Merge branch 'tomek' into dev

Tomasz Bartosiak
2 parents 7a9badfd d7dfbb47
Showing 3 changed files with 79 additions and 7 deletions
semantics/phraseology_generator.py
semantics/static/js/semantics_lexical_units.js
semantics/views.py
@@ -20,6 +20,12 @@ def lexicalisation(argument, categories, base):
     elif lexicalisation_type == 'prepnp': #prepnp(prep, case), number, nouns, atr
         prepnps = get_prepnps(get_preposition(lexicalisation_parameters[0]), get_case(lexicalisation_parameters[1], subj), get_number(attributes[1], subj), get_words(attributes[2]), attributes[3])
         return (prepnps, [])
+    elif lexicalisation_type == 'adjp': # adjp(case), number, gender, degree, adjectives, atr
+        adjps = get_adjps(get_case(lexicalisation_parameters[0], subj), get_number(attributes[1], subj), get_gender(attributes[2]), get_degree(attributes[3]), get_words(attributes[4]), attributes[5])
+        return (adjps, get_verb(base, get_number(attributes[1], subj), subj))
+    elif lexicalisation_type == 'prepadjp': #prepadjp(prep, case), number, gender, degree, adjectives, atr
+        prepadjps = get_prepadjps(get_preposition(lexicalisation_parameters[0]), get_case(lexicalisation_parameters[1], subj), get_number(attributes[1], subj), get_gender(attributes[2]), get_degree(attributes[3]), get_words(attributes[4]), attributes[5])
+        return (prepadjps, [])
     else:
         return ([], [])
     return ([], [])
@@ -33,6 +39,9 @@ def is_subj(categories):
 def get_preposition(attribute):
     return attribute.values.all()[0].parameter.type.name
+def get_numerals(attribute):
+    return get_words(attribute)
+
 def get_words(attribute):
     words = [word.text[1:-1] for word in attribute.values.all()]
     return words
@@ -53,6 +62,22 @@ def get_number(attribute, is_subj):
             number = u'sg'
     return number
+def get_gender(attribute):
+    gender = attribute.values.all()[0].parameter.type.name
+    if gender == u'_':
+        gender = u'n'
+    if gender == u'n':
+        gender = u'n1'
+    if gender == u'm':
+        gender = u'm1'
+    return gender
+
+def get_degree(attribute):
+    degree = attribute.values.all()[0].parameter.type.name
+    if degree == u'_':
+        degree = u'pos'
+    return degree
+
 def get_nps(case, number, nouns, _atr):
     result = []
     for noun in nouns:
@@ -61,23 +86,69 @@ def get_nps(case, number, nouns, _atr):
             filtered = []
             for option in options:
                 (orth, tag) = option
-                if case in tag:
+                if u':' + case + u':' in tag:
                     filtered.append(option)
             options = filtered
         if number != u'_':
             filtered = []
             for option in options:
                 (orth, tag) = option
-                if number in tag:
+                if u':' + number + u':' in tag:
                     filtered.append(option)
             options = filtered
-    return [orth for orth, _ in options]
+        result += options
+    return [orth for orth, _ in result]
 def get_prepnps(prep, case, number, nouns, _atr):
     # ala["ma"] = kot
     nps = get_nps(case, number, nouns, _atr)
     return [prep + ' ' + np for np in nps]
+def get_adjps(case, number, gender, degree, adjectives, _atr):
+    result = []
+    for adjective in adjectives:
+        options = [(interp.orth, interp.getTag(MORFEUSZ2)) for interp in MORFEUSZ2.generate(adjective.encode('utf8'))]
+        filtered = []
+        for option in options:
+            (orth, tag) = option
+            if u'adj:' in tag:
+                filtered.append(option)
+        options = filtered
+        if case != u'_':
+            filtered = []
+            for option in options:
+                (orth, tag) = option
+                if u':' + case + u':' in tag:
+                    filtered.append(option)
+            options = filtered
+        if number != u'_':
+            filtered = []
+            for option in options:
+                (orth, tag) = option
+                if u':' + number + u':' in tag:
+                    filtered.append(option)
+            options = filtered
+        if gender != u'_':
+            filtered = []
+            for option in options:
+                (orth, tag) = option
+                if u':' + gender + u':' in tag or  u'.' + gender + u':' in tag or  u':' + gender + u'.' in tag or  u'.' + gender + u'.' in tag:
+                    filtered.append(option)
+            options = filtered
+        if degree != u'_':
+            filtered = []
+            for option in options:
+                (orth, tag) = option
+                if u':' + degree in tag:
+                    filtered.append(option)
+            options = filtered
+        result += options
+    return [orth for orth, _ in result]
+
+def get_prepadjps(prep, case, number, gender, degree, adjectives, _atr):
+     adjps = get_adjps(case, number, gender, degree, adjectives, _atr)
+     return [prep + ' ' + adjp for adjp in adjps]
+
 def get_verb(inf, number, is_subj):
     if not is_subj:
         return None
@@ -86,7 +157,7 @@ def get_verb(inf, number, is_subj):
         filtered = []
         for option in options:
             (orth, tag) = option
-            if u'fin' in tag and u'sg' in tag and u'ter' in tag:
+            if u'fin:' in tag and u':sg:' in tag and u':ter:' in tag:
                 filtered.append(option)
         options = filtered
         return [orth for orth, _ in options]
@@ -331,7 +331,7 @@ function lexicalisationForm(lemma, pre, tokenised) {
     }
     $.ajax({
-        type: "GET",
+        type: "POST",
         dataType: "json",
         url: ajax_get_mwes,
         data: {"options": JSON.stringify(list)},
@@ -646,10 +646,11 @@ def relational_preference_form(request):
     form = RelationalSelPrefForm()
     return {'form': form}
-@ajax(method='get', encode_result=True)         
+@ajax(method='post', encode_result=True)         
 def ajax_get_mwes(request, options):
     results = []
-    for term in options:
+    bases = sorted(list(set(options)))
+    for term in bases:
         term = unicode(term)
         if len(term) > 0:
             obj_results = LexicalUnit.objects.filter(base=term)