fix(go): replace emirpasic/gods with Go stdlib in article solutions

[neetcode-gh]

Problem

Article Go solutions that used github.com/emirpasic/gods cannot run on the judge.

On the platform, an article's Go snippet is spliced into a per-problem driver that already declares package main and owns the entire import block:

package main

import (
    "bufio"; "fmt"; "os"; "strings"
    "math"; "sort"; "strconv"; "regexp"
    "container/heap"; "container/list"
    "math/rand"; "math/bits"; "unicode"; "bytes"
    "encoding/json"; "runtime/debug"
)

[userCodeGoesHere]

A snippet therefore cannot add an import, and a third-party module could not be resolved even if it could. Every gods-based Go tab failed to compile.

An earlier approach — teaching the drivers to import gods — was abandoned. Rewriting to the standard library is the maintainer-confirmed fix.

Change

All 28 affected Go tabs rewritten to stdlib only:

gods construct	stdlib replacement
priorityqueue.NewWith(...)	container/heap + a local sort.Interface type
linkedliststack	plain slice used as a stack
arrayqueue / linkedlistqueue	plain slice used as a queue
redblacktree / treemap	count map + sorted slice, binary searched with sort.Search / sort.SearchInts

Because the driver already provides container/heap, sort, math and the rest ambiently, none of the new snippets declare imports. Published structure and naming are kept as close to the originals as possible, and behaviour is preserved throughout.

Also removes a stray import "container/heap" from the design-a-leaderboard Heap for top-K tab — not a gods tab, but broken on the judge for exactly the same reason (heap redeclared in this block / "container/heap" imported and not used).

Two latent bugs disappear as a side effect:

• hand-of-straights now bounds-checks the heap before peeking (the old code would panic on a type-asserted nil peek of an empty queue).
• design-a-leaderboard Reset no longer decrements the count for score 0 when the player is absent.

Where Go lacks an ordered-map equivalent (design-a-leaderboard, longest-continuous-subarray-…), a short comment explains the count-map + sorted-slice modelling so the article still reads as a teaching text.

Verification

Every rewritten tab was executed against the production judge via test-problem-article-solutions.ts --lang go, sequentially with a 250 ms delay. 29/29 pass.

Article	Rewritten tab(s)	Result
binary-tree-diameter	Iterative DFS	PASS
cheapest-flight-path	Dijkstra's Algorithm	PASS
depth-of-binary-tree	Breadth First Search	PASS
design-a-leaderboard	Using a TreeMap / SortedMap	PASS
design-a-leaderboard	Heap for top-K (stray-import fix)	PASS
design-twitter-feed	Heap; Heap (Optimal)	PASS; PASS
find-median-in-a-data-stream	Heap	PASS
hand-of-straights	Heap	PASS
invert-a-binary-tree	Breadth First Search	PASS
k-closest-points-to-origin	Min-Heap; Max Heap	PASS; PASS
kth-largest-element-in-an-array	Min-Heap	PASS
kth-largest-integer-in-a-stream	Min-Heap	PASS
last-stone-weight	Heap	PASS
longest-continuous-subarray-with-absolute-diff-…	Sorted Dict	PASS
meeting-schedule-ii	Min Heap	PASS
merge-k-sorted-linked-lists	Heap	PASS
min-cost-to-connect-points	Prim's Algorithm	PASS
minimum-interval-including-query	Sweep Line Algorithm; Min Heap	PASS; PASS
minimum-stack	Brute Force; Two Stacks	PASS; PASS
network-delay-time	Dijkstra's Algorithm	PASS
sliding-window-maximum	Heap	PASS
swim-in-rising-water	Dijkstra's Algorithm	PASS
task-scheduling	Max-Heap	PASS
top-k-elements-in-list	Min-Heap	PASS
trapping-rain-water	Stack	PASS
validate-parentheses	Stack	PASS

The sweep ran every Go tab of each affected problem, not just the rewritten ones. The only failures anywhere were pre-existing Brute Force time-limit-exceeded results on minimum-interval-including-query, sliding-window-maximum, swim-in-rising-water and trapping-rain-water — all untouched by this PR and already failing on main.

Judge config the runs were made against

Run against a neetcode.io working tree carrying in-progress per-language time-limit calibration. Per-problem comparison against neetcode.io main:

• trapping-rain-water (go: 0.4) and meeting-schedule-ii (go: 0.5) — limits identical to main.
• task-scheduling and sliding-window-maximum — no go limit on either branch, so both ran on the default 2.5 s CPU allowance, same as main.
• minimum-interval-including-query — go: 0.45 exists only on the unmerged calibration branch; main has no go entry and would use the 2.5 s default. This tab was therefore verified under a stricter limit than main enforces.
• All other problems have no timeLimits entry at all and ran on the default, identical to main.

So every result above holds under main's configuration.

Out of scope

Three articles still use gods — data-stream-as-disjoint-intervals, design-a-food-rating-system, seat-reservation-manager. None has a Go judge directory, so their rewrites could not be verified and are deliberately left alone.

Separately, ~44 Go blocks across ~30 other articles still carry import lines and will fail on the judge for the same reason as the design-a-leaderboard tab fixed here. That is a broader cleanup, tracked separately rather than bundled into this PR.

[neetcode-gh]

@coderabbitai review

[neetcode-gh]

@greptileai review

[neetcode-gh]

bugbot run

[cursor]

Skipping Bugbot: Bugbot is disabled for this repository. Visit the Bugbot dashboard to update your settings.