plantuml-compose 1.0.0

A Python library that makes creating PlantUML diagrams intuitive via type hints

plantuml-compose

A Python library that makes creating PlantUML diagrams intuitive and type-safe.

Why plantuml-compose?

PlantUML is powerful but its text syntax can be cryptic. What does A -[#red,dashed]-> B mean? What options are available for a state transition?

plantuml-compose solves this by providing:

• Discoverable APIs: Use your IDE's autocomplete to explore options
• Type safety: Catch errors before rendering, not after
• Pure factory functions: Compose diagrams declaratively with nesting
• Full PlantUML coverage: Every diagram type, every feature

Quick Start

from plantuml_compose import sequence_diagram, render

d = sequence_diagram(title="API Call")
p = d.participants
e = d.events

client = p.actor("Client")
server = p.participant("Server")
d.add(client, server)

d.phase("Request", [
    e.message(client, server, "GET /users"),
    e.reply(server, client, "200 OK"),
])

print(render(d))

The Pattern

Every diagram follows the same shape:

# 1. Create diagram — get namespace(s)
d = state_diagram(title="Lifecycle")
el = d.elements
t = d.transitions

# 2. Create elements via namespace factories
idle = el.state("Idle")
active = el.state("Active")

# 3. Register elements
d.add(idle, active)

# 4. Create and register connections
d.connect(
    t.transitions(
        (el.initial(), idle, "start"),
        (idle, active, "go"),
        (active, el.final(), "done"),
    )
)

# 5. Render
print(render(d))

Namespaces organize the API — type d.elements. or d.transitions. and your IDE shows every available factory method.

Bulk helpers reduce repetition:

# Instead of individual calls:
t.transition(a, b, label="x"),
t.transition(b, c, label="y"),
t.transition(c, d, label="z"),

# Use the plural form:
t.transitions(
    (a, b, "x"),
    (b, c, "y"),
    (c, d, "z"),
)

# Or fan-out from one source:
r.arrows_from(api,
    (db, "queries"),
    (cache, "reads"),
    (queue, "publishes"),
)

Styling

Two levels of styling are available on every diagram type.

Inline styles on individual elements and connections:

# Element styling
el.state("Error", style={"background": "#FFCDD2", "line_color": "red"})

# Connection styling — string shorthand or dict
t.transition(a, b, style="dashed")
t.transition(a, b, style={"color": "red", "pattern": "dotted"})

# Arrow length (layout hint)
r.arrow(a, b, length=1)   # short: ->
r.arrow(a, b, length=3)   # long:  --->

# Custom arrow heads
r.arrow(a, b, left_head="o", right_head=">>")  # o-->>

Diagram-wide styling via diagram_style=:

d = class_diagram(
    title="Model",
    diagram_style={
        "class_": {"background": "#E3F2FD", "round_corner": 10},
        "arrow": {"line_color": "gray", "font_size": 10},
        "note": {"background": "#FFF9C4"},
        "stereotypes": {
            "important": {"background": "pink", "font_style": "bold"},
        },
    },
)

Every element property is documented in ElementStyleDict — hover over it in your IDE to see all available keys (background, line_color, font_color, font_name, font_size, padding, margin, round_corner, shadowing, max_width, and more).

Supported Diagram Types

Behavioral

Type	Use When	Guide
Sequence	Message flows between participants	API calls, protocols
State	Entity lifecycle tracking	Order status, workflows
Activity	Process flow with decisions and parallelism	Business processes, algorithms
Use Case	System boundaries and actor goals	Requirements, features
Timing	State changes over time	Hardware signals, protocol timing

Structural

Type	Use When	Guide
Class	Types, attributes, and relationships	Domain models, API types
Object	Specific instances and values	Test scenarios, snapshots
Component	System architecture	Services, modules, dependencies
Deployment	Software on infrastructure	Servers, containers, cloud
Network	Network topology	IP addressing, segments

Hierarchical

Type	Use When	Guide
Mind Map	Brainstorming and organizing ideas	Feature planning, concepts
WBS	Breaking down work	Project planning, tasks
Gantt	Scheduling with dependencies	Timelines, sprint planning

Data & UI

Type	Use When	Guide
JSON	Visualizing JSON data	API responses, config
YAML	Visualizing YAML data	Config files, specs
Salt	UI wireframes	Forms, layouts, menus

Composition

Feature	Guide
Sub-diagrams	Embed diagrams inside notes and messages

Installation

pip install plantuml-compose

Requires Python 3.13+.

Development

# Install dependencies
uv sync

# Run tests
uv run pytest

# Validate output against PlantUML
plantuml --check-syntax output.puml

License

MIT