unity-data-struct

name

unity-data-struct

description

Generate tree-format data structure documentation for Unity JSON schemas (db_*.json tables like db_TTE, db_DBSE). Use when the user pastes/shares a table of fields with short keys (a0, b1, c2...), wants a hierarchical tree diagram of a JSON data structure, asks to "create tree format", "data structure doc", "document db_XXX.json", or "generate data struct markdown". Output is a markdown file with a tree showing parent-child nesting (Object arrays → child objects) plus per-level field reference tables.

Unity Data Structure Documentation

Generate tree-format markdown documentation for Unity JSON data schemas (the db_*.json tables the tool team maintains for short-key-encoded game data like events, bosses, tiers, scoring rules).

When to use

Trigger on any of:

• Table image or pasted rows with columns: Name | Short key | Type | Description | Example value | Note
• File named db_*.json referenced in Docs/Features/<feature>/
• User says: "tree format for this data", "data structure documentation", "doc this schema", "create db_XXX tree"

Input assumptions

The source is usually a table where:

• Short key uses compressed letters+digit (a0..a10, b0..b6, c0..c6) — group keys by letter prefix to detect nesting depth.
• Type = string, int, float, TimeStamp, Object array, String array, Localized.
• Object array means the field contains nested child objects — those children appear as a new indented block in the tree.
• The letter prefix usually increases with depth: a* = root, b* = first nested object, c* = second nested object.

Output Format

Two sections, in this order:

1. Tree diagram (inside a fenced code block)

Use box-drawing characters. Match the simplified style below — do not add closing connector arrows (┐, ─) on the right edge; indent with spaces only.

db_TTE.json (Tactics Tactics Event)
│
├── ID  [string]  event_tactics_1  (Auto-generated by Brick)
│
├── a0  StartTime             [TimeStamp]  1783099545
├── a1  EndTime               [TimeStamp]  1791048345
├── a2  GenderDivision        [int]        0:Male / 1:Female / 2:Universal
│
├── a3  Tiers  [Object array]
│   │
│   └── Tier
│       ├── b0  TierId   [string]  "Tier1"
│       ├── b1  BossId   [string]  "boss_id_1"
│       ├── b3  FeatureScoringRule  [Object array]
│       │   │
│       │   └── ScoringRule
│       │       ├── c0  RuleName   [string]  "GemDestroyed"
│       │       ├── c1  Score      [int]     700
│       │
│       └── b4  RequiredSuperstarNumber [int]  1
│
└── a10  VFXDefeatedBoss  [string]  odb/tactics/grayfx

Rules:

• Top-level line: filename + parenthetical human label.
• One blank │ line as separator between logical groupings (root scalars vs nested object arrays).
• Each field: <short-key> <Name> [<type>] <example> with │ column rule alignment for readability.
• Trailing notes from the source (Auto by Brick, Refer from RoadBoss) placed after the example value.
• Object array fields introduce a child block at deeper indent; reuse the child name from the description or context (e.g., Tier, ScoringRule).

2. Field Reference Tables

One markdown table per level. Use the exact heading pattern:

## Field Reference

### Root (Event)

| Short Key | Name | Type | Description | Example |
|---|---|---|---|---|

### Tier Object (`a3[ ]`)

| Short Key | Name | Type | Description | Example |
|---|---|---|---|---|

### ScoringRule Object (`b3[ ]`)

| Short Key | Name | Type | Description | Example |
|---|---|---|---|---|

Rules:

• Columns: Short Key | Name | Type | Description | Example.
• Quote strings in examples ("Tier1"), keep numbers bare (700, 0.45).
• Object array type rows show — in Example column.
• Notes from source table (e.g. Auto-generated by Brick) go inside Description as parenthetical.
• Convert full type names (Object array → Object array), no abbreviations.

Workflow

1. Identify source: image of a table, pasted rows, or existing JSON.
2. Extract fields: name, short key, type, description, example, note per row.
3. Group by nesting: short-key letter prefix indicates depth. Object-array rows open a new indented block with child rows.
4. Draft the tree in a fenced ``` block (3 backticks, no language tag).
5. Draft the reference tables, one per nesting level.
6. Save to <project>/Docs/Features/<feature>/db_<NAME>_TreeStructure.md. If path not given, ask or default to Docs/Features/db_<NAME>_TreeStructure.md.

Style guidance

• Keep <short-key> <Name> pairs aligned with spaces; let the type column visually start at the same column for every row.
• The root file line should use the JSON filename verbatim (db_TTE.json).
• Don't decorate with emoji or color.
• Don't fabricate fields — only document what's in the source.

Anti-patterns

• Do NOT add a closing ─┐ arrow at the end of sub-tree introducer lines (user prefers simplified).
• Do NOT output JSON syntax ({}, []) in the tree — only box-drawing characters.
• Do NOT include source-table markdown that duplicates the reference tables — the tree + reference tables already cover everything once.