Update propbank_analyzer to adjust searching operations

lib/propbank_analyzer.py

@@ -13,6 +13,7 @@
 
 import sys
 import glob
+import re
 
 from bs4 import BeautifulSoup
 
@@ -28,7 +29,7 @@ OUTPUT_DIR = "../outputData/"
 # Data
 PROPBANK_FRAMES_DIR = "../propbankFrames/"
 PBF_DIGITS = 2
-
+AMR_CORE_ROLE_FORM = [':ARG\d$', 'ARG\d$', '\d$']
 
 
 #==============================================================================
@@ -64,7 +65,11 @@ def get_roleset_id_from_amr_predicate(amr_predicate):
 
 
 def get_number_from_amr_role(amr_role):
-    return 1
+    role_number = -1
+    for role_format in AMR_CORE_ROLE_FORM:
+        if re.match(role_format, amr_role):
+            role_number = int(amr_role[-1])
+    return role_number
 
 
 #==============================================================================
@@ -79,16 +84,16 @@ def find_frame_of_lemma(lemma):
     frame_filepath = glob.glob(target_file, recursive=True)
     
     if len(frame_filepath) >= 1:
-        is_found = True
         frame_filepath = frame_filepath[0]
         with open(frame_filepath, 'r') as f:
             xml_data = f.read()
             frame_data = BeautifulSoup(xml_data, 'xml')
     else:
-        is_found = False
         frame_filepath = ''
         frame_data = None
     
+    is_found = frame_data is not None
+    
     return is_found, frame_filepath, frame_data
 
 
@@ -96,35 +101,36 @@ def find_frame_of_lemma(lemma):
 # Functions to analyze a frame data
 #==============================================================================
 
-def get_roleset_in_frame(frame_data, lemma, roleset_id):
-    """ Get a roleset in a given frame data
+def find_roleset_in_frame(frame_data, lemma, roleset_id):
+    """ Find the roleset corresponding to a lemma and an id in a frame data
     """
       
     try:
         lemma_data = frame_data.find('predicate', {'lemma':lemma})
         roleset_data = lemma_data.find('roleset', {'id':roleset_id})
-        is_found = True
     
     except:
+        lemma_data = None
         roleset_data = None
-        is_found = False
+         
+    is_found = (lemma_data is not None) & (roleset_data is not None)
     
     return is_found, roleset_data
 
 
-def get_role_in_roleset(roleset_data, role_number):
+def find_role_in_roleset(roleset_data, role_number):
+    """ Find the role corresponding to a given number in a roleset data
+    """
   
     try:
         role_data = roleset_data.find('role', {'n':role_number})
-        is_found = True
 
     except:
-        roleset_data = None
-        is_found = False
-
-    return is_found, role_data
+        role_data = None
 
+    is_found = (role_data is not None)
 
+    return is_found, role_data
 
     
 
@@ -152,30 +158,37 @@ def main(amr_predicate, amr_role):
     if frame_found:
         print("----- frame xml file found: " + frame_filepath)
     else:
-        print("----- frame xml file not found")
+        print("----- frame xml file not found for lemma " + lemma)
+           
+    if frame_found:      
+        # -- Analyze frame data to get informations
+        print("-- Analyzing frame data")
+        rs_found, rs_data = find_roleset_in_frame(frame_data, lemma, roleset_id)
+        nb_roles = -1
         
-    # -- Analyze frame data to get informations
-    print("-- Analyzing frame data")
-    rs_found, rs_data = get_roleset_in_frame(frame_data, lemma, roleset_id)
-    nb_roles = -1
-    if rs_found:
-        print("----- roleset id: " + rs_data.get('id'))
-        print("----- roleset name: " + rs_data.get('name'))
-        nb_roles = len(rs_data.find_all('role'))
-        print("----- number of roles: " + str(nb_roles))
-        for n in range(nb_roles):
-            _, role_data = get_role_in_roleset(rs_data, n)
-            print("----- role " + str(n) + ': ' + role_data.get('f') + 
+        if rs_found:
+            print("----- roleset id: " + rs_data.get('id'))
+            print("----- roleset name: " + rs_data.get('name'))
+            nb_roles = len(rs_data.find_all('role'))
+            print("----- number of roles: " + str(nb_roles))
+            for n in range(nb_roles):
+                _, role_data = find_role_in_roleset(rs_data, n)
+                print("----- role " + str(n) + ': ' + role_data.get('f') + 
                                                  ', ' + role_data.get('descr'))
-    else:
-        print("----- roleset not found")
+        else:
+            print("----- roleset " + roleset_id + " not found")
         
-    # -- Analyze frame data to get informations
-    if rs_found & role_number in range(nb_roles):
-        print("-- Finding role")
-        _, role_data = get_role_in_roleset(rs_data, role_number)
-        print("----- role found: " + role_data.get('f') + 
-                                             ', ' + role_data.get('descr'))
+        # -- Analyze frame data to get informations
+        if rs_found & role_number in range(nb_roles):
+            print("-- Finding role")
+            print("----- role number: " + str(role_number))
+            r_found, role_data = find_role_in_roleset(rs_data, role_number)
+            if r_found:
+                print("----- role " + str(role_number) + " found: " + 
+                      role_data.get('f') + 
+                      ', ' + role_data.get('descr'))
+            else:
+                print("----- role " + str(role_number) + " not found")    
     
     # -- Ending print
     print("\n" + "[SSC] Done")

requirements.txt

@@ -4,3 +4,5 @@ argparse
 numpy
 rdflib
 graphviz
+bs4
+lxml