(moepius) How can I highlight certain keywords in chat but leave the prefix or suffix of the word

[feral needle]

A bit hard to phrase this in english for me, but I hope you'll understand what I'm trying to do.
This script is a part of my chat handler https://paste.denizenscript.com/View/114897. It highlights every keyword I have defined in lists with color or other stuff, if the message contains a keyword.

At the end, I have a choose command where I do specific commands/hovers for different keywords I defined in a list.
The goal is to have words like rules, support, etc. be highlighted in chat and also have a command run when clicked.
Right now this works fine BUT in german you can conjugate words to act as verbs, which lets them end with an "en". So I want to highlight the word "support" when someone writes the word support as is in the chat, but not if it is part of another word that got conjugated or sth. (like in supportING, or supprtEN). This problem also applies to the other tests where I find matches to player names, or server locations.

So TL;DR my keywords should only be highlighted if they begin with a space and end with a space OR if they end with punctuations like ?,:!,etc.
Example:
someword abcdefKEYWORDblabla someword. - should not be highlighted
someword KEYWORD someword - should be highlighted
someword KEYWORD! - should be highlighted (but only the keyword, not the exclamation mark)
someword !KEYWORD someword - should not be highlighted

[rapid iris]

(moepius) How can I highlight certain keywords in chat but leave the prefix or suffix of the word

[rapid iris]

Hi I'm AutoThreadBot! Don't mind me, I'll just be adding the helper team to this thread so they can see it. A human will get to you soon.

[radiant notch]

this might be possible with regex. there is maybe a way to do this nicer with denizen string stuff, but i'm not too sure and it sounds very very complicated to me. someone else could probably find a better way but regex sounds like the best way to me.

this is just a quickly thrown together thing but something like this \b(?<!!|\?)KEYWORD seems to do what you want

again, it could be possible with normal denizen text manipulation but just wanted to give another option

[feral needle]

Oof, I hate regex, but I will try it with that first

And yes, maybe someone has a better solution for this. I will leave this open for now ^^

[chilly marsh]

thats easy

you already looping through a list of words (context.message.split)

so you could just check if the word starts with your words

and if true color it

!t element.starts_with

[runic gazelle]

Tag: <ElementTag.starts_with[<element>]>

Returns whether the element starts with a specified element.

Group

element checking

Returns

ElementTag(Boolean)

[chilly marsh]

you might need an additional check for the length of the element, to make sure it doesn't has a weird 'en' ending

if the word typed in chat is longer than your word + 1, do not highlight it

[feral needle]

With "words" you mean I can check with starts_with if my argument from context.message.split starts with a word from the lists I have defined?

On line 18 with- if <[linkstart].filter_tag[<[arg].starts_with[<[filter_value]>]>].any> || <[linkend].filter_tag[<[arg].ends_with[<[filter_value]>]>].any>: I thought I already did sth. like this, but I can still apply some nonsense before the argument and it will geht highlighted ... like xyzwww.website.com will be underlined when testing.

Hm, and right now I go through each list seperatly, not using context.message.split everywhere

[chilly marsh]

Yeah in that loop you should be able to put your highlight word-logic

[feral needle]

hmm, this works https://paste.denizenscript.com/View/114899

Is this what you meant?

[feral needle]

Maybe I can make it more compact, since I test many different lists on their own and when adding the starts_with bit to every single one, it will get much messier than it already is

[chilly marsh]

You technically don't need that if, if you do just an element.replace

since replace text does simply what it says

basically what youre doing is to check if such a word exists, before you replace that word

[feral needle]

Maybe I can do sth. like

foreach <context.message.split> as:arg:
test if it starts with a word from list1:

[chilly marsh]

if list1 contains arg

or starts with

whatever fits ya logic better

[feral needle]

Woudln't this result in the problem I want to solve? when I have sth. like xyzkeyword and don't want this to be highlighted?

[chilly marsh]

ye

need an additonal check either way

[feral needle]

https://paste.denizenscript.com/View/114900 implemented the starts_with check now ... now I need to add the length one ^^
Can you check if this matches what you intended what I should do?

[chilly marsh]

You have too many filter tags

[chilly marsh]

^ have a simple if

You already filter 6x times for each word a player says

can have a map with the key being the word you'd want to replace

and the value being the value you'd want to replace with

[feral needle]

Okay, not too sure yet on how to use a map for this problem, but I'll look into it ... never worked with maps yet, not gotten the hang of the concept yet ^^

[chilly marsh]

something like

# Script sample
- definemap words:
    Discord: <&color[#5865F2]>Discord<&co> <&n>https://is.gd/cdiscord<&r>
    word: value
- foreach <context.message.split> as:word:
  - foreach <[words].keys> as:replaceable:
    - if <[word].substring[0,<[replaceable].length>]> == <[replaceable]>:
      - Element.replace[<[word].substring[0,<[replaceable].length>]>].with[<[words.<[replaceable]>]>]>
      - foreach stop

that should do?

