graphable

graphable is a lightweight, type-safe Python library for building, managing, and visualizing dependency graphs. It provides a simple API for defining nodes and their relationships, performing topological sorts, and exporting graphs to various formats like Mermaid, Graphviz, and ASCII text trees.

Features

• Type-Safe: Built with modern Python generics and type hints.
• Topological Sorting: Easily get nodes in dependency order.
• Cycle Detection: Built-in protection against circular dependencies.
• Filtering & Tagging: Create subgraphs based on custom predicates or tags.
• Graph Syncing: Automatically expand graphs to include all reachable nodes with discover().
• Transitive Reduction: Automatically remove redundant edges while preserving reachability.
• Reachability Analysis: Easily find ancestors or descendants of any node.
• Node Ordering: Compare nodes using standard operators (a < b means a is an ancestor of b).
• Container Protocols: Use Pythonic idioms like len(graph), node in graph, and for node in graph.
• Integrity: Generate and validate deterministic BLAKE2b checksums of graph structure and metadata.
• Parallel Sorting: Group nodes into execution layers for parallel processing with parallelized_topological_order().
• Orchestration Ready: Track node status and duration for workflow execution.
• Edge Attributes: Support for weights, labels, and custom metadata on relationships.
• Diffing: Compare two graphs structurally and visually with diff() and diff_graph().
• Dependency Algorithms: Native implementations for:
  • Critical Path Method (CPM): Identify bottlenecks based on node durations.
  • BFS & DFS: Perform breadth-first or depth-first traversals with generators.
  • Longest Path: Find the deepest chain of dependencies.
  • All Paths: Explore every possible connection between two nodes.
  • Transitive Closure: Analyze the full "blast radius" of dependencies.
  • Cycle Resolution: Get suggestions for breaking circular dependencies with suggest_cycle_breaks().
• Advanced Slicing: Extract focused subgraphs using upstream_of(), downstream_of(), or subgraph_between().
• Unified I/O: Easy Graph.read() and graph.write() methods with automatic format detection.
• Parsing: Reconstruct graphs from JSON, YAML, TOML, CSV, or GraphML files and strings.
• Clustering: Group nodes into subgraphs/clusters in visualizations based on tags.
• Visualizations:
  • Mermaid: Generate flowchart definitions or export directly to SVG. Supports clustering and color-coded diffs.
  • Graphviz: Generate DOT definitions or export to SVG with custom styling. Supports clustering.
  • D2: Generate D2 definitions or export to SVG with modern styling and layouts. Supports clustering.
  • PlantUML: Generate component or deployment diagram definitions. Supports clustering.
  • TikZ: Generate high-quality LaTeX definitions for academic documents.
  • GraphML: Industrial-standard XML export for professional analysis tools (Gephi, yEd).
  • Interactive HTML: Generate a single, portable HTML file with zooming, panning, and Live Reload when used with the CLI.
  • JSON, YAML & TOML: Export graph structure as machine-readable data.
  • CSV: Export simple edge lists for data processing.
  • NetworkX: Seamless integration with the NetworkX library for advanced analysis.
  • Text Tree & ASCII Flowchart: Generate beautiful ASCII representations.
• Modern Tooling: Managed with uv and just.

Installation

As a Library

Use your preferred package manager to add graphable to your project:

# Using uv (recommended)
uv add graphable

# Using pip
pip install graphable

As a Command Line Tool

To use the graphable CLI globally, we recommend using pipx. You can choose between the bare-bones version or the full-featured "Rich" version:

# Full version with beautiful terminal output (recommended)
pipx install "graphable[cli]"

# Bare-bones version (standard library only)
pipx install graphable

---

Command Line Interface

graphable provides a powerful CLI for analyzing and transforming graph files.

Basic Usage

# Get summary information about a graph
graphable info topology.json

# Validate a graph for cycles and consistency
graphable check topology.yaml

# Convert between graph data formats
graphable convert input.json output.yaml

# Render a graph as an image (SVG or PNG)
graphable render input.json output.png --engine graphviz

# Simplify a graph using transitive reduction
graphable reduce complex.graphml simple.svg

# Compare two versions of a graph
graphable diff old.json new.json

# Generate a visual diff (added elements in green, removed in red)
graphable diff old.json new.json -o diff.svg

# Serve a live-reloading interactive visualization
graphable serve architecture.yaml

Automation & CI/CD

If you have the cli extra installed but need plain-text output for logging or automation, use the --bare flag:

graphable --bare info topology.json

Supported Formats

The CLI automatically detects formats based on file extensions:

• Parsers: .json, .yaml/.yml, .toml, .csv, .graphml
• Exporters: .json, .yaml, .toml, .csv, .graphml, .dot, .mmd, .d2, .puml, .html, .tex, .txt, .ascii, .cy.json
• Image Engines: Mermaid, Graphviz, D2, PlantUML (Supports .svg and .png)

---

Quick Start

from graphable.graph import Graph
from graphable.graphable import Graphable
from graphable.views.texttree import create_topology_tree_txt

# 1. Define your nodes
a = Graphable("Database")
b = Graphable("API Service")
c = Graphable("Web Frontend")

# 2. Build the graph
g = Graph()
g.add_edge(a, b)  # API Service depends on Database
g.add_edge(b, c)  # Web Frontend depends on API Service

# 3. Get topological order
for node in g.topological_order():
    print(node.reference)
# Output: Database, API Service, Web Frontend

