1+ /*
2+ * The contents of this file are subject to the Mozilla Public License
3+ * Version 1.1 (the "License"); you may not use this file except in
4+ * compliance with the License. You may obtain a copy of the License at
5+ * http://www.mozilla.org/MPL/
6+ *
7+ * Software distributed under the License is distributed on an "AS IS"
8+ * basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See the
9+ * License for the specific language governing rights and limitations
10+ * under the License.
11+ *
12+ * The Original Code is "SharpSimpleNLG".
13+ *
14+ * The Initial Developer of the Original Code is Ehud Reiter, Albert Gatt and Dave Westwater.
15+ * Portions created by Ehud Reiter, Albert Gatt and Dave Westwater are Copyright (C) 2010-11 The University of Aberdeen. All Rights Reserved.
16+ *
17+ * Contributor(s): Ehud Reiter, Albert Gatt, Dave Wewstwater, Roman Kutlak, Margaret Mitchell, Saad Mahamood, Nick Hodge
18+ */
19+
20+ /* Additional Notes:
21+ * - Original Java source is SimpleNLG from 12-Jun-2016 https://github.com/simplenlg/simplenlg
22+ * - This is a port of the Java version to C# with no additional features
23+ * - I have left the "Initial Developer" section to reflect this fact
24+ * - Any questions, comments, feedback on this port can be sent to Nick Hodge <[email protected] > 25+ */
26+
27+ using System . Text ;
28+ using SimpleNLG . Extensions ;
29+
30+ namespace SharpNLG . Extensions
31+ {
32+ /**
33+ * This class is used to parse numbers that are passed as figures, to determine
34+ * whether they should take "a" or "an" as determiner.
35+ *
36+ * @author bertugatt
37+ *
38+ */
39+
40+ public class DeterminerAgrHelper
41+ {
42+ /*
43+ * An array of strings which are exceptions to the rule that "an" comes
44+ * before vowels
45+ */
46+ private static string [ ] AN_EXCEPTIONS = { "one" , "180" , "110" } ;
47+
48+ /*
49+ * Start of string involving vowels, for use of "an"
50+ */
51+ private static string AN_AGREEMENT = @"\A(a|e|i|o|u).*" ;
52+
53+ /*
54+ * Start of string involving numbers, for use of "an" -- courtesy of Chris
55+ * Howell, Agfa healthcare corporation
56+ */
57+ // private static final string AN_NUMERAL_AGREEMENT =
58+ // "^(((8((\\d+)|(\\d+(\\.|,)\\d+))?).*)|((11|18)(\\d{3,}|\\D)).*)$";
59+
60+ /**
61+ * Check whether this string starts with a number that needs "an" (e.g.
62+ * "an 18% increase")
63+ *
64+ * @param string
65+ * the string
66+ * @return <code>true</code> if this string starts with 11, 18, or 8,
67+ * excluding strings that start with 180 or 110
68+ */
69+
70+ public static bool requiresAn ( string stringa )
71+ {
72+ var req = false ;
73+
74+ var lowercaseInput = stringa . toLowerCase ( ) ;
75+
76+ if ( lowercaseInput . matches ( AN_AGREEMENT ) && ! isAnException ( lowercaseInput ) )
77+ {
78+ req = true ;
79+
80+ }
81+ else
82+ {
83+ var numPref = getNumericPrefix ( lowercaseInput ) ;
84+
85+ if ( numPref != null && numPref . length ( ) > 0
86+ && numPref . matches ( @"^(8|11|18).*$" ) )
87+ {
88+ var num = int . Parse ( numPref ) ;
89+ req = checkNum ( num ) ;
90+ }
91+ }
92+
93+ return req ;
94+ }
95+
96+ /*
97+ * check whether a string beginning with a vowel is an exception and doesn't
98+ * take "an" (e.g. "a one percent change")
99+ *
100+ * @return
101+ */
102+
103+ private static bool isAnException ( string stringa )
104+ {
105+ foreach ( var ex in AN_EXCEPTIONS )
106+ {
107+ if ( stringa . matches ( "^" + ex + ".*" ) )
108+ {
109+ // if (string.equalsIgnoreCase(ex)) {
110+ return true ;
111+ }
112+ }
113+
114+ return false ;
115+ }
116+
117+ /*
118+ * Returns <code>true</code> if the number starts with 8, 11 or 18 and is
119+ * either less than 100 or greater than 1000, but excluding 180,000 etc.
