chart Operation - Ensemble Edge

The chart operation generates data visualizations from your data without external services or client-side rendering. Output as server-rendered SVG (default), QuickChart.io image URLs, or Vega-Lite specs for interactive client-side rendering.

The chart operation is fully Workers-compatible - no DOM, Canvas, or external dependencies. All rendering is pure string manipulation.

Quick Start

Simple bar chart:

agents:
  - name: revenue-chart
    operation: chart
    config:
      type: bar
      data: ${sales.output}
      x: month
      y: revenue
      title: Monthly Revenue

Line chart with multiple series:

agents:
  - name: metrics-chart
    operation: chart
    config:
      type: line
      data: ${metrics.output}
      x: date
      y: [users, sessions, conversions]
      smooth: true

Pie chart:

agents:
  - name: distribution-chart
    operation: chart
    config:
      type: pie
      data: ${categories.output}
      labels: category
      values: amount

Configuration

config:
  # Chart type (required unless auto: true)
  type: string           # bar, line, area, pie, donut, scatter, sparkline
  auto: boolean          # Auto-detect best chart type from data

  # Data and axes
  data: array            # Array of data objects (required)
  x: string              # Field name for x-axis (for cartesian charts)
  y: string | string[]   # Field name(s) for y-axis (for cartesian charts)
  labels: string         # Field for labels (for pie/donut)
  values: string         # Field for values (for pie/donut)

  # Output format
  output: string         # svg (default), url, vega, html

  # Dimensions
  width: number          # Chart width in pixels (default: 600)
  height: number         # Chart height in pixels (default: 400)

  # Labels and titles
  title: string          # Chart title
  subtitle: string       # Chart subtitle
  xAxis:
    label: string        # X-axis label
    format: string       # Number format: compact, percent, currency
  yAxis:
    label: string        # Y-axis label
    format: string       # Number format: compact, percent, currency

  # Styling
  theme: string          # light (default), dark, minimal, brand
  colors: string[]       # Custom color palette

  # Chart options
  legend: boolean | string  # Show legend: true, false, top, bottom, right
  stacked: boolean       # Stack bars/areas (for multi-series)
  horizontal: boolean    # Horizontal bars (for bar charts)
  smooth: boolean        # Smooth curves (for line/area charts)
  showValues: boolean    # Show values on chart (default: true)

  # Reference lines
  referenceLine:
    value: number        # Y-value for the reference line (or 'average', 'median')
    label: string        # Label text
    style: string        # solid, dashed, dotted
    color: string        # Line color

  # Annotations
  annotations: array     # Array of annotation configs
    - type: string       # line, band, point, label
      value: number | [number, number]  # Value or range for bands
      label: string      # Annotation label
      color: string      # Annotation color
      style: string      # solid, dashed, dotted (for lines)

Chart Types

Bar Chart

Vertical or horizontal bar charts with optional grouping or stacking.

agents:
  - name: sales-chart
    operation: chart
    config:
      type: bar
      data:
        - { month: "Jan", revenue: 45000, costs: 32000 }
        - { month: "Feb", revenue: 52000, costs: 35000 }
        - { month: "Mar", revenue: 48000, costs: 31000 }
      x: month
      y: [revenue, costs]
      title: Revenue vs Costs
      legend: true

Stacked bar chart:

config:
  type: bar
  data: ${regional-sales.output}
  x: quarter
  y: [north, south, east, west]
  stacked: true
  title: Regional Sales by Quarter

Horizontal bar chart:

config:
  type: bar
  data: ${rankings.output}
  x: score
  y: name
  horizontal: true

Line Chart

Line charts with single or multiple series, optional smoothing.

agents:
  - name: trends-chart
    operation: chart
    config:
      type: line
      data:
        - { date: "2024-01", users: 1000, sessions: 2500 }
        - { date: "2024-02", users: 1500, sessions: 3200 }
        - { date: "2024-03", users: 1200, sessions: 2800 }
      x: date
      y: [users, sessions]
      smooth: true
      title: User Metrics

Area Chart

Filled area charts with gradient fills, similar to line charts.

agents:
  - name: traffic-chart
    operation: chart
    config:
      type: area
      data: ${traffic.output}
      x: hour
      y: requests
      title: Requests per Hour

Stacked area:

config:
  type: area
  data: ${sources.output}
  x: date
  y: [organic, paid, direct]
  stacked: true

Pie & Donut Charts

Circular charts for showing proportions.