!t element.substring

[runic gazelle]

Tag: <ElementTag.substring[<#>(,<#>)]>

Returns the portion of an element between two element indices.
If no second index is specified, it will return the portion of an
element after the specified index.
For example: <element[hello].substring[2,4]> returns "ell"

Group

element manipulation

Returns

ElementTag

[feral needle]

Ooh, doesn't look too complicated. Thx

[feral needle]

So in this version I use a map: https://paste.denizenscript.com/View/114918
two things I noticed:

1. The list of keywords will make my map with the current formatting long af, is there some way to make it more compact? I have many words that will result in the same color highlighting, so it would be nice to be able to bundle them together under one key with one value or sth. like that
2. When I test for the word ending with a symbol (- define symbols <list[?|,|:|;|!|.]>), I test if the last letter of the word ends with one of the symbols and then I want the last occurence of that symbol to be replaced with the standard color ... problem is that it currently replaces all occurences of the symbol to be replaced ... has the replace tag some sort of index function where I can determine that only the last found matching symbol is replaced? When I give ".last" a number like ".last[2]" it does the opposite of what I want, returning the first 2, but I want only the last

[chilly marsh]

What's the goal of 2

What's the purpose of line 20

!disablereplies

[runic gazelle]

+++ Info disablereplies: Please disable reply-pings when responding to helpers. This relates to Rule #3: https://denizenscript.com/discord_rules#rule3
https://denizenscript.com/discimg/disable_replies.png

[chilly marsh]

for 1, you could have a data script and store it somewhere else

If you want to keep everything else default color you can use

!t element.color

[runic gazelle]

Tag: <ElementTag.color[<color>]>

Makes the input text colored by the input color. Equivalent to "<COLOR><ELEMENT_HERE><COLOR.end_format>"
Color can be a color name, color code, hex, or ColorTag... that is: ".color[gold]", ".color[6]", and ".color[#AABB00]" are all valid.
Note that this is a magic Denizen tool - refer to !language Denizen Text Formatting.

Group

text manipulation

Returns

ElementTag

[chilly marsh]

!t element.custom_color

[runic gazelle]

Tag: <ElementTag.custom_color[<custom_color_name>]>

Makes the input text colored by the custom color value based on the common base color names defined in the Denizen config file.
If the color name is unrecognized, returns the value of color named 'default'.
Default color names are 'base', 'emphasis', 'warning', 'error'.
Note that this is a magic Denizen tool - refer to !language Denizen Text Formatting.

Group

text manipulation

Returns

ElementTag

[chilly marsh]

that one

so if you have "<green>Hello <element[World].custom_color[emphasis]>! What's up?“

only the word „World“ will be emphasis

everything else stays green

[feral needle]

The goal of 2. is that when I have a keyword like "hello.com" or "word" and it only highlights the keyword itself as intended, I want an exception for symbols that commonly are at the end of a sentence or used without a space like colons. So as result the keyword gets highlighted but the symbol does not. On line 20 I test if the current word ends with one of the symbols I defined in a list.

[feral needle]

Okay, thank you. So then I can still use the logic in my script above but use different key/value pairs to find the words and replace them?

[feral needle]

Used this now in a new version: https://paste.denizenscript.com/View/114964
For links I didn't find a good solution, since I need to test if a word is a link in general (starts with www, etc. or ends with .com, .net, etc.), so I didn't know how to use a map for this to be able to only highlight the link itself and don't highlight symbols like ?,!,:,etc. at the end

This test did not work

- if <[symbols].filter_tag[<[word].ends_with[<[filter_value]>]>].any>: - define textrpl <[textrpl].replace[<[word].to_list.last>].with[<&f><[word].to_list.last><&r>]>

I thought when I only tried to replace the last letter of the word by converting it to a list, it would only replace the color of the last letter in the word, but when it founds a dot in betwenn it also resets

[chilly marsh]

A link doesnt necessarily ends with .de/.at or whatever

[feral needle]

oh, true ... youtube.com/directory

[chilly marsh]

you could them detect the same way as you do with replaceables

if it starts with http:// its a link and ends with space

or https://

for stuff like meta.denizenscript.com or google.de it's more complicated

i can replace meta with whatever i want

[feral needle]

I would have to define what a valid link in general is ... I guess testing for starts_with https://,http:// and www. is good enough ^^

[chilly marsh]

well you could check the top level domain after last dot

if the string after last dot starts with com / de / at its a link

[feral needle]

Oh, yes that should work ... only have to make a list of what top-level domains are valid ^^

or data script

[feral needle]

Can you elaborate on this? I can't use my map to match words and their lengths from the keys

[chilly marsh]

no map use list, no lengthy check, already explained

[feral needle]

would this also fix this problem? I highlight a link when it starts with http...etc. but when the last letter of the link is a sentence symbol, this shouldn't be highlighted

[chilly marsh]

!tias

[runic gazelle]

Info: tias

Try it and see!

If somebody pulled this up for you, you're probably asking a question of the public channel that's easier and faster to figure out by just attempting your idea in-game and looking at the result of that attempt.

[feral needle]

I tried and saw that it didn't work.
But I added a test and some overcomplicated stuff to get what I want with my highlighted links:

https://paste.denizenscript.com/View/114965

Without the redundant use of a flag lol: https://paste.denizenscript.com/View/114966