Added card `matcher` functions and related. by AllUniversal · Pull Request #1299 · Steamodded/smods

AllUniversal · 2026-04-02T20:06:59Z

+Title, I moved this over from #838, the latest two comments on that PR explain card matchers.

Additional Info:

I didn't modify api's or I've made a PR to the wiki repo.
I didn't modify api's or I've updated lsp definitions.
I didn't make new lovely files or all new lovely files have appropriate priority.

+Title

*/-Title, also removed some minor features, will re-add once Steamodded#838 is merged. :)

…d stuff. +/*Title, the new subcondition `count` allows for matching according to the total number of cards that share a property, for example; A matcher that matches any card, as long as its seal appears at least twice. +Also added `Card:get_suits()`, as I had a need for it.

AllUniversal · 2026-04-03T00:49:29Z

I now added a new subcondition count. It should make it possible to match for cards based on how many cards share a property, like matching a card only if its seal appears exactly once, or at_least twice, or at_most (=> subflags of count) three times, or if a custom func (subflag of count) returns true when called with the total number of the card's property.

EDIT: Oh, and just to mention it here, I also added Card:get_suits().

+/*Title

+Title

AllUniversal · 2026-04-03T02:17:22Z

Realized stuff ain't working right, will look at it tomorrow.

*Title, the `count` subcondition is now evaluated at the same time as the others. -> This was necessary for it to not mess up if `condition.invert` was true. (Or in general)

AllUniversal · 2026-04-03T09:58:52Z

Fixed (and tested) it, .count should now be working!

*Title, modified an early return condition.

*Title, realized I didn't like an early initilization of a map.

+Title

*Title, nuff said.

*/-Title, removed relic.

*Title, forgor to put `check_function` into the `SMODS.insert_card_matcher_condition()` whitelist

Aurelius7309 · 2026-04-03T10:41:31Z

+    return matchers_met_cards, cards_met_matchers
+end
+
+function SMODS.get_hand_from_matching(matchers_to_cards, cards_to_matchers, deduplicate_matches, all_matched_cards_score)


style: this has too many arguments. table args would be preferred

Currently you can call it like this;

The first to params of SMODS.get_hand_from_matching() match the return values from SMODS.match_cards(),
should I turn the two boolean params into one args param?

My bad, I phrased that sloppily. That's exactly what I mean.

…ent) as `SMODS.card_matcher_nil_sentinel` *Title

*Title

Aurelius7309 · 2026-04-03T21:27:32Z

Ran into some hard crashes while testing it out, so I just went ahead and fixed those ^^
I'm still not close to having fully tested it, but I did notice this not working as expected:
SMODS.create_card_matcher { enhancement = { any = { m_gold = true, m_steel = true }, count = { at_least = 3 } } } - in fact I couldn't get it to fire at all. Is this user error on my part or a bug?

AllUniversal · 2026-04-03T21:40:45Z

@Aurelius7309 The following code works for me;

Specifically, this matcher would match a card if;

It is either a Gold or Steel card
Its enhancement (no matter what) must appear at least three times.
-> Thus, in a hand of three or more Gold cards and less than three Steel cards, it matches every Gold card, but no Steel card.

If you want it to match three cards that are either Gold or Steel each, I don't think this is currently supported actually, in which case I'll look into how I'd add it.

Aurelius7309 · 2026-04-03T21:45:56Z

Huh, I placed this matcher into the flush Poker Hand Part and it did nothing even with 3 gold + 3 steel cards. Might have done it wrong, I guess I'll experiment some more tomorrow. Also without looking at code I'd have expected the count to relate to the property and not the actual matched card - though either can be useful.

I suppose this means there's also currently no support for linking properties, e.g., two cards of the same rank and suit?

AllUniversal · 2026-04-03T21:47:33Z

Indeed, hmm. I'll see about that too.

+Title, this new boolean subflag makes it so that a card's property being counted is propagated to all other properties in the `any` flag of a matcher. (As long as the card matches the `any` flag). -> This allows counting e.g. a Steel card matching an `any = {m_steel = true, m_gold = true}` as a Gold card too.

…into card-matchers

*Title

AllUniversal · 2026-04-03T22:37:32Z

EDIT 2026.5.28: This is outdated now, the any_related subflag has been removed. The below functionality should be achievable by checking the number of cards matched instead.