agents:
  - name: market-share
    operation: chart
    config:
      type: pie
      data:
        - { company: "A", share: 35 }
        - { company: "B", share: 28 }
        - { company: "C", share: 22 }
        - { company: "Other", share: 15 }
      labels: company
      values: share
      title: Market Share

Donut chart (pie with hole in center):

config:
  type: donut
  data: ${expenses.output}
  labels: category
  values: amount
  title: Expense Breakdown

Scatter Plot

Show relationships between two numeric variables.

agents:
  - name: correlation-chart
    operation: chart
    config:
      type: scatter
      data:
        - { price: 100, rating: 4.5, brand: "A" }
        - { price: 200, rating: 4.0, brand: "B" }
        - { price: 150, rating: 4.8, brand: "A" }
      x: price
      y: rating
      title: Price vs Rating

Sparkline

Minimal inline charts for dashboards and tables.

agents:
  - name: trend-spark
    operation: chart
    config:
      type: sparkline
      data: [10, 15, 12, 18, 14, 22, 19]
      width: 100
      height: 30

From objects:

config:
  type: sparkline
  data: ${daily-metrics.output}
  y: value
  width: 120
  height: 40

Output Formats

SVG (Default)

Server-rendered SVG string. Best for embedding directly in HTML.

config:
  type: bar
  data: ${data.output}
  x: category
  y: value
  output: svg  # Default, can be omitted

URL (QuickChart.io)

Generate an image URL for use in emails, Slack, etc.

config:
  type: line
  data: ${metrics.output}
  x: date
  y: value
  output: url

Returns: https://quickchart.io/chart?c={...}

QuickChart URLs are generated client-side. The chart config is URL-encoded, so very large datasets may exceed URL length limits.

Vega-Lite Spec

Generate a Vega-Lite specification for client-side interactive rendering.

config:
  type: bar
  data: ${data.output}
  x: category
  y: value
  output: vega

Returns a JSON object you can render with Vega-Embed.

HTML

Self-contained HTML document with embedded SVG.

config:
  type: pie
  data: ${data.output}
  labels: name
  values: amount
  output: html

Useful for generating standalone chart pages or PDF conversion.

Theming

Built-in Themes

# Light theme (default) - white background, dark text
config:
  theme: light

# Dark theme - dark background, light text
config:
  theme: dark

# Minimal theme - no borders, subtle colors
config:
  theme: minimal

# Brand theme - blue primary color
config:
  theme: brand

Custom Colors

Override the color palette:

config:
  type: bar
  data: ${data.output}
  x: category
  y: [series1, series2, series3]
  colors:
    - "#FF6B6B"  # Coral
    - "#4ECDC4"  # Teal
    - "#45B7D1"  # Sky blue

Reference Lines

Add horizontal reference lines for targets, thresholds, or averages.

config:
  type: line
  data: ${performance.output}
  x: date
  y: score
  referenceLine:
    value: 80
    label: Target
    style: dashed
    color: "#FF0000"

Reference lines can also use dynamic values:

referenceLine:
  value: average  # or median
  label: Average

Annotations

Add annotations to highlight specific data points or ranges.

Line Annotation

Draw a horizontal or vertical line at a specific value:

config:
  type: bar
  data: ${sales.output}
  x: month
  y: revenue
  annotations:
    - type: line
      value: 50000
      label: Target
      color: "#FF0000"
      style: dashed

Band Annotation

Highlight a range of values:

config:
  type: line
  data: ${metrics.output}
  x: date
  y: score
  annotations:
    - type: band
      value: [70, 90]       # Range from 70 to 90
      label: Target Range
      color: "#00FF00"

Point Annotation

Mark a specific point:

config:
  type: line
  data: ${performance.output}
  x: date
  y: score
  annotations:
    - type: point
      value: 85
      label: Peak
      color: "#0000FF"

Multiple Annotations

Combine multiple annotations:

config:
  type: bar
  data: ${sales.output}
  x: quarter
  y: revenue
  annotations:
    - type: line
      value: 100000
      label: Goal
      color: "#00AA00"
      style: dashed
    - type: band
      value: [80000, 100000]
      label: Target Zone
      color: "#FFCC00"
    - type: line
      value: 50000
      label: Minimum
      color: "#FF0000"
      style: dotted

Auto Detection

Let the chart agent automatically choose the best chart type:

config:
  auto: true
  data: ${unknown-data.output}
  labels: category
  values: value

Auto-detection rules:

Small categorical data (2-6 items with string + number) → pie
Time-series data (date field + numeric) → line
Two numeric fields without categories → scatter
Single numeric array → sparkline
Default → bar

Number Formatting

Format axis values and labels:

config:
  type: bar
  data: ${revenue.output}
  x: month
  y: amount
  yAxis:
    label: Revenue
    format: currency  # $1.2K, $45.3M, etc.

