Sonnex is an alternative to Soundex for french language.

This package implements an alternative to the Soundex algorithms for french language. It does so by approximating what the word should sound in french. Since it is very basic, it has no other dependencies than base.

Sonnex

Sonnex is an alternative to Soundex for french language

The string must contain only one word. The Sonnex code contains the following characters:

• 1 ← un, ein, in, ain
• 2 ← en, an
• 3 ← on
• a ← a, à, â
• b ← b, bb
• C ← ch
• d ← d, dd
• e ← e, eu
• E ← ê, é, è, ai, ei
• f ← f, ff, ph
• g ← gu
• i ← î, i, ille
• j ← j, ge
• k ← k, c, qu, ck
• l ← l, ll
• m ← m, mm
• n ← n, nn
• o ← o, ô
• p ← p, pp
• r ← r, rr
• s ← s, ss
• t ← t, tt
• u ← u, ù, û
• v ← v, w
• z ← z, s
• U ← ou

Examples

Here are a few examples of sonnex results:

• balade | ballade → balad
• basilic | basilique → bazilik
• boulot | bouleau → bUlo
• cane | canne → kan
• censé | sensé → s2sé
• compte | comte | conte → k3t
• cygne | signe → sin
• date | datte → dat
• dessin | dessein → dés1
• différend | différent → difér2
• cric | crique → krik
• champ | chant → C2

Testing

The test directory contains two files:

• test-sonnex-homonymes.txt: a list of french homonyms
• test-sonnex-streets.txt: a list of streets from Rouen (Normandy)

You can try these files against Sonnex by using the following command:

cat test-sonnex-homonymes.txt | runhaskell test-sonnex.hs.