/// <summary>
/// Convert the given digit token into (word) Items in the WordRelation.
/// </summary>
/// <param name="tokenVal">The digit string.</param>
private void DigitsToWords(string tokenVal)
{
FeatureSet featureSet = _tokenItem.Features;
string nsw = "";
if (featureSet.IsPresent("nsw"))
{
nsw = featureSet.GetString("nsw");
}
if (nsw.Equals("nide"))
{
NumberExpander.ExpandId(tokenVal, _wordRelation);
}
else
{
string rName = featureSet.GetString("name");
string digitsType = null;
if (tokenVal.Equals(rName))
{
digitsType = (string)_cart.Interpret(_tokenItem);
}
else
{
featureSet.SetString("name", tokenVal);
digitsType = (string)_cart.Interpret(_tokenItem);
featureSet.SetString("name", rName);
}
if (digitsType.Equals("ordinal"))
{
NumberExpander.ExpandOrdinal(tokenVal, _wordRelation);
}
else if (digitsType.Equals("digits"))
{
NumberExpander.ExpandDigits(tokenVal, _wordRelation);
}
else if (digitsType.Equals("year"))
{
NumberExpander.ExpandId(tokenVal, _wordRelation);
}
else
{
NumberExpander.ExpandNumber(tokenVal, _wordRelation);
}
}
}
/// <summary>
/// process the utterance
/// </summary>
/// <param name="text">The text.</param>
/// <exception cref="IllegalStateException"></exception>
/// <returns>The utterance contain the tokens</returns>
public virtual List <string> Expand(string text)
{
string simplifiedText = SimplifyChars(text);
CharTokenizer tokenizer = new CharTokenizer();
tokenizer.WhitespaceSymbols = UsEnglish.WhitespaceSymbols;
tokenizer.SingleCharSymbols = UsEnglish.SingleCharSymbols;
tokenizer.PrepunctuationSymbols = UsEnglish.PrePunctuationSymbols;
tokenizer.PostpunctuationSymbols = UsEnglish.PunctuationSymbols;
tokenizer.SetInputText(simplifiedText);
Utterance utterance = new Utterance(tokenizer);
Relation tokenRelation;
if ((tokenRelation = utterance.GetRelation(Relation.Token)) == null)
{
throw new IllegalStateException("token relation does not exist");
}
_wordRelation = WordRelation.CreateWordRelation(utterance, this);
for (_tokenItem = tokenRelation.Head; _tokenItem != null; _tokenItem =
_tokenItem.GetNext())
{
FeatureSet featureSet = _tokenItem.Features;
string tokenVal = featureSet.GetString("name");
// convert the token into a list of words
TokenToWords(tokenVal);
}
List <string> words = new List <string>();
for (Item item = utterance.GetRelation(Relation.Word).Head; item != null; item =
item.GetNext())
{
if (!string.IsNullOrEmpty(item.ToString()) && !item.ToString().Contains("#"))
{
words.Add(item.ToString());
}
}
return(words);
}
/// <summary>
/// Returns true if the given token is the name of a US state. If it is, it
/// will add the name of the state to (word) Items in the WordRelation.
/// </summary>
/// <param name="tokenVal">The token string.</param>
private bool IsStateName([In] string tokenVal)
{
string[] state = UsStatesMap.Get(tokenVal);
if (state != null)
{
bool expandState = false;
// check to see if the state initials are ambiguous
// in the English language
if (state[1].Equals("ambiguous"))
{
string previous = (string)_tokenItem.FindFeature("p.name");
string next = (string)_tokenItem.FindFeature("n.name");
int nextLength = next.Length;
FeatureSet featureSet = _tokenItem.Features;
// check if the previous word starts with a capital letter,
// is at least 3 letters long, is an alphabet sequence,
// and has a comma.
bool previousIsCity =
(char.IsUpper(previous[0]) &&
previous.Length > 2 &&
Matches(AlphabetPattern, previous) && _tokenItem
.FindFeature("p.punc").Equals(","));
// check if next token starts with a lower case, or
// this is the end of sentence, or if next token
// is a period (".") or a zip code (5 or 10 digits).
bool nextIsGood =
(char.IsLower(next[0])) ||
_tokenItem.GetNext() == null ||
featureSet.GetString("punc").Equals(".") || ((nextLength == 5 || nextLength == 10) && Matches(
DigitsPattern, next));
if (previousIsCity && nextIsGood)
{
expandState = true;
}
else
{
expandState = false;
}
}
else
{
expandState = true;
}
if (expandState)
{
for (int j = 2; j < state.Length; j++)
{
if (state[j] != null)
{
_wordRelation.AddWord(state[j]);
}
}
return(true);
}
}
return(false);
}
/// <summary>
/// Converts the given string containing "St" and "Dr" to (word) Items in the WordRelation.
