File Formats & Attributes

File Formats

`.ilu` — Native Format

The native Illuminator binary format. This is the only format that preserves all graph data including node positions, styles, attributes, and edge information. Load with load file.ilu or File > Open. Save with File > Save As.

Use .ilu files as your primary working format when you want to save and resume a session with all visual state intact.

`.csv` — Comma-Separated Values

Standard comma-delimited text files. For File > Open and the CSV dialog, the first row is normally a header with column names. For the load and import commands, the same applies by default (header=true). If your file has no header row, pass header=false and refer to columns by 0-based index instead of names.

Default delimiter: , (comma)
Requires a target parameter in the load command to define which columns represent edge endpoints
Additional columns can be imported as node or edge attributes

`.tsv` — Tab-Separated Values

Same as CSV but with tab characters as the default delimiter.

Default delimiter: \t (tab)
All other rules are the same as CSV

Delimiter Override

Both load and import accept a delimiter parameter to override the default:

// Use semicolons
load data.csv target=(a, b, name) delimiter=";"

// Explicit tab for a .csv file
load data.csv target=(a, b, name) delimiter=tab

Attributes

Graph objects (nodes and edges) carry typed attributes that power labelling, filtering with the query language, and data-driven styling with the map command.

Supported Types

Type	Storage	Description	Use Cases
float	32-bit float	Floating-point numbers	Scores, weights, continuous values
int32	Signed 32-bit	Signed integers	Counts, ranks, signed identifiers
uint32	Unsigned 32-bit	Unsigned integers	Node IDs, categories, flags
string	Variable-length	Per-object string	Names, labels, unique text
shared_string	Interned	Shared/interned string	Categories, groups (memory-efficient for repeated values)

Usage by Feature

Feature	Accepted Types	Details
map (data-driven styling)	float, int32, uint32	Only numeric attributes can drive visual mappings
set label content	string, shared_string	Only string attributes can be used as label text sources
Query expressions	All types	Numeric attributes compare against numbers; string attributes compare against quoted strings
import	All types	Any type can be imported from CSV columns

Specifying Types

Attribute types are specified when loading or importing data. They appear as the middle field in attribute tuples:

// In the attributes parameter: (column_name, type, attribute_name)
load data.csv target=(a, b, name) attributes=[(weight, float, weight), (category, shared_string, category)]

← Query Language