Algorithms

Field contracts and worked examples for the algorithms blocks — converted at build time from the repo's authoring-skill reference. Every example body is YAML; in a real doc the fence language is the block type (for example sequence … ). Schemas are strict — unknown fields are rejected. See the field contract for the at-a-glance shape of all 77 blocks, or the block catalog for live previews.

Algorithms & data structures

Purpose-built for algorithm walkthroughs and CS explainers — one step per block (freeze a moment, don't animate). All four share the same tone enum: active (navy — the element under examination) · visited (light blue — already processed) · target (green — the goal) · muted (gray — out of play). For graph algorithms (BFS / DFS / Dijkstra), use graph with node state + edge weight instead.

array — array cells for algorithm walkthroughs

A row of square cells. value is a string — quote numbers (value: "19"). Indices render above each cell (showIndex, default true); a pointer label ("i", "lo", "mid") renders below its cell with a ▲ tick; window draws a navy-dashed outline around a 0-based inclusive index range (out-of-bounds values clamp). Use array for binary search, two pointers, and sliding windows; use table for genuinely tabular data.

linkedlist — pointer-chain diagram

Boxed nodes (a value cell + a pointer-dot cell) joined by arrows. kind: doubly adds a second, lower back-arrow per link; nullEnd (default true) terminates the chain in a ∅ ground symbol. A node label ("head", "curr") renders above its node with a ▼ tick. Use linkedlist for pointer manipulation (reversal, insertion, fast/slow); use flow for control flow.

bintree — binary tree

Nodes reference their parent by id and must set side: left | right when they do — a parent without a side is a schema error, and so is placing two children on one side. Parentless nodes are roots; multiple roots lay out side by side (handy for showing rotations). A parent centres over its children and single-child nodes offset toward their side, so unbalanced chains slant instead of stacking. Tint the chain down to a target to show a search path. Use bintree for BSTs, heaps, and traversals; use tree for file hierarchies (and tree with variant: issue for issue breakdowns).

hashmap — buckets + chained entries

A vertical column of bucket slots (indices 0..buckets-1); each bucket's entries chain rightward as rounded key: value pills joined by arrows — collision chains read left → right in entry order. Empty buckets show a dim "—". Entries whose bucket falls outside 0..buckets-1 are skipped, not clamped — they simply don't render, so fix the index. Rendering caps at 12 buckets with a "+N more" note; keep the count small enough to read. Use hashmap for hashing / collision walkthroughs; use table for plain key-value listings.