If you want it to match three cards that are either Gold or Steel each, I don't think this is currently supported actually, in which case I'll look into how I'd add it.

This is now supported through the use of the new count subflag any_related; It makes it so that if a card matches a matcher.condition's any flag, and that condition (e.g. enhancement) is counted, the matcher also adds 1 to all the other values of the any flag, thus essentially merging their counts into one.

So the above example of matching at_least three cards that are either Steel or Gold is now possible with this matcher;

local matcher = SMODS.create_card_matcher({enhancement = {any = {m_gold = true, m_steel = true}, count = {at_least = 3, any_related = true}}})

…ax_cards ` +Title, this allows easy `matcher` duplication for hand detection, without having to calculate the same `matcher` multiple times.

…toring +/*Title, see LSP defs or PR for details on `overlap`.

AllUniversal · 2026-04-04T01:45:24Z

@Aurelius7309 Aight so I added a new count subflag overlap; It too has the all, any or none (new) subflags, but for conditions, not keys. This is a matcher for matching two or more cards, as long as they have a Red seal and share enhancements;

SMODS.create_card_matcher({seal = {any = {Red = true}, count = {at_least = 2, overlap = {any = {enhancement = true}}}}})

One for matching exactly two cards with a Red seal and shared enhancement;

SMODS.create_card_matcher({seal = {any = {Red = true}, count = {exact = 2, overlap = {any = {enhancement = true}}}}})

(This one would not match two cards with Red seals if there's a third one, unless that one has a different enhancement)

+Title

AllUniversal · 2026-05-20T12:53:06Z

Was just inspired to try implementing Holy Fool's Faust Joker's Devil's Hand pokerhand using card matchers, this is the code (replacing Flush for testing):

SMODS.PokerHandPart {
    key = '_flush',
    func = function(hand) 
        local matcher = SMODS.create_card_matcher({suit = {count = {exact = 1}}, rank = {count = {exact = 1}}})
        local m_t_c, c_t_m = SMODS.match_cards(hand, {matcher})
        return SMODS.get_hand_from_matching(m_t_c, c_t_m, {matcher_max_cards = {[matcher] = 4}})
    end -- get_flush(hand) end,
}

Making these four cards count as a Devil's Hand / Flush:

…s + added some warn messages -/+Title; -`count.any_related` subflag has been removed -> Instead check for the number of cards matched after the fact. + `count.overlap` has gained new subflags, allowing even more fine-grained matching. Example in the PR. +Added some warn messages when a matcher is created with invalid flags.

AllUniversal · 2026-05-28T12:35:17Z

I have now added subflags to count.overlap[condition];

"any" = boolean             -> any [condition]s overlap
"all" = boolean             -> all [condition]s of the primary card checked overlap with the card checked against
"all_either" = boolean      -> all [condition]s of the card with fewer [condition]s overlap with the other card 
"none" = boolean            -> none of the [condition]s overlap
"exact" = integer           -> [exact] number of cards overlap
"at_least" = integer        -> [at_least] number of cards overlap
"at_most" = integer         -> [at_most] number of cards overlap
"func" = functions          -> see `count.func`

These allow checking overlap more precisely, e.g.;

{suit = {count = {at_least = 3, overlap = {any = {enhancement = {none = true}, seal = {none = true}}}}}}

^ This is a matcher to match at least 3 cards which share a suit, as long as they either share no enhancements or no seals.

{suit = {count = {at_least = 3, overlap = {any = {enhancement = {none = true}, seal = {any = true, at_least = 3}}}}}}

^ This is a matcher to match at least 3 cards which share a suit, as long as they either share no enhancements or a seal with 2 other cards (with which they share a suit).

*Title

…cherCondition` + fixed/refactored some stuff */+Title, `SMODS.MatcherCondition` allows creating new conditions easily, for e.g. Paperback's paperclips.

*Title

+Title

AllUniversal added 2 commits April 2, 2026 21:48

Moved code from Steamodded#838.

6a5813b

+Title

De-Quantumified Rank related stuff

a6c7396

*/-Title, also removed some minor features, will re-add once Steamodded#838 is merged. :)

AllUniversal requested review from Aurelius7309, Eremel and GauntletGames-2086 as code owners April 2, 2026 20:06

AllUniversal added 2 commits April 3, 2026 03:00

LSP defs

bb61a25

+/*Title

Allow a condition to check for nil, e.g. no enhancement/seal/etc.

