erickcan / sound-change-applier

A command-line sound change applier written in Python.

sound-change-applier

A command-line sound change applier written in Python.

Setup

To begin using this sound change applier, you just need to download the executable (.exe) file in the Releases page and use it through the command-line prompt:

$ sca sca-options

You can also clone this repo and use the sca.bat or sca.sh file:

$ git clone https://github.com/erickcan/sound-change-applier.git
$ sca sca-options       # if Windows
$ ./sca.sh sca-options  # if Linux

Command line

sca (-f rules-file words-file | -n named-rules-json named-rule words)
    [--csv-output] [-s sound-classes-json | --no-sound-classes]

Usage

• --file-based-sound-change | -f to apply a set of rules to a set of words, each defined in a separate file, creating a text file with the changed words;
  • rules-file: text file with the rules
  • words-file: text file with the words
• --named-sound-change | -n to apply a named sound change to words passed on the command line, and then prints the words after the change;
  • named-rules-json: JSON file where sound classes are defined
  • named-rule: name of the rule to apply (defined in named-rules-json)
  • words: words to apply the sound change
• --csv-output to create a before and after of the changed words;
• --sound-classes-file | -s to specify the JSON file where the sound classes are defined (see also Default sound classes);
• --no-sound-classes to not make use of sound classes.

Examples

Suppose the following files:

named_rules.json

{
  "terminal-devoicing": "[bdgz] > [ptks] / _#",
  "h-dropping": "h -> _ / _",
  "z-rhotacization": "z -> r / V_V",
  "l-vocalization": "l => w / _!V",
  "/æ/-raising": "æ => eə / _N",
  "th-fronting": "[θð] -> [fv] / _"
}

rules.txt

d => ð / V_V
[gk] > _ / #_n
a -> ə / _#
l -> w / _!V
[aou]e > [æøy] / _
sw > s / _[ou]
[ao]N -> [ãõ] / _
s => z / [bdg]_
e -> _ / CC_#

words.txt

aeon
beds
clue
daemon
knowledge
laterals
media
sand
swords

If this were passed to the command line,

$ sca -f rules.txt words.txt --csv-output

the following file would be created:

sound-change-YYYY-MM-DD-HH-MM-SS.csv

aeon,æõ
beds,bedz
clue,cly
daemon,dæmõ
knowledge,nowledg
laterals,lateraws
media,meðiə
sand,sãd
swords,sordz

And if the following were passed into the command line,

$ sca -n named_rules.json h-dropping "here he had hallucinated"

it would print:

ere e ad allucinated

Rule notation

The rules can be written in any of the following forms:

x  > y / a_b
x -> y / a_b
x => y / a_b

where x becomes y when x is between a and b. (Spaces between >, ->, => and / are insignificant.)

Symbols

symbol	meaning	example
#	word boundary	_#: end of word, #_: start of word
A..Z	sound class	Gt: class G followed by t
[xyz]	ad-hoc class	u[rl]: u followed by r or l
> _	sound-eraser	h -> _: deletes h
_	everywhere, when	_x: when followed by x, _: everywhere
!_	not preceded by	u!_: when not preceded by u
_!	not followed by	_!n: when not followed by n

Appendix

Default sound classes

If neither --sound-classes-file nor --no-sound-classes are used, the following sound classes are used:

{
  "V": "aeiou",
  "C": "bcdfghjklmnpqrstvwxyz",
  "P": "pbtdkg",
  "F": "fvsz",
  "N": "mn",
  "S": "sz"
}