/// <summary>
/// If a matcher in the sequence doesn't match, then it doesn't match
/// </summary>
/// <param name="context"></param>
/// <param name="start"></param>
/// <returns></returns>
public override MatchResult Matches(MatchContext context, TokenEntity startToken, PatternMatcher nextPatternMatcher)
{
var tokenEntity = startToken;
// try to match each element in the sequence.
int start = startToken?.Start ?? 0;
int end = 0;
for (int iPattern = 0; iPattern < PatternMatchers.Count; iPattern++)
{
var matchResult = new MatchResult(false, this, tokenEntity);
var patternMatcher = PatternMatchers[iPattern];
if (patternMatcher.ContainsWildcard())
{
// look ahead to non wild card
nextPatternMatcher = PatternMatchers.Skip(iPattern).Where(pm => !pm.ContainsWildcard()).FirstOrDefault();
// run wildcard pattern matcher
matchResult = patternMatcher.Matches(context, tokenEntity, nextPatternMatcher);
// if the match was not the wildcard pattern, then advance to that.
if (matchResult.NextPatternMatch != null)
{
Debug.Assert(matchResult.NextPatternMatch.PatternMatcher == nextPatternMatcher);
matchResult = matchResult.NextPatternMatch;
iPattern = PatternMatchers.IndexOf(matchResult.PatternMatcher);
Debug.Assert(iPattern >= 0);
}
}
else
{
matchResult = patternMatcher.Matches(context, tokenEntity, nextPatternMatcher);
}
// if the element did not match, then sequence is bad, return failure
if (matchResult.Matched == false)
{
return(new MatchResult(false, this, tokenEntity));
}
tokenEntity = matchResult.NextToken;
end = Math.Max(matchResult.End, end);
}
return(new MatchResult(true, this, tokenEntity, start, end));
}
/// <summary> /// Match and add entity as appropriate. /// </summary> /// <param name="matchContext">match context.</param> /// <param name="tokenEntity">token entity</param> /// <returns>matchresult</returns> public abstract MatchResult Matches(MatchContext matchContext, TokenEntity tokenEntity, PatternMatcher nextPatternMatcher);
/// <summary>
///
/// </summary>
/// <param name="pattern">pattern to parse</param>
/// <returns></returns>
public PatternMatcher Parse(string pattern, bool defaultFuzzyMatch = false, Ordinality ordinality = Ordinality.One, int maxMatches = 16)
{
OrdinalityPatternMatcher ordinalityPatternMatcher = new OrdinalityPatternMatcher(ordinality, maxMatches);
SequencePatternMatcher sequence = new SequencePatternMatcher();
StringBuilder sb = new StringBuilder();
var fuzzyMatch = defaultFuzzyMatch;
var chars = pattern.GetEnumerator();
while (chars.MoveNext())
{
char ch = chars.Current;
bool repeatChar;
do
{
repeatChar = false;
switch (ch)
{
case '(':
Ordinality modifierOrdinality = Ordinality.One;
AddTextToSequence(sequence, sb, fuzzyMatch);
var subText = GetPatternGroup(chars).Trim();
bool inModifiers = true;
while (inModifiers && chars.MoveNext())
{
ch = chars.Current;
switch (ch)
{
case '~':
fuzzyMatch = !defaultFuzzyMatch;
break;
case '?':
modifierOrdinality = Ordinality.ZeroOrOne;
break;
case '+':
modifierOrdinality = Ordinality.OneOrMore;
break;
case '*':
modifierOrdinality = Ordinality.ZeroOrMore;
break;
default:
if (byte.TryParse(ch.ToString(), out byte num))
{
maxMatches = num;
}
else
{
var patternMatcher = Parse(subText, fuzzyMatch, modifierOrdinality, maxMatches);
sequence.PatternMatchers.Add(patternMatcher);
// break out of modifier loop
inModifiers = false;
repeatChar = true;
}
break;
}
}
if (inModifiers)
{
// paren was end of string.
var patternMatcher = Parse(subText, fuzzyMatch, modifierOrdinality, maxMatches);
sequence.PatternMatchers.Add(patternMatcher);
}
maxMatches = 16;
fuzzyMatch = defaultFuzzyMatch;
break;
case '|':
{
AddTextToSequence(sequence, sb, fuzzyMatch);
if (sequence.PatternMatchers.Count == 1)
{
ordinalityPatternMatcher.PatternMatchers.Add(sequence.PatternMatchers.Single());
}
else
{
ordinalityPatternMatcher.PatternMatchers.Add(sequence);
}
sequence = new SequencePatternMatcher();
}
break;
default:
sb.Append(ch);
break;
}
} while (repeatChar);
}
AddTextToSequence(sequence, sb, fuzzyMatch);
if (sequence.PatternMatchers.Any())
{
if (sequence.PatternMatchers.Count == 1)
{
ordinalityPatternMatcher.PatternMatchers.Add(sequence.PatternMatchers.Single());
}
else
{
ordinalityPatternMatcher.PatternMatchers.Add(sequence);
}
}
// if this is a oneOf, maxMatches with a single pattern matcher, just the inner patternmatcher.
