avm_controls_tag:n {3} } }
    \cs_if_exist:NTF \4
      {  \RenewDocumentCommand{\4}{}{    \__avm_controls_tag:n {4} } }
      {  \NewDocumentCommand{\4}{}{      \__avm_controls_tag:n {4} } }
    \cs_if_exist:NTF \5
      {  \RenewDocumentCommand{\5}{}{    \__avm_controls_tag:n {5} } }
      {  \NewDocumentCommand{\5}{}{      \__avm_controls_tag:n {5} } }
    \cs_if_exist:NTF \6
      {  \RenewDocumentCommand{\6}{}{    \__avm_controls_tag:n {6} } }
      {  \NewDocumentCommand{\6}{}{      \__avm_controls_tag:n {6} } }
    \cs_if_exist:NTF \7
      {  \RenewDocumentCommand{\7}{}{    \__avm_controls_tag:n {7} } }
      {  \NewDocumentCommand{\7}{}{      \__avm_controls_tag:n {7} } }
    \cs_if_exist:NTF \8
      {  \RenewDocumentCommand{\8}{}{    \__avm_controls_tag:n {8} } }
      {  \NewDocumentCommand{\8}{}{      \__avm_controls_tag:n {8} } }
    \cs_if_exist:NTF \9
      {  \RenewDocumentCommand{\9}{}{    \__avm_controls_tag:n {9} } }
      {  \NewDocumentCommand{\9}{}{      \__avm_controls_tag:n {9} } }
    \cs_if_exist:NTF \type 
      {  \RenewDocumentCommand{\type}{s m}
        {  
          \IfBooleanTF { ##1 } 
            {  \__avm_controls_type_starred:n {##2}  }
            {  \__avm_controls_type:n {##2} }
         } 
       }
       {  \NewDocumentCommand{\type}{s m}
         {
           \IfBooleanTF { ##1 }
             {  \__avm_controls_type_starred:n {##2}  }
             {  \__avm_controls_type:n {##2}  }
          } 
        }
    \cs_if_exist:NTF \punk 
      {  \RenewDocumentCommand{\punk}{m m}
        { \__avm_controls_punk:nn {##1}{##2} } }
      {  \NewDocumentCommand{\punk}{m m}
        {  \__avm_controls_punk:nn {##1}{##2} } }
  }
%    \end{macrocode}
%    \end{macro}

%    \begin{macro}{\__avm_wrap:n}
% The wrapper that first splits the input to \cs{avm} at each occurrence of \cs{__avm_mode_switch_character} and then inverses \cs{l__avm_mode_bool}. It then calls the parser (\cs{__avm_parse:n}) for each splitted sequence. This wrapping is necessary because there is no known expandable way to switch a boolean.
%    \begin{macrocode}
\cs_new_protected:Npn \__avm_wrap:n #1
  {
    \seq_set_split:NVn \l__avm_wrapper_seq 
    \__avm_mode_switch_character { #1 }
    \seq_map_inline:Nn \l__avm_wrapper_seq 
      {  
        \exp_args:No \exp_not:o 
          {  \__avm_parse:n {##1} }  
        \bool_set_inverse:N \l__avm_mode_bool 
      }
  }
%    \end{macrocode}
%    \end{macro}

%    \begin{macro}{\__avm_parse:n}
% Finnaly, the parser. It is build on \cs{@@@@_act:NNNnn} from \cmd{l3tl} (see the sub-section \emph{Token by token changes}). Many thanks to Phelype Oleinik for help on this, and in particular on help with expansion. 
%    \begin{macrocode}
\cs_new:Npn \__avm_parse:n #1
  {
      \exp:w
      \group_align_safe_begin:
        \__avm_parse_loop:w #1 
        \q_recursion_tail \q_recursion_stop 
        \__avm_result:n { }
  }

\cs_new:Npn \__avm_end:w \__avm_result:n #1
  {
    \group_align_safe_end:
    \exp_end:
    #1
  }

\cs_new:Npn \__avm_parse_loop:w #1 \q_recursion_stop
  {
    \tl_if_head_is_N_type:nTF {#1}
      {
        \__avm_N_type:N #1 \q_recursion_stop 
      }
      {
        \tl_if_head_is_group:nTF {#1}
          { \__avm_replace_group:nw #1 \q_recursion_stop }
          { \__avm_replace_space:w #1 \q_recursion_stop }
      }
  }
  
\cs_new:Npn \__avm_N_type:N #1
  {
    \quark_if_recursion_tail_stop_do:Nn #1 { \__avm_end:w }
    \bool_if:NTF \l__avm_mode_bool 
      { \__avm_replace:N #1 } 
      { \__avm_replace_none:N #1 } 
  }
  
\cs_new:Npn \__avm_replace_none:N #1 
  { 
    \__avm_parse_output:nw {#1}
  }

\cs_new:Npn \__avm_replace:N #1
  {
    \str_case:nnF {#1}
      {
        { \+ }{ \__avm_replace_plus: }
        { \- }{ \__avm_replace_minus: }
        { \shuffle }{ \__avm_replace_circle: }
        { [  }{ \__avm_replace_lbrack: }
        { ]  }{ \__avm_replace_rbrack: }
        { (  }{ \__avm_replace_lparen: }
        { )  }{ \__avm_replace_rparen: }
        { \{ }{ \__avm_replace_lbrace: }        
        { \} }{ \__avm_replace_rbrace: }                
        { <  }{ \__avm_replace_langle: } 
        { >  }{ \__avm_replace_rangle: } 
      }
      { \__avm_replace_none:N #1 }
  }

\cs_new:Npn \__avm_replace_group:nw #1
  { \exp_args:NNo \exp_args:No \__avm_replace_group:n { \__avm_parse:n {#1} } }

\cs_new:Npn \__avm_replace_group:n #1 { \__avm_parse_output:nw { {#1} } }

\exp_last_unbraced:NNo
\cs_new:Npn \__avm_replace_space:w \c_space_tl { \__avm_parse_output:nw { ~ } }

\cs_new:Npn \__avm_parse_output:nw #1 #2 \q_recursion_stop \__avm_result:n #3
  { \__avm_parse_loop:w #2 \q_recursion_stop \__avm_result:n {#3  #1 } }
%    \end{macrocode}
%    \end{macro}
%    \begin{macrocode}
%</package>
%    \end{macrocode}
%
% \end{implementation}

% \PrintIndex