924fae1

+Title

AllUniversal marked this pull request as draft April 3, 2026 02:17

Refactored deferred_count to _pre_count to make it function

016cc11

*Title, the `count` subcondition is now evaluated at the same time as the others. -> This was necessary for it to not mess up if `condition.invert` was true. (Or in general)

AllUniversal marked this pull request as ready for review April 3, 2026 09:58

AllUniversal added 6 commits April 3, 2026 12:01

Small hand detection change

87585eb

*Title, modified an early return condition.

Tiniest refactor.

ac14973

*Title, realized I didn't like an early initilization of a map.

LSP def

df782e6

+Title

Tinierest refactor.

2a06df4

*Title, nuff said.

Refactoring 3: Tokyo drift

1aa1fbf

*/-Title, removed relic.

Fixed oversight

78145a5

*Title, forgor to put `check_function` into the `SMODS.insert_card_matcher_condition()` whitelist

Aurelius7309 reviewed Apr 3, 2026

View reviewed changes

AllUniversal and others added 5 commits April 3, 2026 12:53

Exposed _nil_sentinel (for matching a card WITHOUT e.g. an enhancem…

56bca8d

…ent) as `SMODS.card_matcher_nil_sentinel` *Title

Refactoring of SMODS.get_hand_from_matching() params as requested

5dc4bba

*Title

args should be optional here

c382871

missing pairs

deeea3c

there is another

7dff650

AllUniversal added 3 commits April 4, 2026 00:25

Merge branch 'card-matchers' of https://github.com/AllUniversal/smods …

d598214

…into card-matchers

Fix potential double counting of card property values.

8b6e02a

*Title

AllUniversal added 2 commits April 4, 2026 00:59

Added new SMODS.get_hand_from_matching() args argument `matcher_m…

aa1ba99

…ax_cards ` +Title, this allows easy `matcher` duplication for hand detection, without having to calculate the same `matcher` multiple times.

Added count subflag overlap to allow linking conditions + refac…

56684f0

…toring +/*Title, see LSP defs or PR for details on `overlap`.

AllUniversal added 7 commits April 4, 2026 11:43

Missing LSP def

fa0c00b

+Title

Merge remote-tracking branch 'upstream/main' into card-matchers

cfdc493

Merge remote-tracking branch 'upstream/main' into card-matchers

349302c

Merge remote-tracking branch 'upstream/main' into card-matchers

32f16b5

Merge remote-tracking branch 'upstream/main' into card-matchers

9ef6fb9

Merge remote-tracking branch 'upstream/main' into card-matchers

a909b79

Merge remote-tracking branch 'upstream/main' into card-matchers

d21af49

AllUniversal added 5 commits May 28, 2026 15:05

Small fixes and refactors

184e7c8

*Title

Moved everything into own card_matchers.lua file + Added `SMODS.Mat…

261c32e

…cherCondition` + fixed/refactored some stuff */+Title, `SMODS.MatcherCondition` allows creating new conditions easily, for e.g. Paperback's paperclips.

More refactoring

70ecc51

*Title

Moved and updated LSP defs.

0b8a74c

*Title

Missing LSP def

682684c

+Title

Conversation

AllUniversal commented Apr 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Additional Info:

Uh oh!

AllUniversal commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

AllUniversal commented Apr 3, 2026

Uh oh!

AllUniversal commented Apr 3, 2026

Uh oh!

Aurelius7309 Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

AllUniversal Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Aurelius7309 Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Aurelius7309 commented Apr 3, 2026

Uh oh!

AllUniversal commented Apr 3, 2026

Uh oh!

Aurelius7309 commented Apr 3, 2026

Uh oh!

AllUniversal commented Apr 3, 2026

Uh oh!

AllUniversal commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

AllUniversal commented Apr 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

AllUniversal commented May 20, 2026

Uh oh!

AllUniversal commented May 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

AllUniversal commented Apr 2, 2026 •

edited

Loading

AllUniversal commented Apr 3, 2026 •

edited

Loading

AllUniversal commented Apr 3, 2026 •

edited

Loading

AllUniversal commented Apr 4, 2026 •

edited

Loading

AllUniversal commented May 28, 2026 •

edited

Loading