blmille1 / dtsearchparser Goto Github PK

How can we support unicode characters to be searched?

Hi,

I would like to ask You for a fix with SubExpressions:

So the problem is when I try to parse a regular expression like this: ##(\b[a-z]+[0-9]+\b)
It’s parsed like this: ##

I think it’s because there is a rule which is applied to an expression like this:
(\b[a-z]+[0-9]+\b) which is this expression: '(' expression ')'
That expression is called SubExpression and I don’t think we use it correctly because in that Lucene Query Visitor it is just ignored as I can see.

Instead of ignoring that it would be good to follow an order something like this:

• parentheses
• w/n
• OR
• AND

So I would like to ask help in resolving these issues:

• One is that RegExp expression parsing is not working
  (if we remove the subexpression then it works),
• Other one is that SubExpressions are not handled correctly
  (which signed with parentheses).

I saw an article which describes this situation but it’s slightly different then the one used here, so it would be helpful if someone can show me how can this be applied to the parser's grammar file:
https://stackoverflow.com/questions/5575452/parsing-expressions-with-subexpressions-in-antlr

I found some other missing expressions from that Grammar file which I would like to share with You:

Expressions:

•   RangeStart expression RangeEnd  # RangeExpression
  

•   '<' Exists '>'  # ExistsExpression
  

Tokens:

• Exists : [eE][xX][iI][sS][tT][sS];
• RangeStart : [[{];
• RangeEnd : [\]}];

Kind Regards, Attila

Parsing
( is this ) love

results in
contents:"is this"

It does not throw any error or exception. Can we have the parser throw exception when entire text to be parsed is not consumed properly?

I tried changing grammar as
query : expr;
to
query : expr EOF;

But that didn't help.

test case* written
is expanded as
SpanNear([contents:test, SpanMultiTermQueryWrapper(contents:case*), contents:written], 0, True)

"test case* written" is expanded as
SpanNear([contents:test, contents: , SpanMultiTermQueryWrapper(contents:case*), contents: , contents:written], 0, True)

Are the blank contents: expected?