Migrating from anyascii

disarm's transliterate() replaces anyascii for Unicode-to-ASCII conversion.

Quick migration

# Before
from anyascii import anyascii
result = anyascii("café")

# After
from disarm import transliterate
result = transliterate("café")

Or use the compatibility alias:

from disarm import unidecode as anyascii
result = anyascii("café")

API comparison

anyascii disarm Notes
anyascii(s) transliterate(s)
— transliterate(s, lang="de") New: language profiles
— transliterate(s, errors="ignore") New: error modes
— transliterate(s, errors="preserve") New: preserve unmapped

Behavioral differences

Transliteration approach

anyascii and disarm both provide Unicode → ASCII transliteration, but they use different lookup tables. The core Latin-script mappings are very similar, but edge cases may differ. A detailed character-level comparison across all 83 supported languages shows:

  • 49,089 codepoints across all Unicode blocks tested comprehensively (no sampling)
  • 48,415 mapped by disarm vs 48,761 by anyascii — anyascii has broader coverage of some extended script blocks, while disarm provides language-aware romanization with 83 language profiles and 1,136 characters only disarm maps
  • Most differences are systematic: CJK pinyin casing, Korean romanization, and language-specific national standards
from disarm import transliterate

# Common cases — identical
assert anyascii("café") == 'cafe'
assert transliterate("café") == 'cafe'

# CJK — may differ in romanization style
anyascii("北京")       # romanization varies
assert transliterate("北京") == 'bei jing'

Language awareness

anyascii has no language parameter. disarm provides 83 language-specific profiles:

from disarm import transliterate

# anyascii can't do this
assert transliterate("München", lang="de") == 'Muenchen'
assert transliterate("Malmö", lang="sv") == 'Malmoe'

Error handling

anyascii silently drops characters with no mapping. disarm gives you control:

from disarm import transliterate

assert transliterate("♠", errors="replace", replace_with="?") == '?'
assert transliterate("♠", errors="ignore") == ''
assert transliterate("♠", errors="preserve") == '♠'

New features in disarm

Beyond basic transliteration, disarm also provides:

  • slugify() — URL slug generation
  • sanitize_filename() — OS-safe filenames
  • normalize_confusables() — homoglyph normalization
  • TextPipeline — composable text processing
  • strip_accents(), fold_case() — granular text cleaning

All in a single package with one consistent API.