/// <summary>
/// Extracts a block of text delimited by the specified open and close
/// characters. It is assumed the parser is positioned at an
/// occurrence of the open character. The open and closing characters
/// are not included in the returned string. On return, the parser is
/// positioned at the closing character or at the end of the text if
/// the closing character was not found.
/// </summary>
/// <param name="parser">ParsingHelper object</param>
/// <param name="openChar">Start-of-block delimiter</param>
/// <param name="closeChar">End-of-block delimiter</param>
/// <returns>The extracted text</returns>
internal static string ExtractBlock(ParsingHelper parser, char openChar, char closeChar)
{
// Track delimiter depth
int depth = 1;
// Extract characters between delimiters
parser.MoveAhead();
int start = parser.Index;
while (!parser.EndOfText)
{
char ch = parser.Peek();
if (ch == openChar)
{
// Increase block depth
depth++;
}
else if (ch == closeChar)
{
// Decrease block depth
depth--;
// Test for end of block
if (depth == 0)
{
break;
}
}
else if (ch == '"' || ch == '\'')
{
// Don't count delimiters within quoted text
parser.MoveAhead();
parser.SkipWhile(c => c != ch);
}
// Move to next character
parser.MoveAhead();
}
return(parser.Extract(start, parser.Index));
}
/// <summary>
/// Parses a query segment and converts it to an expression
/// tree.
/// </summary>
/// <param name="query">Query segment to be converted.</param>
/// <param name="defaultConjunction">Implicit conjunction type.</param>
/// <returns>Root node of expression tree.</returns>
internal INode?ParseNode(string?query, ConjunctionType defaultConjunction)
{
ConjunctionType conjunction = defaultConjunction;
TermForm termForm = TermForm.Inflectional;
bool termExclude = false;
bool resetState = true;
INode? root = null;
INode? node;
string term;
ParsingHelper parser = new ParsingHelper(query);
while (!parser.EndOfText)
{
if (resetState)
{
// Reset modifiers
conjunction = defaultConjunction;
termForm = TermForm.Inflectional;
termExclude = false;
resetState = false;
}
parser.SkipWhitespace();
if (parser.EndOfText)
{
break;
}
char ch = parser.Peek();
if (Punctuation.Contains(ch))
{
switch (ch)
{
case '"':
case '\'':
termForm = TermForm.Literal;
parser.MoveAhead();
term = parser.ParseWhile(c => c != ch);
root = AddNode(root, term.Trim(), termForm, termExclude, conjunction);
resetState = true;
break;
case '(':
// Parse parentheses block
term = ExtractBlock(parser, '(', ')');
node = ParseNode(term, defaultConjunction);
root = AddNode(root, node, conjunction, true);
resetState = true;
break;
case '<':
// Parse angle brackets block
term = ExtractBlock(parser, '<', '>');
node = ParseNode(term, ConjunctionType.Near);
root = AddNode(root, node, conjunction);
resetState = true;
break;
case '-':
// Match when next term is not present
termExclude = true;
break;
case '+':
// Match next term exactly
termForm = TermForm.Literal;
break;
case '~':
// Match synonyms of next term
termForm = TermForm.Thesaurus;
break;
default:
break;
}
// Advance to next character
parser.MoveAhead();
}
else
{
// Parse this query term
term = parser.ParseWhile(c => !Punctuation.Contains(c) && !char.IsWhiteSpace(c));
// Allow trailing wildcard
if (parser.Peek() == '*')
{
term += parser.Peek();
parser.MoveAhead();
termForm = TermForm.Literal;
}
// Interpret term
StringComparer comparer = StringComparer.OrdinalIgnoreCase;
if (comparer.Compare(term, "AND") == 0)
{
conjunction = ConjunctionType.And;
}
else if (comparer.Compare(term, "OR") == 0)
{
conjunction = ConjunctionType.Or;
}
else if (comparer.Compare(term, "NEAR") == 0)
{
conjunction = ConjunctionType.Near;
}
else if (comparer.Compare(term, "NOT") == 0)
{
termExclude = true;
}
else
{
root = AddNode(root, term, termForm, termExclude, conjunction);
resetState = true;
}
}
}
return(root);
}