# 4. Visualize as a text tree
print(create_topology_tree_txt(g))
# Output:
# Web Frontend
# └─ API Service
#    └─ Database

Visualizing with ASCII Flowchart

from graphable.views.asciiflow import create_topology_ascii_flow

print(create_topology_ascii_flow(g))
# Output:
# +----------+
# | Database |
# +----------+
#   v
#   +--> API Service
#
# +-------------+
# | API Service |
# +-------------+
#   v
#   +--> Web Frontend

Visualizing with Mermaid

from graphable.views.mermaid import create_topology_mermaid_mmd

mmd = create_topology_mermaid_mmd(g)
print(mmd)
# Output:
# flowchart TD
# Database --> API Service
# API Service --> Web Frontend

Visualizing with Graphviz

from graphable.views.graphviz import create_topology_graphviz_dot

dot = create_topology_graphviz_dot(g)
print(dot)
# Output:
# digraph G {
#     "Database" [label="Database"];
#     "Database" -> "API Service";
#     "API Service" [label="API Service"];
#     "API Service" -> "Web Frontend";
#     "Web Frontend" [label="Web Frontend"];
# }

Advanced Analysis

Edge Attributes & Orchestration

Track durations and weights for workflow optimization:

# Add weighted edges
g.add_edge(task_a, task_b, weight=10, label="critical path")

# Set task metadata
task_a.duration = 5.0
task_a.status = "completed"

Critical Path Method (CPM)

Identify the longest chain of dependencies:

analysis = g.cpm_analysis()
print(f"Project duration: {max(n['EF'] for n in analysis.values())}")

critical_nodes = g.critical_path()
print(f"Critical Path: {[n.reference for n in critical_nodes]}")

Graph Diffing

Compare two versions of a graph:

diff_data = g_v1.diff(g_v2)
print(f"Added nodes: {diff_data['added_nodes']}")

# Generate a visual diff graph for rendering
dg = g_v1.diff_graph(g_v2)
print(dg.render(create_topology_mermaid_mmd))

Advanced Slicing

Focus on specific parts of the graph:

# All nodes required by 'Web UI'
sub = g.upstream_of(ui)

# All nodes impacted by a change in 'Database'
sub = g.downstream_of(db)

# Everything between 'Source' and 'Sink'
sub = g.subgraph_between(source, sink)

Native Traversals (BFS & DFS)

Perform idiomatic iterations over your graph:

from graphable.enums import Direction

# Visit every node reachable from 'Task A' breadth-first
for node in g.bfs(task_a):
    print(f"Visiting: {node.reference}")

# Explore dependencies (upstream) depth-first
for node in g.dfs(task_a, direction=Direction.UP):
    print(f"Required by: {node.reference}")

Advanced Analysis with NetworkX

If you have networkx installed, you can convert your graph for advanced analysis:

import networkx as nx

# Convert to NetworkX DiGraph
dg = g.to_networkx()

# Use NetworkX algorithms
print(nx.dag_longest_path(dg))
# Output: ['Database', 'API Service', 'Web Frontend']

Visualizing with D2

from graphable.views.d2 import create_topology_d2

d2 = create_topology_d2(g)
print(d2)
# Output:
# Database: Database
# Database -> API Service
# API Service: API Service
# API Service -> Web Frontend
# Web Frontend: Web Frontend

Advanced Usage

Pythonic Protocols

print(f"Nodes: {len(g)}")
if "Database" in g:
    node = g["Database"]
    
for node in g:  # Iterates in topological order
    print(node.reference)

Transitive Reduction

Clean up your graphs by removing redundant edges:

reduced_g = g.transitive_reduction()
# Or render directly
print(g.render(create_topology_mermaid_mmd, transitive_reduction=True))

Clustering by Tags

Group nodes in visualizations:

from graphable.views.mermaid import MermaidStylingConfig

a.add_tag("backend")
b.add_tag("backend")
config = MermaidStylingConfig(cluster_by_tag=True)

print(g.render(create_topology_mermaid_mmd, config=config))

Parsing Graphs

Reconstruct graphs from exported data:

# From a file
g = Graph.from_json("graph.json")

# From a string (YAML)
yaml_data = """
nodes:
  - id: Database
  - id: API
edges:
  - source: Database
    target: API
"""
g = Graph.from_yaml(yaml_data)

Equality Comparison

Compare graphs easily:

g1 = Graph.from_json("topology.json")
g2 = Graph.from_yaml("topology.yaml")

if g1 == g2:
    print("Graphs are identical in structure and tags.")

Node Ordering

Nodes support comparison based on reachability:

if db < api:
    print("Database is an ancestor of API")
    
if ui > api:
    print("UI is a descendant of API")

Checksums & Integrity

Verify your graph hasn't changed:

digest = g.checksum()
# ... later ...
if g.validate_checksum(digest):
    print("Graph integrity verified.")

Unified I/O

Read and write any format without thinking about parsers:

# Automatically detects JSON
g = Graph.read("topology.json")

# Automatically detects YAML
g.write("topology.yaml")

Documentation

Full documentation is available in the docs/ directory. You can build it locally:

just docs-view

Security

graphable takes security seriously. We implement strict sanitization for all external inputs and use secure defaults for parsing and export. For more details, see our Security Policy.

If you discover a vulnerability, please report it via email to the maintainer.

Development

This project uses uv for dependency management and just as a command runner.

just install    # Install dependencies
just check      # Run linting, type checking, and tests
just coverage   # Run tests with coverage report

License

This project is licensed under the MIT License - see the LICENSE file for details.