120+ */
121+
122+ private static bool checkNum ( int num )
123+ {
124+ var needsAn = false ;
125+
126+ // eight, eleven, eighty and eighteen
127+ if ( num == 11 || num == 18 || num == 8 || ( num >= 80 && num < 90 ) )
128+ {
129+ needsAn = true ;
130+
131+ }
132+ else if ( num > 1000 )
133+ {
134+ // num = Math.Round(num / 1000);
135+ num = num / 1000 ;
136+ needsAn = checkNum ( num ) ;
137+ }
138+
139+ return needsAn ;
140+ }
141+
142+ /*
143+ * Retrieve the numeral prefix of a string.
144+ */
145+
146+ private static string getNumericPrefix ( string stringa )
147+ {
148+ var numeric = new StringBuilder ( ) ;
149+
150+ if ( stringa != null )
151+ {
152+ stringa = stringa . Trim ( ) ;
153+
154+ if ( stringa . length ( ) > 0 )
155+ {
156+
157+ var buffer = new StringBuilder ( stringa ) ;
158+ var first = buffer . charAt ( 0 ) ;
159+
160+ if ( first . isDigit ( ) )
161+ {
162+ numeric . append ( first ) ;
163+
164+ for ( var i = 1 ; i < buffer . length ( ) ; i ++ )
165+ {
166+ var next = buffer . charAt ( i ) ;
167+
168+ if ( next . isDigit ( ) )
169+ {
170+ numeric . append ( next ) ;
171+
172+ // skip commas within numbers
173+ }
174+ else if ( next . Equals ( ',' ) )
175+ {
176+ continue ;
177+
178+ }
179+ else
180+ {
181+ break ;
182+ }
183+ }
184+ }
185+ }
186+ }
187+
188+ return numeric . length ( ) == 0 ? null : numeric . ToString ( ) ;
189+ }
190+
191+
192+ /**
193+ * Check to see if a string ends with the indefinite article "a" and it agrees with {@code np}.
194+ * @param text
195+ * @param np
196+ * @return an altered version of {@code text} to use "an" if it agrees with {@code np}, the original string otherwise.
197+ */
198+
199+ public static string checkEndsWithIndefiniteArticle ( string text , string np )
200+ {
201+
202+ var tokens = text . Split ( ' ' ) ;
203+
204+ var lastToken = tokens [ tokens . Length - 1 ] ;
205+
206+ if ( lastToken . equalsIgnoreCase ( "a" ) && DeterminerAgrHelper . requiresAn ( np ) )
207+ {
208+
209+ tokens [ tokens . Length - 1 ] = "an" ;
210+
211+ return stringArrayToString ( tokens ) ;
212+
213+ }
214+
215+ return text ;
216+
217+ }
218+
219+ // Turns ["a","b","c"] into "a b c"
220+ private static string stringArrayToString ( string [ ] sArray )
221+ {
222+
223+ var buf = new StringBuilder ( ) ;
224+
225+ for ( var i = 0 ; i < sArray . Length ; i ++ )
226+ {
227+
228+ buf . Append ( sArray [ i ] ) ;
229+
230+ if ( i != sArray . Length - 1 )
231+ {
232+
233+ buf . Append ( " " ) ;
234+
235+ }
236+
237+ }
238+
239+ return buf . ToString ( ) ;
240+
241+ }
242+
243+ }
244+ }
0 commit comments