Skip to content

Script, syllables and names

The Odia character inventory, a syllable splitter, and a generator of Odia names. Every example on this page is run by the test suite.

Alphabet

alphabet exposes the Odia Unicode block as plain list objects of single characters: all_letters, vowels, consonants, matras (vowel signs and other combining marks), numbers and punctuations.

>>> from openodia import alphabet
>>> alphabet.vowels
['ଅ', 'ଆ', 'ଇ', 'ଈ', 'ଉ', 'ଊ', 'ଋ', 'ଌ', 'ଏ', 'ଐ', 'ଓ', 'ଔ']
>>> len(alphabet.consonants), len(alphabet.numbers)
(34, 10)
>>> "କ" in alphabet.consonants, "କ" in alphabet.vowels
(True, False)
>>> alphabet.matras[:6]
['ଁ', 'ଂ', 'ଃ', '଼', 'ଽ', 'ା']

Gotcha: matras includes more than vowel signs

The list also holds the candrabindu, anusvara, visarga, nukta, avagraha and virama, plus the ai length mark ୖ, the au length mark ୗ, the isshar ୰, the letter wa ୱ and the fraction sign ୲. It is the set of combining marks and signs, not only the matras of the textbook chart.

Gotcha: the lists are shared

Each attribute is one list object for the whole process. Copy it with list(alphabet.vowels) before you sort or mutate it.

A common use is counting how much of a string is Odia script:

>>> odia = set(alphabet.all_letters)
>>> sum(ch in odia for ch in "ଓଡ଼ିଆ ଭାଷା abc")
9

Syllables

syllable.split breaks a word into orthographic syllables: a consonant with its vowel sign, or a conjunct (consonant + virama + consonant) with its vowel sign. count returns how many there are and hyphenate joins them with a separator.

>>> from openodia import syllable
>>> syllable.split("ନମସ୍କାର")
['ନ', 'ମ', 'ସ୍କା', 'ର']
>>> syllable.split("ପ୍ରଥମ")
['ପ୍ର', 'ଥ', 'ମ']
>>> syllable.count("ବିଦ୍ୟାଳୟ")
4
>>> syllable.hyphenate("ବିଦ୍ୟାଳୟ")
'ବି-ଦ୍ୟା-ଳ-ୟ'
>>> syllable.hyphenate("ଓଡ଼ିଆ ଭାଷା", separator="·")
'ଓ·ଡ଼ି·ଆ ଭା·ଷା'
>>> syllable.split("")
[]

Gotcha: the split is orthographic

split follows the written script, so it does not model schwa deletion or pronunciation. Characters that are not Odia are returned one by one, and spaces are returned as their own item.

Names

name produces random Odia names. Part generators sample without replacement, and asking for more than the list holds returns the whole list.

>>> from openodia import name
>>> len(name.generate_firstnames(count=3))
3
>>> len(name.generate_surnames(count=4))
4
>>> len(name.generate_prefixes(count=2)), len(name.generate_middlenames(count=2))
(2, 2)
>>> set(name.generate_firstnames(count=3, name_type="male")) <= set(name.generate_firstnames(count=10**6, name_type="male"))
True
>>> len(name.generate_names(count=5))
5

name_type accepts "male", "female" or "unisex" (case-insensitive). An empty value mixes all first names. Anything else raises ValueError:

>>> name.generate_firstnames(name_type="robot")
Traceback (most recent call last):
  ...
ValueError: Invalid name_type provided

Gotcha: two sources, two behaviours

The part generators (generate_firstnames, generate_surnames, ...) sample from lists shipped in the package with random.sample. generate_names instead calls the or_IN locale of Faker once per name and does not de-duplicate.

For use cases such as masking and anonymizing names in a text, see Applications.

Next