Changed reflex validation rules in semantic validation.

Bartłomiej Nitoń
1 parent 74bcb2d2
Showing 4 changed files with 53 additions and 24 deletions
dictionary/management/commands/get_payments.py
dictionary/models.py
semantics/validation.py
wordnet/models.py
 #-*- coding:utf-8 -*-
 import codecs
+import datetime
 from django.contrib.auth.models import User
 from django.core.management.base import BaseCommand
@@ -14,7 +15,8 @@ class Command(BaseCommand):
         get_payments()
 def get_payments():
-    payments_path = 'data/payments_20150907.csv'
+    now = datetime.datetime.now().strftime('%Y%m%d')
+    payments_path = 'data/payments_%s.csv' % now
     payments_file = codecs.open(payments_path, 'wt', 'utf-8')
     users = User.objects.order_by('username')
@@ -406,9 +406,6 @@ class Frame(Model):
         positions_str_tab = [position['position'].text_rep for position in sorted_positions]
         return u'%s:%s' % (':'.join(frame_chars_str_tab), ' + '.join(positions_str_tab))
-    def has_refl_args(self):
-        return self.positions.filter(arguments__type__in=['refl', 'recip']).exists()
-    
     def __unicode__(self):
         return '%s' % (self.text_rep)
@@ -766,7 +763,10 @@ class Argument(Model):
         return False        
 def sort_arguments(arguments):
-    return sortArguments(arguments)                     
+    return sortArguments(arguments)
+
+def reflex_phrase_types():
+    return ['refl', 'recip']               
 class ArgRealization(Model):
 #    # !NOWE! nie dodalem tego jeszcze na produkcyjnym
 # -*- coding: utf-8 -*-
-from dictionary.models import Lemma
+from django.db.models import Max
+
+from dictionary.models import Lemma, reflex_phrase_types
 from semantics.models import LexicalUnitExamples
 from semantics.utils import get_matching_frame
@@ -9,12 +11,12 @@ def validate_frames(lemma_id):
     actual_frames = lemma.entry_obj.actual_frames()
     error_msg = u''
     for frame in actual_frames.all():
-        error_msg = frame_valid(frame, actual_frames)
+        error_msg = frame_valid(lemma, frame, actual_frames)
         if error_msg:
             break
     return error_msg
-def frame_valid(frame, actual_frames):
+def frame_valid(lemma, frame, actual_frames):
     error_msg = ''
     complements = frame.complements.all()
     if not arguments_exists(complements):
@@ -29,8 +31,8 @@ def frame_valid(frame, actual_frames):
         error_msg = u'Semantyka: Rama semantyczna %d nie ma dopiętych przykładów.' % frame.id
     elif duplicates_exists(frame, actual_frames):
         error_msg = u'Semantyka: Rama semantyczna %d posiada duplikaty.' % frame.id
-    elif not schemas_reflex_agreed(frame):
-        error_msg = u'Semantyka: Rama semantyczna %d ma dopięte schematy o niezgodnej ze znaczeniami zwrotności.' % frame.id
+    elif not schemas_reflex_agreed(lemma, frame):
+        error_msg = u'Semantyka: Rama semantyczna %d ma dopięte elementy o niezgodnej zwrotności.' % frame.id
     elif nonch_pinned(frame):
         error_msg = u'Semantyka: Rama semantyczna %d jest dopięta do typu frazy nonch.' % frame.id
     elif multiplied_same_arg_in_schema(frame):
@@ -88,23 +90,41 @@ def duplicates_exists(frame, actual_frames):
         return True
     return False
-def schemas_reflex_agreed(frame):
+def schemas_reflex_agreed(lemma, frame):
     complements = frame.complements.all()
     lexical_units = frame.lexical_units.all()
-    for compl in complements:
-        for real in compl.realizations.all():
-            schema_reflex = real.frame.get_char_value('ZWROTNOŚĆ').value
-            if (not reflex_in_lexical_units(lexical_units, schema_reflex)):
+    for lex_unit in lexical_units:
+        for schema in lemma.frames.all():
+            if not schema_lex_unit_reflex_agree(lex_unit, schema, complements):
                 return False
     return True
-#def refl_like_phrase_type_pinned(compl, real):
-#    same_frame_realizations = compl.realizations.filter(frame=real.frame,
-#                                                        alternation=real.alternation)
-            
-def reflex_in_lexical_units(lexical_units, reflex):
-    for lex_unit in lexical_units:
-        if lex_unit.is_reflexive() == bool(reflex):
+def schema_lex_unit_reflex_agree(lexical_unit, schema, complements):
+    if complements.filter(realizations__frame=schema).exists():
+        if (not reflex_with_self_mark_agreed(lexical_unit, schema) and 
+            not (lexical_unit.is_reflexive() and not lexical_unit.is_new() and 
+             reflex_with_phrase_types_agreed(lexical_unit, schema, complements))):
+            return False
+    return True
+
+def reflex_with_self_mark_agreed(lexical_unit, schema):
+    schema_self_mark = schema.get_char_value('ZWROTNOŚĆ').value
+    if not lexical_unit.is_reflexive() == bool(schema_self_mark):
+        return False
+    return True
+
+def reflex_with_phrase_types_agreed(lexical_unit, schema, complements):
+    max_alternations = complements.all().aggregate(Max('realizations__alternation'))['realizations__alternation__max']
+    for alternation in range(1, max_alternations+1):
+        if not reflex_with_alternation_phrase_types_agreed(complements, schema, alternation):
+            return False
+    return True
+
+def reflex_with_alternation_phrase_types_agreed(complements, schema, alternation):
+    for compl in complements:
+        if compl.realizations.filter(argument__type__in=reflex_phrase_types(),
+                                     alternation=alternation,
+                                     frame=schema).exists():
             return True
     return False
@@ -155,13 +175,15 @@ def validate_lexical_units(lemma_id):
     lexical_units = lemma.entry_obj.lexical_units()
     for lex_unit in lexical_units.all():
         if not examples_reflex_agreed(lex_unit):
-            error_msg = u'Semantyka: Znaczenie %s ma podpięte przykłady ze schematów o niezgodnej zwrotności.' % unicode(lex_unit)
+            error_msg = u'Semantyka: Znaczenie %s ma podpięte przykłady o niezgodnej zwrotności.' % unicode(lex_unit)
     return error_msg
 def examples_reflex_agreed(lexical_unit):
     lex_examples = LexicalUnitExamples.objects.filter(lexical_unit=lexical_unit)
     for lex_example in lex_examples:
         schema_reflex = lex_example.example.frame.get_char_value('ZWROTNOŚĆ').value
-        if not (lexical_unit.is_reflexive() == bool(schema_reflex)):
+        if (not (lexical_unit.is_reflexive() == bool(schema_reflex)) and 
+            not (lexical_unit.is_reflexive() and not lexical_unit.is_new() and
+                 lex_example.arguments.filter(arguments__type__in=reflex_phrase_types()).exists())):
             return False
     return True
@@ -40,6 +40,11 @@ class LexicalUnit(models.Model):
     def actual_frames(self):
         return self.frames.filter(next__isnull=True, removed=False)
+    def is_new(self):
+        if self.luid < 0:
+            return True
+        return False
+    
     def is_reflexive(self):
         base_parts = self.base.split()
         if len(base_parts) > 1 and base_parts[1] == u'się':