/// </summary>
/// <param name="drStString">The string with "St" and "Dr".</param>
private void DrStToWords(string drStString)
{
string street = null;
string saint = null;
char c0 = drStString[0];
if (c0 == 's' || c0 == 'S')
{
street = "street";
saint = "saint";
}
else
{
street = "drive";
saint = "doctor";
}
FeatureSet featureSet = _tokenItem.Features;
string punctuation = featureSet.GetString("punc");
string featPunctuation = (string)_tokenItem.FindFeature("punc");
if (_tokenItem.GetNext() == null || punctuation.IndexOf(',') != -1)
{
_wordRelation.AddWord(street);
}
else if (featPunctuation.Equals(","))
{
_wordRelation.AddWord(saint);
}
else
{
string pName = (string)_tokenItem.FindFeature("p.name");
string nName = (string)_tokenItem.FindFeature("n.name");
char p0 = pName[0];
char n0 = nName[0];
if (char.IsUpper(p0) && char.IsLower(n0))
{
_wordRelation.AddWord(street);
}
else if (char.IsDigit(p0) && char.IsLower(n0))
{
_wordRelation.AddWord(street);
}
else if (char.IsLower(p0) && char.IsUpper(n0))
{
_wordRelation.AddWord(saint);
}
else
{
string whitespace = (string)_tokenItem.FindFeature("n.whitespace");
if (whitespace.Equals(" "))
{
_wordRelation.AddWord(saint);
}
else
{
_wordRelation.AddWord(street);
}
}
}
if (punctuation != null && punctuation.Equals("."))
{
featureSet.SetString("punc", "");
}
}
/// <summary>
/// Converts the given Token into (word) Items in the WordRelation.
/// </summary>
/// <param name="tokenVal">the string value of the token, which may or may not be
/// same as the one in called "name" in flite</param>
private void TokenToWords(string tokenVal)
{
FeatureSet tokenFeatures = _tokenItem.Features;
string itemName = tokenFeatures.GetString("name");
int tokenLength = tokenVal.Length;
if (tokenFeatures.IsPresent("phones"))
{
_wordRelation.AddWord(tokenVal);
}
else if ((tokenVal.Equals("a") || tokenVal.Equals("A")) &&
((_tokenItem.GetNext() == null) ||
!(tokenVal.Equals(itemName)) || !(((string)_tokenItem
.FindFeature("punc")).Equals(""))))
{
/* if A is a sub part of a token, then its ey not ah */
_wordRelation.AddWord("_a");
}
else if (Matches(AlphabetPattern, tokenVal))
{
if (Matches(RomanNumbersPattern, tokenVal))
{
/* XVIII */
RomanToWords(tokenVal);
}
else if (Matches(IllionPattern, tokenVal) &&
Matches(UsMoneyPattern,
(string)_tokenItem.FindFeature("p.name")))
{
/* $ X -illion */
_wordRelation.AddWord(tokenVal);
_wordRelation.AddWord("dollars");
}
else if (Matches(DrStPattern, tokenVal))
{
/* St Andrew's St, Dr King Dr */
DrStToWords(tokenVal);
}
else if (tokenVal.Equals("Mr"))
{
_tokenItem.Features.SetString("punc", "");
_wordRelation.AddWord("mister");
}
else if (tokenVal.Equals("Mrs"))
{
_tokenItem.Features.SetString("punc", "");
_wordRelation.AddWord("missus");
}
else if (tokenLength == 1 &&
char.IsUpper(tokenVal[0]) &&
((string)_tokenItem.FindFeature("n.whitespace"))
.Equals(" ") &&
char.IsUpper(((string)_tokenItem
.FindFeature("n.name"))[0]))
{
tokenFeatures.SetString("punc", "");
string aaa = tokenVal.ToLower();
if (aaa.Equals("a"))
{
_wordRelation.AddWord("_a");
}
else
{
_wordRelation.AddWord(aaa);
}
}
else if (IsStateName(tokenVal))
{
/*
* The name of a US state isStateName() has already added the
* full name of the state, so we're all set.