# Available formats:
# - compact: 1.2K, 45.3M (default)
# - percent: 45%, 123%
# - currency: $1.2K, $45.3M
# - none: raw numbers

Complete Dashboard Example

name: sales-dashboard

trigger:
  - type: http
    path: /dashboard
    methods: [GET]

agents:
  # Fetch data
  - name: fetch-sales
    operation: data
    config:
      type: d1
      database: analytics
      query: |
        SELECT month, revenue, costs, profit
        FROM monthly_sales
        WHERE year = 2024
        ORDER BY month

  # Main revenue chart
  - name: revenue-chart
    operation: chart
    config:
      type: bar
      data: ${fetch-sales.output}
      x: month
      y: [revenue, costs]
      title: Revenue vs Costs
      theme: dark
      legend: bottom
      yAxis:
        format: currency

  # Profit trend
  - name: profit-chart
    operation: chart
    config:
      type: area
      data: ${fetch-sales.output}
      x: month
      y: profit
      title: Profit Trend
      smooth: true
      referenceLine:
        value: 0
        label: Breakeven
        style: dashed

  # Sparkline for header
  - name: revenue-spark
    operation: chart
    config:
      type: sparkline
      data: ${fetch-sales.output}
      y: revenue
      width: 100
      height: 30

flow:
  - agent: fetch-sales
  - parallel:
      - agent: revenue-chart
      - agent: profit-chart
      - agent: revenue-spark

output:
  body:
    revenueChart: ${revenue-chart.output}
    profitChart: ${profit-chart.output}
    sparkline: ${revenue-spark.output}
  headers:
    Content-Type: text/html

Performance

Chart generation is fast since it’s pure string manipulation:

Chart Type	Data Size	Typical Speed
Bar	10 items	<5ms
Bar	100 items	<15ms
Line	100 points	<10ms
Line	1000 points	<50ms
Pie	6 slices	<5ms
Scatter	100 points	<10ms
Sparkline	50 points	<2ms

Error Handling

Missing Required Fields

# This will throw an error:
config:
  type: bar
  data: ${data.output}
  x: month
  # y is missing

Error: chart: bar chart requires config.x and config.y

Invalid Chart Type

# This will throw an error:
config:
  type: histogram
  data: ${data.output}

Error: chart: invalid type "histogram". Must be one of: bar, line, area, pie, donut, scatter, sparkline

Empty Data

# This will throw an error:
config:
  type: bar
  data: []
  x: category
  y: value

Error: chart: config.data must be a non-empty array

transform - Prepare and transform data for charts
data - Fetch data from databases
http - Fetch data from APIs
html - Embed charts in HTML templates
pdf - Generate PDF reports with charts

Conductor

Getting Started

Core Concepts

Building

Components

Operations Reference

Plugins

Starter Kit

Playbooks

Reference

​Quick Start

​Configuration

​Chart Types

​Bar Chart

​Line Chart

​Area Chart

​Pie & Donut Charts

​Scatter Plot

​Sparkline

​Output Formats

​SVG (Default)

​URL (QuickChart.io)

​Vega-Lite Spec

​HTML

​Theming

​Built-in Themes

​Custom Colors

​Reference Lines

​Annotations

​Line Annotation

​Band Annotation

​Point Annotation

​Multiple Annotations

​Auto Detection

​Number Formatting

​Complete Dashboard Example

​Performance

​Error Handling

​Missing Required Fields

​Invalid Chart Type

​Empty Data

​Related Operations

Quick Start

Configuration

Chart Types

Bar Chart

Line Chart

Area Chart

Pie & Donut Charts

Scatter Plot

Sparkline

Output Formats

SVG (Default)

URL (QuickChart.io)

Vega-Lite Spec

HTML

Theming

Built-in Themes

Custom Colors

Reference Lines

Annotations

Line Annotation

Band Annotation

Point Annotation

Multiple Annotations

Auto Detection

Number Formatting

Complete Dashboard Example

Performance

Error Handling

Missing Required Fields

Invalid Chart Type

Empty Data

Related Operations