Graph I/O¶

PyGraphina supports multiple formats for saving and loading graphs.

Supported Formats¶

Format	Read	Write	Best For
Edge List	Yes	Yes	Simple, human-readable
JSON	Yes	Yes	Data interchange
Binary	Yes	Yes	Speed, compact storage
GraphML	No	Yes	Standard graph format

Edge List Format¶

Simple text format with edges, one per line.

# Save as edge list
g.save_edge_list("graph.txt", sep=" ")

# Load from edge list
g = pg.PyGraph()
num_nodes, num_edges = g.load_edge_list("graph.txt", sep=" ")

File format:

source target weight
0 1 1.0
1 2 2.0
2 3 1.5

JSON Format¶

Structured format with nodes and edges.

# Save as JSON
g.save_json("graph.json")

# Load from JSON
g = pg.PyGraph()
g.load_json("graph.json")

Binary Format¶

Compact, fast format for large graphs.

# Save as binary
g.save_binary("graph.bin")

# Load from binary
g = pg.PyGraph()
g.load_binary("graph.bin")

GraphML Format¶

Standard XML-based graph format.

# Save as GraphML
g.save_graphml("graph.graphml")

Can be loaded in other tools (Gephi, Cytoscape, etc.).

Example: Loading and Saving¶

import pygraphina as pg

# Create and modify a graph
g = pg.PyGraph()
for i in range(100):
    g.add_node(i)

for i in range(99):
    g.add_edge(i, i + 1, 1.0)

# Save in multiple formats
g.save_edge_list("graph.txt")
g.save_json("graph.json")
g.save_binary("graph.bin")
g.save_graphml("graph.graphml")

# Load in different format
g2 = pg.PyGraph()
g2.load_edge_list("graph.txt")

# Verify
assert g2.node_count() == g.node_count()
assert g2.edge_count() == g.edge_count()

Separator Options for Edge Lists¶

Common separators:

" " (space) - default, human-readable
"," (comma) - CSV format
"\t" (tab) - tab-separated

# CSV format
g.save_edge_list("graph.csv", sep=",")
g2.load_edge_list("graph.csv", sep=",")

# Tab-separated
g.save_edge_list("graph.tsv", sep="\t")
g3.load_edge_list("graph.tsv", sep="\t")

Format Comparison¶

Property	Edge List	JSON	Binary
Human-Readable	Yes	Yes	No
Compact	Moderate	No	Yes
Fast	Moderate	Slow	Fast
Standard	Yes	Yes	No

Tips¶

Use edge list for portability - Works with most tools
Use binary for speed - Fastest I/O for large graphs
Use JSON for interchange - Inspect and modify easily
Use GraphML for visualization - Standard format for graph tools