*/
}
else if (tokenLength > 1 && !IsPronounceable(tokenVal))
{
/* Need common exception list */
/* unpronouncable list of alphas */
NumberExpander.ExpandLetters(tokenVal, _wordRelation);
}
else
{
/* just a word */
_wordRelation.AddWord(tokenVal.ToLower());
}
}
else if (Matches(DottedAbbrevPattern, tokenVal))
{
/* U.S.A. */
// remove all dots
NumberExpander.ExpandLetters(tokenVal.Replace(".", ""),
_wordRelation);
}
else if (Matches(CommaIntPattern, tokenVal))
{
/* 99,999,999 */
NumberExpander.ExpandReal(tokenVal.Replace(",", "").Replace("'", ""), _wordRelation);
}
else if (Matches(SevenPhoneNumberPattern, tokenVal))
{
/* 234-3434 telephone numbers */
int dashIndex = tokenVal.IndexOf('-');
string aaa = tokenVal.JSubString(0, dashIndex);
string bbb = tokenVal.Substring(dashIndex + 1);
NumberExpander.ExpandDigits(aaa, _wordRelation);
_wordRelation.AddBreak();
NumberExpander.ExpandDigits(bbb, _wordRelation);
}
else if (MatchesPartPhoneNumber(tokenVal))
{
/* part of a telephone number */
var punctuation = (string)_tokenItem.FindFeature("punc");
if (punctuation.Equals(""))
{
_tokenItem.Features.SetString("punc", ",");
}
NumberExpander.ExpandDigits(tokenVal, _wordRelation);
_wordRelation.AddBreak();
}
else if (Matches(NumberTimePattern, tokenVal))
{
/* 12:35 */
int colonIndex = tokenVal.IndexOf(':');
string aaa = tokenVal.JSubString(0, colonIndex);
string bbb = tokenVal.Substring(colonIndex + 1);
NumberExpander.ExpandNumber(aaa, _wordRelation);
if (!(bbb.Equals("00")))
{
NumberExpander.ExpandId(bbb, _wordRelation);
}
}
else if (Matches(Digits2DashPattern, tokenVal))
{
/* 999-999-999 */
DigitsDashToWords(tokenVal);
}
else if (Matches(DigitsPattern, tokenVal))
{
DigitsToWords(tokenVal);
}
else if (tokenLength == 1 &&
char.IsUpper(tokenVal[0]) &&
((string)_tokenItem.FindFeature("n.whitespace"))
.Equals(" ") &&
char.IsUpper(((string)_tokenItem
.FindFeature("n.name"))[0]))
{
tokenFeatures.SetString("punc", "");
string aaa = tokenVal.ToLower();
if (aaa.Equals("a"))
{
_wordRelation.AddWord("_a");
}
else
{
_wordRelation.AddWord(aaa);
}
}
else if (Matches(DoublePattern, tokenVal))
{
NumberExpander.ExpandReal(tokenVal, _wordRelation);
}
else if (Matches(OrdinalPattern, tokenVal))
{
/* explicit ordinals */
string aaa = tokenVal.JSubString(0, tokenLength - 2);
NumberExpander.ExpandOrdinal(aaa, _wordRelation);
}
else if (Matches(UsMoneyPattern, tokenVal))
{
/* US money */
UsMoneyToWords(tokenVal);
}
else if (tokenLength > 0 && tokenVal[tokenLength - 1] == '%')
{
/* Y% */
TokenToWords(tokenVal.JSubString(0, tokenLength - 1));
_wordRelation.AddWord("percent");
}
else if (Matches(NumessPattern, tokenVal))
{
NumberExpander.ExpandNumess(tokenVal.JSubString(0, tokenLength - 1), _wordRelation);
}
else if (Matches(DigitsSlashDigitsPattern, tokenVal) && tokenVal.Equals(itemName))
{
DigitsSlashDigitsToWords(tokenVal);
}
else if (tokenVal.IndexOf('-') != -1)
{
DashToWords(tokenVal);
}
else if (tokenLength > 1 && !Matches(AlphabetPattern, tokenVal))
{
NotJustAlphasToWords(tokenVal);
}
else if (tokenVal.Equals("&"))
{
// &
_wordRelation.AddWord("and");
}
else if (tokenVal.Equals("-"))
{
// Skip it
}
else
{
// Just a word.
_wordRelation.AddWord(tokenVal.ToLower());
}
}