Domain-specific diagram builders and graph analyzers on top of Yog.
Choreo is a family of Elixir libraries that let you model, analyze, and render complex systems as graphs. Instead of drawing boxes and arrows by hand, you write code. Instead of static pictures, you get live analysis — reachability, cycles, bottlenecks, threat generation, and more.
alias Choreo.Dataflow
# A dataflow pipeline with one line of analysis
pipeline =
Dataflow.new()
|> Dataflow.add_source(:sensor, label: "IoT Sensor")
|> Dataflow.add_transform(:parse, label: "JSON Parser")
|> Dataflow.add_sink(:db, label: "TimescaleDB")
|> Dataflow.connect(:sensor, :parse, data_type: "raw bytes")
|> Dataflow.connect(:parse, :db, data_type: "event")
Dataflow.Analysis.cyclic?(pipeline) #=> false
Dataflow.to_mermaid(pipeline) #=> Mermaid diagramgraph TD
classDef default color:white
parse[["JSON Parser"]]
db["TimescaleDB"]
sensor(["IoT Sensor"])
style parse fill:#3b82f6
style db fill:#f43f5e
style sensor fill:#10b981
sensor -->|raw bytes| parse
parse -->|event| db
Add choreo to your mix.exs:
def deps do
[
{:choreo, "~> 0.9"}
]
endChoreo supports 13 different artifact modeling vocabularies. They are categorized and detailed in the following guides:
-
Architecture & Design Modeling
- System Architecture (
Choreo) - Cloud Network Topology (
Choreo.Infrastructure) - C4 Model Architecture (
Choreo.C4) - STRIDE Threat Modeling (
Choreo.ThreatModel) - Domain-Driven Design & Event Storming (
Choreo.Domain) - Database ERD Design (
Choreo.ERD) - UML Class & Struct Diagrams (
Choreo.UML)
- System Architecture (
-
- Finite State Machines (
Choreo.FSM) - Sequence Diagrams (
Choreo.Sequence) - Saga Task Orchestration (
Choreo.Workflow) - Dataflow Pipelines (
Choreo.Dataflow)
- Finite State Machines (
-
- Software Dependency Graphs (
Choreo.Dependency) - Decision Trees (
Choreo.DecisionTree) - Concept Mapping / Mind Maps (
Choreo.MindMap) - Project Task Planning (
Choreo.Planner)
- Software Dependency Graphs (
Choreo provides powerful graph analysis tools to identify "hotspots" in your architecture, workflows, and pipelines. Use heatmap/2 to automatically color nodes based on importance or performance metrics.
| Metric | Measure | Question | Best for |
|---|---|---|---|
| Structural Importance | Betweenness Centrality | "Which nodes are critical bridges/connectors?" | Choreo, Dependency |
| Connectivity | Degree Centrality | "Which nodes have the most connections?" | MindMap, Dependency |
| SPOF Detection | Articulation Points | "Which nodes would disconnect the system if they failed?" | Choreo, Dataflow |
| Nucleus Detection | K-Core Decomposition | "Which nodes form the most tightly-coupled core?" | Choreo, Dependency |
| Dependency Reduction | Transitive Reduction | "What is the minimal set of dependencies that preserve reachability?" | Dependency |
| Path Analysis | Dijkstra / Widest Path | "What is the fastest or highest-throughput path between two points?" | Workflow, Dataflow |
| Execution Hotspots | Latency Heatmap | "Which tasks slow down the entire workflow?" | Workflow |
| Volume Hotspots | Throughput Heatmap | "Which stages handle the most data volume?" | Dataflow |
| Security Hotspots | Risk Heatmap | "Which components have the most security threats?" | ThreatModel |
All modules render to DOT (Graphviz) and Mermaid.js via a shared theming pipeline.
# DOT output (Graphviz)
Choreo.to_dot(system, theme: :default)
Choreo.to_dot(system, theme: :dark)
# Mermaid.js output (GitHub, GitLab, Notion, Livebook)
Choreo.to_mermaid(system, theme: :default)
Choreo.to_mermaid(system, theme: :ocean)
# Custom theme
theme = Choreo.Theme.custom(
colors: %{database: "#ff0000", service: "#00ff00"},
graph_bgcolor: "#0f172a",
node_fontcolor: "white"
)
Choreo.to_dot(system, theme: theme)
Choreo.to_mermaid(system, theme: theme)mix testAll modules ship with comprehensive ExUnit test suites covering builders, analysis, rendering, and doctests.
MIT