Schema Configuration

The columns prop defines the data schema that drives editor suggestions. Keys are column names, values are column definitions.

Basic Schema

const columns = {
    level: { type: 'enum', suggest: true, autocomplete: true, values: ['debug', 'info', 'error'] },
    service: { type: 'string', suggest: true, autocomplete: true },
    status_code: { type: 'number', suggest: true },
    host: { type: 'string', suggest: true },
}

Column Definition Fields

Field	Type	Default	Description
type	string	—	Column type: 'string', 'number', 'enum', 'object'
suggest	boolean	true	Show this column in suggestions
autocomplete	boolean	false	Enable value autocomplete for this column
values	string[] or number[]	—	Static values for autocomplete (used before onAutocomplete callback)
children	object	—	Nested column definitions for object-type columns
display_name	string	''	Optional user-friendly display name for the column. On parsed columns, the alias overrides this value

Column Types

string

General text column. All comparison operators are available including regex (~, !~).

{ type: 'string', suggest: true }

number

Numeric column. Regex operators (~, !~) are excluded from operator suggestions.

{ type: 'number', suggest: true }

enum

Enumerated column with a fixed set of values. Like number, regex operators are excluded. Use values to provide the set:

{ type: 'enum', suggest: true, autocomplete: true, values: ['debug', 'info', 'warning', 'error', 'critical'] }

object

Nested column with children. When the user types the column name followed by ., child columns are suggested:

{
    type: 'object',
    suggest: true,
    children: {
        labels: {
            type: 'object',
            suggest: true,
            children: {
                tier: { type: 'string', suggest: true },
                env: { type: 'string', suggest: true },
            },
        },
        version: { type: 'string', suggest: true },
    },
}

Value Autocomplete

Static Values

Provide a values array for columns with known options:

{
    level: {
        type: 'enum',
        suggest: true,
        autocomplete: true,
        values: ['debug', 'info', 'warning', 'error', 'critical'],
    },
    status_code: {
        type: 'number',
        suggest: true,
        autocomplete: true,
        values: [200, 201, 204, 400, 404, 500],
    },
}

Async Values

Set autocomplete: true without values and provide an onAutocomplete callback on the editor component. The callback is called when the user reaches the value position for that column.

// Column definition — no static values
{ type: 'string', suggest: true, autocomplete: true }

// Editor callback fetches values from API
async function onAutocomplete(key, value) {
    const resp = await fetch(`/api/values?column=${key}`)
    const data = await resp.json()
    return { items: data.values }
}

Nested Columns (JSONSchema)

Define nested structures using children on object-type columns:

const columns = {
    metadata: {
        type: 'object',
        suggest: true,
        children: {
            labels: {
                type: 'object',
                suggest: true,
                children: {
                    tier: { type: 'string', suggest: true, autocomplete: true, values: ['dev', 'staging', 'prod'] },
                    env: { type: 'string', suggest: true, autocomplete: true },
                },
            },
            version: { type: 'string', suggest: true },
        },
    },
}

Typing metadata. suggests labels and version. Typing metadata.labels. suggests tier and env. There is no depth limit.

Remote Key Discovery (Schemaless)

For object columns where the nested keys are not known at schema definition time, omit children and provide an onKeyDiscovery callback:

const columns = {
    request: { type: 'object', suggest: true }, // no children — triggers discovery
}

async function onKeyDiscovery(columnName, segments) {
    const resp = await fetch(`/api/keys?column=${columnName}&path=${segments.join(',')}`)
    const data = await resp.json()
    return data.keys // Array<{ name: string, type?: string, hasChildren?: boolean }>
}

When the user types request., the callback is called with ('request', ['request']). Discovered keys are cached per activation session.

If a column has both children and onKeyDiscovery, the static children take priority.

Full Example

const columns = {
    // Enum with static values
    level: { type: 'enum', suggest: true, autocomplete: true, values: ['debug', 'info', 'error'] },

    // String with async autocomplete
    service: { type: 'string', suggest: true, autocomplete: true },

    // Number — excludes regex operators
    status_code: { type: 'number', suggest: true, autocomplete: true, values: [200, 400, 500] },

    // Simple string
    host: { type: 'string', suggest: true },

    // Nested with known children
    metadata: {
        type: 'object',
        suggest: true,
        children: {
            labels: {
                type: 'object',
                suggest: true,
                children: {
                    tier: { type: 'string', suggest: true, autocomplete: true, values: ['dev', 'staging', 'prod'] },
                    env: { type: 'string', suggest: true, autocomplete: true },
                },
            },
            version: { type: 'string', suggest: true },
        },
    },

    // Schemaless — nested keys discovered at runtime
    request: { type: 'object', suggest: true },
}