[Zaf-commit] SF.net SVN: zaf:[9642] trunk/dict/zu/hunspell/zu_aff.py

[<fri...@us...>]

Revision: 9642
          http://zaf.svn.sourceforge.net/zaf/?rev=9642&view=rev
Author:   friedelwolff
Date:     2008-12-24 13:40:35 +0000 (Wed, 24 Dec 2008)

Log Message:
-----------
Rewrite verb_rules to accept **kwargs and work with Affix and AffixGroup objects directly

Modified Paths:
--------------
    trunk/dict/zu/hunspell/zu_aff.py

Modified: trunk/dict/zu/hunspell/zu_aff.py
===================================================================
--- trunk/dict/zu/hunspell/zu_aff.py	2008-12-24 11:57:35 UTC (rev 9641)
+++ trunk/dict/zu/hunspell/zu_aff.py	2008-12-24 13:40:35 UTC (rev 9642)
@@ -139,7 +139,7 @@
     prefix = re.sub(r"[aeiou]([aeiou])", r"\1", prefix)
     return prefix
 
-def verb_rules(prefix):
+def verb_rules(group, **kwargs):
     """Generate the necessary rules to prepend the given prefix to a verb
 
     It receives a string with the already built (complete) prefix and 
@@ -147,65 +147,70 @@
     consequence of vowel verbs are taken into account here, and no users of 
     this function need to take vowel verbs into account.
     """
-    changed = []
+    if isinstance(group, basestring):
+        #old API - this is actually the prefix
+        prefix = group
+        #Let's make a temporary group to work with
+        group = AffixGroup(flag="tmp", suffix=False)
+    else:
+        prefix = kwargs.pop('prefix')
     #normal verb starting on consonant:
-    changed.append(["0", prefix, "[^y]"])
-    #monosyllabic verbs: in the dictionary in imperative form, e.g. yidla
-    changed.append(["yi", prefix, "yi"])
+    group.add_rule(affix=prefix, condition="[^aeiou]", **kwargs)
 
-    #now for the complicated part: verbs starting on vowels. (e.g. yakha)
+    #now for the complicated part: verbs starting on vowels. (e.g. akha)
     if prefix[-1] == 'u':
         #the 'u' always needs to be removed, we probably need a 'w'
         prefix = prefix[0:-1]
         #if the original prefix ended on 'wu', we don't want to add 
         #another 'w' as this will result in 'ww'
         if len(prefix) > 0 and prefix[-1] == 'w':
-            changed.append(["y", prefix, "y[ae]"])
+            group.add_rule(affix=prefix, condition="[ae]", **kwargs)
         else:
-            changed.append(["y", prefix + 'w', "y[ae]"])
+            group.add_rule(affix=prefix + 'w', condition="[ae]", **kwargs)
         if len(prefix) == 0:
             #if prefix == 'u' before 'o' we change to 'w' as above
-            changed.append(["y", 'w', "yo"])
+            group.add_rule(affix='w', condition="o", **kwargs)
         else:
             #for a prefix ending on 'u' before 'o' we simply remove the 'u' 
             #without adding 'w', e.g. lu + osa -> losa
-            changed.append(["y", prefix, "yo"])
-        return changed
+            group.add_rule(affix=prefix, condition="o", **kwargs)
+        return group.rules
     if prefix[-1] == 'i':
         #if the complete prefix is 'i' before a vowel verb, we can 
         #ignore it, as it is simply the same as the imperative form
         #example: i + enza = yenza
         #
-        #otherwise, change 'i' to 'y'
         if len(prefix) > 1:
             prefix = prefix[0:-1]
-            changed.append(["y", prefix, "y[^i]"])
-            #changed.append(["y", prefix, "y"])
-            #TODO: this just made the "yi", prefix "yi" rule (above) unnecessary
-        return changed
+            group.add_rule(affix=prefix, condition="[aeiou]", **kwargs)
+        else:
+            #otherwise, change 'i' to 'y'
+            group.add_rule(affix="y", condition="[aeiou]", **kwargs)
+        return group.rules
     if prefix[-1] == 'a':
         prefix = prefix[0:-1]
-        changed.append(["y", prefix, "y[^i]"])
-        #changed.append(["y", prefix, "y"])
-        #TODO: this just made the "yi", prefix "yi" rule (above) unnecessary
-        return changed
+        #prefix could now be "", which is probably ok:
+        # amadoda enza lokhu
+        group.add_rule(affix=prefix, condition="[aeiou]", **kwargs)
+        return group.rules
     if prefix == "o":
         #Although other prefixes can end on 'o' ('zo' or 'yo'), these
         #are only used with consonant verbs and not with monosyllabics.
         #TODO: verify if others are possible
-        #TODO: verify if 'zo' and 'yo' are only used before consonants
 
-        return changed
+        return group.rules
     if prefix[-1] == "e":
         #TODO: situative e- and be- should be handled (entirely thrown away?). 
         #We can probably just generate them, and duplicate handling should take
         #care of them.  Then we need to handle relatives with ending on e.
         pass
-    return changed
+    return group.rules
 
 ################################################################################
 
 for i in concords:
+#    A_rules.extend(verb_rules(i)) #XXX:DONE
+
     A_rules.extend(verb_rules(i+"nga"))
     A_rules.extend(verb_rules(i+"sa"))
 
@@ -242,6 +247,7 @@
 
 #Mode specific ones:
 for i in subject_morphemes:
+#    a_rules.extend(verb_rules(i+"ya"))#XXX:DONE
     #TODO: be- and se- forms
 
     #Negative future tenses:
@@ -328,8 +334,13 @@
 #        flag, cross_product, type = rule_set[0]
 #        group = aff.add_group(suffix=type=="SFX", cross_product=cross_product, flag=flag)
 #        for rule in rule_set[1:]:
-#            strip, affix, condition = rule
-#            group.add_rule(strip=strip, affix=affix, condition=condition)
+#            if isinstance(rule, Affix):
+#                group.add_rule(rule)
+#            else:
+#                strip, affix, condition = rule
+#                group.add_rule(strip=strip, affix=affix, condition=condition)
+#    print aff.hunspell()
+#    exit()
 
     main_subjects = aff.add_group(suffix=False, flag="B")
     # This group will contain all the major subject morphemes


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.