Alphabetify makes learning new alphabets easy. Copypaste any text – the breaking news, your favourite book, or a piece of homework. Alphabetify will transform the text for you. It starts with all symbols in your native alphabet, then slowly introduces more and more foreign characters. You will get used to them and learn the alphabet without any effort. Try it here!
import { alphabetify } from 'alphabetify'
const text =
'Tell me, O muse, of that ingenious hero who travelled far and wide after he had sacked the famous town of Troy. Many cities did he visit, and many were the nations with whose manners and customs he was acquainted; moreover he suffered much by sea while trying to save his own life and bring his men safely home; but do what he might he could not save his men, for they perished through their own sheer folly in eating the cattle of the Sun-god Hyperion; so the god prevented them from ever reaching home. Tell me, too, about all these things, O daughter of Jove, from whatsoever source you may know them.'
alphabetify(text, 'grek-grc', 'en')
.then(result => console.log(result))
// Tell me, O muse, of thαt ingenious hero who trαvelled fαr ἀnd wiδe ἀfter he ἁδ sαckεδ thε fαmous town of Troy. Mαny citiεs δiδ ἑ visit, ἀnδ mαny wεrε θε nαtions wιθ whosε mαnnεrs ἀnδ κustoms ἑ wαs ἀκquαιntεδ; morεovεr ἑ suffεrεδ muκh βι sεα whιλε trιιγγ to sαvε ἱs owν λιfε ἀνδ βrιγγ ἱs μεν sαfελι ὁμε; βut δο whαt ἑ μιγht hε κοuλδ νοt sαvε hις μεν, fορ θει περισhεδ θροuγh θειρ ὀwν σhεερ fολλι ἰν ἐατιγγ θε καττλε ὀφ θε Σουν-γοδ Hιπεριον; σο θε γοδ πρεουεντεδ θεμ φρομ εουερ ῥεαχιγγ ὁμε. Τελλ με, τοο, ἀβοουτ ἀλλ θεσε θιγγς, O δαυχτερ οφ Dιοουε, φρομ ὀυχατσοεουερ σοουρκε ἰοου μει κνοου θεμ.alphabetify(text, alphabet, [lang, [pre, [post]]])
text
The string of the original text which should be transliterated, in the source alphabet. It may be very long.
alphabet
The code string specifying the target alphabet:
| Code | Alphabet | Quality |
|---|---|---|
cyrl-ru |
Russian | ⭐⭐ |
grek-el |
Modern Greek | ⭐⭐⭐ |
grek-grc |
Ancient Greek | ⭐⭐⭐⭐ |
hira |
Japanese Hiragana | ⭐ |
kana |
Japanese Katakana | ⭐ |
lang optional
The code string specifying the original language. (Eurocentrically, the original alphabet is always Latin.)
Specifying the original language adds some minor language-specific rules. For example, in English the letter v will be processed in a similar way to the letter w, while in German v will be processed in a similar way to f.
If the original language is unspecified, English will be assumed.
| Code | Language |
|---|---|
| de | German |
| en | English |
pre optional
The number 0 ≤ m ≤ (1 - n) specifying the proportion of text at the beginning of the text string which should not be transliterated at all. 0 by default.
post optional
The number 0 ≤ n ≤ (1 - m) specifying the proportion of text at the end of the text string which should be transliterated completely. 0 by default.
A promise, which on resolution returns the string with the increasingly transliterated text. You can process it by appending something like .then(output => process(output)).catch(error => throw error) to the function call.
This module makes use of the fs module, which is available in Node JS, but not in the browser. For usage in the browser, use bundling with Webpack or an equivalent tool and have a look at the configuration in this repository in webpack.config.js, docs/webpack-entry.js and docs/index.js.
Transliteration rules are converted from a short form (e. g., only involving lowercase letters) in alphabets/src/ to a long form in alphabets/build. This is done with the alphabets/preprocess module, which is run by npm run preprocess. The code, including the resulting long form rules, is bundled for web use by running npm run bundle.
This may be a bit abstract. Have a look into some of the JSON files in the alphabets folder to get a better understanding of the notation.
The short form files are found in the folder alphabets/src. If you would like to improve the rule sets, this is the place to look at. The files consist of a specification of the alphabet in regex terms (e. g., a-z for Latin), some optional macros (that is rules to be run on the rules) and of the list of rule blocks. Each rule block consists of rules in the short form.
The short form is a tuple [a, b, lang]:
-
ais a regular expression string,bis a regular expression replacement string. -
langis an optional specifier of the language to which the rule is restricted, such asenorde. -
Only lowercase letters are used in
aandb. Uppercase and mixed-case will be handled automatically. -
The rules will be applied in the order of their length (excluding bracketed parts), starting with the longest rule. It may sometimes be appropriate to give higher priority to certain rules. In this case, underscores
_can be inserted at one's convenience. They will increase the length for the priority sorting and will be ignored afterwards. -
The tuples are grouped into array fields. All tuples of one field will always be applied together. That means: Either none of them will be applied or all of them will be applied.
The long form will be generated automatically from the short form by running npm run preprocess.
The long form is a pair [a, b] where alphabetify will apply text.replace(new RegExp(a, 'g'), b) to the input text for each pair, in the order of their appearance:
-
The pairs are organized in lists, where each element contains all the rules that apply at a transliteration level.
E. g.: There is a text of 100 words. There are 5 (short-form) transliteration rule groups.
-
For the first 20 words, the first element of the (long-form) list will be applied. It contains simply the rules from the first (short-form) rule group.
-
For the next 20 words, the second element of the (long-form) list will be applied. It contains both the rules from the first and the second rule group, in the correct order of their priority.
-
For words 41-60, the third element of the (long-form) list will be applied. It contains rules from the first, second, and third (short-form) rule group.
-
Etc. pp. The crucial point is that at each step where an additional rule group is used, the rules have to be sorted again.
-
-
There is a separate folder for each source language in the
alphabets/buildfolder.
-
The aim of the transliteration is not to achieve phonetic equivalence, but rather to establish a simple set of one-to-one (or a-few-to-one) character matchings: I believe this is the easiest way for the reader to spot patterns and to memorize the new alphabet.
-
However, when there are script-specific concepts which could not be expressed by such simple non-phonetic matchings, phonetics may be taken into account.
E. g.: the ancient greek Omega would probably be omitted in a simple non-phonetic matching, and the English O would be matched with Omikron. But the reader also wants to learn the letter Omega. A replacement rule set which distinguishes Omikron from Omega must take phonetics into account.
Deriving phonetics from written input text works to different degrees for different input languages. While in some rough rules can be established for German, there often appear to be almost no rules for English. That is sad.
-
It is aimed to achieve a balance between keeping the rule set small, and introducing special phonetics-related concepts in a sufficiently accurate way.
The preferred way of contributing is via issues & pull requests in this repo. I also made some Reddit threads to get in feedback from non-coders: