How can I match one of two alternate patterns with Rosie Pattern Language?
For example if the string is "(248)" or "(313)" I want to match. Is this possible?
Yes this is possible. Given those two specific patterns the expression would look like this:
eitherAreaCode = "(248)"/"(313)"
This would match either (248) or (313) and nothing else.
Related
For example, my input text are:
You can I talk with someone
on entity I have:
#pron:aboutme = I, Me
#pron:aboutother = someone, anyone, everyone, Richard
So... I want get #pron:aboutother literal
The problem is #pron.literal returns "I" and not "someone"
How can get #pron:aboutother input literal for this case?
#sys-number is a shorthand syntax. In this case, you need to use full syntax <?entities['pron'].get(1).literal?> to get the literal of the second detected entity. It might be good to check if there are two entities of the type detected in the input before (otherwise you get arrayoutofbounds exception).
Trying to do a pretty basic search implementation of partial matching. For instance, I'd like 'ia hu' to return 'Ian Hunter'. I've got first and last name split so we're indexing first, last and combined.
Was reading the suggestion in here, but this just isn't a very elegant or feasible way to solve: https://www.algolia.com/doc/faq/troubleshooting/how-can-i-make-queries-within-the-middle-of-a-word.
I don't think we should have to generate a ton of substring combos for first and last name to get this to return results.
Has anyone implemented a more elegant solution?
In this specific use case (matching "Ian Hunter" with "ia hu"), you can turn prefix matching on all words with queryType=prefixAll (see documentation).
This will not allow infix matching, so "an hu" or "ia un" will not match "Ian Hunter". This cannot therefore be considered a general solution to your question. However, in practice, prefix matching tends to be what people use instinctively; infix matching is relatively rare in my experience.
I'm having a problem with non-latin character sets, and I need to check that a Range is in bounds before performing .substringWithRange. This seems really easy, but I can't find the way to do this.
Given a range:
let lastCharRange = currentString.endIndex.predecessor() ..< currentString.endIndex
How can I check:
let lastExpected = expectedString.substringWithRange(lastCharRange)
is in bounds?
Indexes are tied closely to the String that generated them -- actually to the String's CharacterView, which is a CollectionType. This holds true generally for collections.
So, you simply can't use the Index you got from one String on another String.
Depending on what you are doing, you might have to get a substring from the first and then search the second. You can also get the two Strings' CharacterViews and work with them via their collection-based interface: expectedString.characters.last, for example.
Is it possible to define an extractor that is composed such that a sub-pattern depends on a previously matched sub-pattern?
Consider matching a date pattern, where valid "days" depends on the matched "month".
This is to avoid a guard to compare values bound by the sub-patterns, and also to avoid providing an overly-customized extractor.
Sample syntax:
case r"\d{4}-$month\d{2}-${day filter month.allows}\d{2}" => s"$month $day"
Perhaps you can formulate it under the aegis of this behavior:
https://issues.scala-lang.org/browse/SI-796
That is, before they fix it.
`where client.name.ToLower().Contains(name.ToLower())
Now it's clearer. It's a (badly done) case insensitive search for the name in the client.name. True if name is contained in client.name. Badly done because using international letters (clearly "international letters" don't exist. I mean letters from a culture different from you own. The classical example is the Turkish culture. Read this: http://www.i18nguy.com/unicode/turkish-i18n.html , the part titled Turkish Has An Important Difference), you can break it. The "right" way to do it is: client.name.IndexOf(name, StringComparison.CurrentCultureIgnoreCase) != -1. Instead of StringComparison.CurrentCultureIgnoreCase you can use StringComparison.InvariantCultureIgnoreCase. If you have to use tricks like the ToLower, it has been suggested that it's better to ToUpper both sides of the comparison (but it's MUCH better to use StringComparison.*)
Looks like LINQ to me.
I'm not really up-to-date on .NET these days, but I'd read that as looking for client objects whose name property is a case-insensitive match with the ToString property of the client variable, while allowing additional characters before or after, much like WHERE foo is like '%:some_value%' in SQL. If I'm right, btw, client is a terrible variable name in this instance.
This is a strange piece of code. It would be good to know a bit more about the client object. Essentially it is checking if the case insensitive name value on the client object contains the case insensitive value of the client object (as a string). So if the client name contains the string name of the class itself essentially.
.ToLower() returns the same string you call it on in all lowercase letters. Basically, this statement returns true if name.ToLower() is embedded anywhere within client.name.ToLower().
//If:<br/>
client.name = "nick, bob, jason";
name = "nick";
//Then:<br/>
client.name.ToLower().Contains(name.ToLower());
//would return true