[Snippet] StringIndexer

[Magtheridon96]

This system generates an index for any string below 8191.

/************************************************
*
*   StringIndexer
*   v1.0.0.2
*   By Magtheridon96
*
*   - Retrieves an index for a string below 8191
*   - Helps easy string data attachment
*
*   Optional:
*   ---------
*
*       - Table by Bribe
*           - hiveworkshop.com/forums/jass-resources-412/snippet-new-table-188084/
*
*   API:
*   ----
*
*       - function IndexString takes string s returns integer
*           - Indexes a string and returns the index
*       - function GetStringId takes string s returns integer
*           - Retrieves the index of an indexed string
*       - function GetStringById takes integer id returns string
*           - Returns the string corresponding to a certain Id
*
************************************************/
library StringIndexer requires optional Table

    globals
        private string array strings
        private integer count = 0
    endglobals
    
    static if LIBRARY_Table then
        private module Init
            private static method onInit takes nothing returns nothing
                set data = Table.create()
            endmethod
        endmodule
    endif

    private struct D extends array
        static if LIBRARY_Table then
            static Table data
            
            implement Init
        else
            static hashtable data = InitHashtable()
        endif
    endstruct
    
    function GetStringById takes integer i returns string
        return strings[i]
    endfunction
    
    function GetStringId takes string t returns integer
        static if LIBRARY_Table then
            return D.data[StringHash(t)]
        else
            return LoadInteger(D.data, StringHash(t), 0)
        endif
    endfunction
    
    function IndexString takes string t returns integer
        local integer i = StringHash(t)
        
        static if LIBRARY_Table then
            if D.data[i] == 0 then
                set count = count + 1
                
                set D.data[i] = count

                set strings[count] = t
                return count
            endif

            return D.data[i]
        else
            if LoadInteger(D.data, i, 0) == 0 then
                set count = count + 1

                call SaveInteger(D.data, i, 0, count)

                set strings[count] = t
                return count
            endif

            return LoadInteger(D.data, i, 0)
        endif
    endfunction
    
endlibrary

I think it's quite nifty.

Feel free to comment.

 

[Troll-Brain]

Is this serious ?

I believe in you enough but i fail to guess a practical use of it.
Care to share one example ?

 

[Switch33]

That is what I thought first when I saw it as well.

This might be useful for my map. For example it could be used for setting up cinematics in-game transmissions text during gameplay where I'd actually need to be using strings. Then again an in-game cinematic maker might be a bit too intense for most people.

 

[LuizBills]

Is this serious ?²

 

[Magtheridon96]

This is more of an interface than a system.
I'm using it for Mode Manager (as soon as I make an update)

Using this would look way better than inlining the whole thing within my code.
Plus, I can totally avoid checks for mode strings that have already been registered by making the struct instance the same as the string id. (Overwriting the data would cause memory leaks (boolexprs))

 

[Bribe]

Makes me wish that Blizzard still allowed use to do the GetStringId return bug. Well you win some (hashtable) you lose some (faux code arrays, B2I, GetStringId). I suppose this thing does a job however limited the job is. It's not a requirement to be completely useful for every system in the world, it's just a requirement that it works and that it can be used for at least 1 thing, otherwise this and 90% of Nestharus resources would be pending forever.

Well it does need to have a module initializer if you are using Table and you're gonna index strings from a module. Other than that I can't see anything wrong with it.

 

[Magtheridon96]

I'm also going to need this for a Chat system ^_^ (Curse word struct for filtering out chat strings, string formatting, etc..)

Basically, the main purpose is 'A struct per string'.

I'll fix that module thing ;p (Thanks for noticing that)

 

[Bribe]

Approved 3/5 (works properly).

 

[Bribe]

I just realized that this has a syntax error without Table in the map -_- "if D.data == 0" will not work if it's a hashtable.

 

[Magtheridon96]

Fixed. Sorry about that Bribe

 

[Troll-Brain]

Just because you use a simple StringHash without any string formatting :

- StringHash("a/b") == StringHash("a\\b")
- StringHash("WhatEVER") == StringHash("whatever")
- maybe other stuff that i'm not aware

Well, it could be easily handled with some string formatting, but i don't think it's worth it, i just wanted to highlight it.

 

[Magtheridon96]

That's Blizzard's fault, not mine

What kind of string formatting are you suggesting though?

 

[Troll-Brain]

That's Blizzard's fault, not mine

I just said how StringHash is case and "slash" insentive, nothing more, nothing less.
In fact here, i think that the insensitive case is more a pro than a con.
And for the slashs, meh who would care about it ?

What kind of string formatting are you suggesting though?

Something like that :

WhAtEver -> *wh*at*ever.

Ofc, the special characters used for the string formatting should be defined in constants.
But it will leak many strings, coz of Substring usage and string concatenations.
Futhermore, as said, i don't think it's needed anyway.

 

[Magtheridon96]

I don't really care about string leaks.
They're like 1 byte per character.
Using 2 different strings is theoretically slower than using 2 identical strings. (Assuming that all these strings have the same size)
This is because I'm pretty sure that a string is only hashed the first time you use it.
The second time would likely only be a check for whether it's hashed or not.

-> String leaks are not a problem.

 

[Troll-Brain]

Well, is there a practical limit size for the internal string table ?
And more strings = slower (i don't think it's O(1), right ? )

But the speed and leaks are not such an issue.
Even if i think that Nestharus did manage to create lag with huge strings leaks, and (heavy ?) strings usage.

My point is that adding this feature would be useless, more, remove a pro, but for the precision sake i had to mentionned it.
Plus, a such thing would be used for chat commands or something, right ?
Who would seriously need a case sensitivity feature ?
Again, i think about it more as a pro than a con.

 

[Magtheridon96]

Not only chat commands, in the future I plan to make an anti-chat spam feature
It stops players from spamming the chat with an extremely fast algorithm that I have recently devised by myself :3
It's not useful for that system in particular, but it's useful for one of the plugins: Anti-cursing

First I'm going to have to make a poll to find the average typing speed of warcraft III players nowadays :3

 

[Adiktuz]

otherwise this and 90% of Nestharus resources would be pending forever.

^_^

Hmmmm... This could be really useful...

 

[Magtheridon96]

If you're working with strings, yes ^.^

 

[mckill2009]

so basically it's just like this and works like convertions?...

local integer i
    call IndexString("Hello I'm Me")    
    call BJDebugMsg(GetStringById(1))
    set i = GetStringId(GetStringById(1))
    call BJDebugMsg(I2S(i))

 

[Magtheridon96]

Think of it as a UnitIndexer, but for strings

You can use it to avoid the mess of using Tables when managing string-based structs.
This way, you can just create a struct based on GetStringId ^.^