[Snowball-discuss] missing conjugations in french

066ce286 at free.fr 066ce286 at free.fr
Thu Dec 19 16:08:41 GMT 2024


Hello,

In french, for example the verb 'parler' (to speak) ar any other 1st group verbs are conjugated as :

[I, you (singular), he/she), we, you (plural), they]
« parle parles parle parlons parlez parlent »

But stemming provinded by snowball is : 

« parl parl parl parlon parl parlent »

I suggest following patch in snowball/algorithms/french.sbl

After regenerating stem_UTF_8_french.c and recompiling dict_snowball.so against postgresql, then I have better stemming :

loxodata_text=# select to_tsvector('french','parle parles parle parlons parlez parlent');
    to_tsvector     
--------------------
 'parl':1,2,3,4,5,6
(1 row)



diff --git a/algorithms/french.sbl b/algorithms/french.sbl
index 793d2f0..eca4d27 100644
--- a/algorithms/french.sbl
+++ b/algorithms/french.sbl
@@ -193,7 +193,7 @@ backwardmode (
 
             '{a^}mes' '{a^}t' '{a^}tes' 'a' 'ai' 'aIent' 'ais' 'ait' 'ant'
             'ante' 'antes' 'ants' 'as' 'asse' 'assent' 'asses' 'assiez'
-            'assions'
+            'assions' 'ent' 'ons'
                 (delete
                  try(['e'] delete)
                 )



More information about the Snowball-discuss mailing list