PatternMatcher result = ordinalityPatternMatcher;
if (ordinalityPatternMatcher.PatternMatchers.Count == 1 && ordinalityPatternMatcher.Ordinality == Ordinality.One)
{
result = ordinalityPatternMatcher.PatternMatchers.Single();
}
// if it is a sequence with only one patternMatcher, just the inner patternmatcher
if (result is SequencePatternMatcher spm && spm.PatternMatchers.Count == 1)
{
result = spm.PatternMatchers.Single();
}
return(result);
}
/// <summary>
/// If a matcher in the sequence doesn't match, then it doesn't match
/// </summary>
/// <param name="context"></param>
/// <param name="start"></param>
/// <returns></returns>
public override MatchResult Matches(MatchContext context, TokenEntity startToken, PatternMatcher nextPatternMatcher)
{
var tokenEntity = startToken;
int start = startToken?.Start ?? 0;
int end = 0;
int minMatches = 1;
int maxMatches = this.MaxMatches;
switch (Ordinality)
{
case Ordinality.One:
maxMatches = 1;
break;
case Ordinality.OneOrMore:
break;
case Ordinality.ZeroOrMore:
minMatches = 0;
break;
case Ordinality.ZeroOrOne:
minMatches = 0;
maxMatches = 1;
break;
}
int matched = 0;
MatchResult matchResult = null;
bool found;
do
{
found = false;
for (int iPattern = 0; iPattern < PatternMatchers.Count; iPattern++)
{
var patternMatcher = PatternMatchers[iPattern];
var result = patternMatcher.Matches(context, tokenEntity, nextPatternMatcher);
if (result.Matched)
{
found = true;
matched++;
matchResult = result;
end = Math.Max(result.End, end);
tokenEntity = matchResult.NextToken;
if (matched == maxMatches)
{
// then we are done;
return(new MatchResult(true, this, tokenEntity, start, end)
{
NextPatternMatch = result.NextPatternMatch
});
}
break;
}
else
{
if (matchResult != null)
{
matchResult.NextPatternMatch = result.NextPatternMatch;
}
}
}
} while (found && tokenEntity != null);
if (matched < minMatches)
{
// not matched.
return(new MatchResult(false, this)
{
NextPatternMatch = matchResult?.NextPatternMatch
});
}
return(new MatchResult(true, this, tokenEntity, start, end)
{
NextPatternMatch = matchResult?.NextPatternMatch
});
}
public override MatchResult Matches(MatchContext context, TokenEntity startToken, PatternMatcher nextPatternMatcher)
{
var tokenEntity = startToken;
if (tokenEntity != null)
{
if (nextPatternMatcher != null)
{
MatchResult nextPatternMatch = nextPatternMatcher?.Matches(context, tokenEntity, null);
if (nextPatternMatch.Matched && nextPatternMatch.NextToken != tokenEntity)
{
return(new MatchResult(false, this)
{
NextPatternMatch = nextPatternMatch
});
}
}
if (!context.IsTokenMatched(tokenEntity))
{
// if last child is a wildcard and it's end matches the last token's end
// then we will merge the wildcards together.
var previousToken = context.GetPrevTokenEntity(tokenEntity);
var wildcardEntity = context.CurrentEntity.Children.FirstOrDefault(wildcard => wildcard.Type == this.entityType && wildcard.End == previousToken.End);
if (wildcardEntity != null)
{
var newEntity = new LucyEntity()
{
Type = entityType,
Start = wildcardEntity.Start,
End = tokenEntity.End,
Score = ((float)tokenEntity.End - wildcardEntity.Start) / context.Text.Length / 2,
Text = context.Text.Substring(wildcardEntity.Start, tokenEntity.End - wildcardEntity.Start),
Resolution = context.Text.Substring(wildcardEntity.Start, tokenEntity.End - wildcardEntity.Start),
};
// remove old entity
context.CurrentEntity.Children.Remove(wildcardEntity);
// add new merged wildcard entity "joe" "smith" => "joe smith"
context.AddToCurrentEntity(newEntity);
}
else
{
var newEntity = new LucyEntity()
{
Type = entityType,
Start = tokenEntity.Start,
End = tokenEntity.End,
Score = ((float)tokenEntity.End - tokenEntity.Start) / context.Text.Length / 2,
Text = context.Text.Substring(tokenEntity.Start, tokenEntity.End - tokenEntity.Start),
Resolution = context.Text.Substring(tokenEntity.Start, tokenEntity.End - tokenEntity.Start)
};
context.AddToCurrentEntity(newEntity);
}
return(new MatchResult(true, this, context.GetNextTokenEntity(tokenEntity), tokenEntity.Start, tokenEntity.End));
}
}
return(new MatchResult(false, this));
}
public override MatchResult Matches(MatchContext context, TokenEntity startToken, PatternMatcher nextPatterMatcher)
{
var tokenEntity = startToken;
var matchResult = new MatchResult(false, this, tokenEntity);
if (tokenEntity != null)
{
var entity = context.Entities.FirstOrDefault(le => le.Start == tokenEntity.Start && le.Type == EntityType);
if (entity != null)
{
// add the matched entity to the children of the currentEntity.
context.AddToCurrentEntity(entity);
matchResult.Matched = true;
matchResult.End = entity.End;
matchResult.NextToken = context.GetFirstTokenEntity(entity.End);
}
}
return(matchResult);
}