LICA Dataset

The LICA dataset (paper link) is a collection of graphic design layouts, released to promote research in the field of AI for Design. Each layout captures the complete rendering specification of a design — component positions, typography, images, and background — alongside rich natural-language annotations at both the layout and template level.

Layouts are organized by template: a template is a design theme that can produce multiple layout variations (slides). Each template folder contains all of its layouts, rendered images, and per-layout annotations.

Getting started

1. Download the lica-data folder and place it in the root of this repository so the structure looks like:

lica-dataset/
├── lica-data/          # <-- place here
├── lica_dataset.py
├── requirements.txt
└── README.md

2. Install dependencies:

pip install -r requirements.txt

Python 3.9 or later is required.

Dataset structure

lica-data/
├── metadata.csv                              # per-layout metadata
├── layouts/
│   └── <template_id>/
│       └── <layout_id>.json                  # component-level layout spec
├── images/
│   └── <template_id>/
│       └── <layout_id>.png or .mp4           # rendered layout (image or video)
└── annotations/
    ├── template_annotations.json             # template-level annotations
    └── <template_id>/
        └── <layout_id>.json                  # per-layout annotation

metadata.csv

Layout JSON (layouts/<template_id>/<layout_id>.json)

Each layout file is a flat JSON object with the canvas specification and a list of components:

{
  "components": [ ... ],
  "background": "rgb(252, 252, 252)",
  "width": "1920px",
  "height": "1080px",
  "duration": 3
}

Component types

Each component has a type field and CSS-like positioning/visual properties directly on the object.

TEXT — positioned text element

{
  "type": "TEXT",
  "text": "Hello World",
  "left": "108px", "top": "200px", "width": "400px", "height": "50px",
  "color": "rgb(255, 255, 255)",
  "fontSize": "48px",
  "fontFamily": "League Spartan--400",
  "fontWeight": "400",
  "textAlign": "center",
  "lineHeight": "52px",
  "letterSpacing": "0em",
  "textTransform": "none",
  "fontStyle": "normal",
  "transform": "none"
}

IMAGE — positioned image

{
  "type": "IMAGE",
  "src": "https://storage.googleapis.com/lica-video/<uuid>.png",
  "left": "0px", "top": "0px", "width": "1920px", "height": "1080px",
  "transform": "none",
  "opacity": 1,
  "overflow": "hidden"
}

GROUP — container/shape element with optional clip path

{
  "type": "GROUP",
  "left": "108px", "top": "463px", "width": "555px", "height": "508px",
  "background": "rgb(255, 255, 255)",
  "backgroundColor": "rgb(255, 255, 255)",
  "clipPath": "path(\"M0,0 ...\")",
  "transform": "none"
}

Annotations

Per-layout (annotations/<template_id>/<layout_id>.json):

{
  "description": "Visual description of the specific layout",
  "aesthetics": "Notes on design style, composition, visual hierarchy",
  "tags": "comma, separated, keyword, tags",
  "user_intent": "Inferred purpose or goal of the design",
  "raw": "Concatenation of all fields above"
}

Template-level (annotations/template_annotations.json):

A JSON object keyed by template UUID. Each entry has the same fields (description, aesthetics, tags, user_intent, raw) but describes the overall design theme shared by all layouts in the template.

---

Quick start

from lica_dataset import LicaDataset

ds = LicaDataset("lica-data")
print(ds)
# LicaDataset(n=1183, categories=['Business Cards', 'Cards & Invitations', ...])

Inspect categories and metadata

print(ds.categories)
# ['Business Cards', 'Cards & Invitations', 'Education', 'Flyers', ...]

print(ds.templates)
# ['3b919d2e-...', '831589c4-...', ...]

# Full metadata as a Pandas DataFrame
print(ds.metadata.head())

# Per-category summary
print(ds.summary())

Filter layouts

Filtering methods return a new LicaDataset view and are chainable.

# All presentation layouts
presentations = ds.by_category("Presentations")
print(len(presentations))

# All layouts from a specific template
template_layouts = ds.by_template("3b919d2e-539f-4b2c-8d86-7709ef65b496")
print(template_layouts.metadata[["layout_id", "template_layout_index"]])

# Filter by canvas dimensions
widescreen = ds.by_dimensions(1920, 1080)

# Filter by aspect ratio
portrait = ds.by_aspect_ratio("portrait")    # height > width
landscape = ds.by_aspect_ratio("landscape")  # width > height
square = ds.by_aspect_ratio("square")        # width == height

Access individual records

layout_id = "gsessHF2ev5r4ZgwPUh5"

# Load the layout JSON
layout = ds.get_layout(layout_id)
print(layout.get("background"))            # "rgb(252, 252, 252)" or None
print(len(layout["components"]))           # number of components

# Per-layout annotation
annotation = ds.get_annotation(layout_id)
print(annotation["tags"])

# Template-level annotation
tmpl_ann = ds.get_template_annotation("3b919d2e-539f-4b2c-8d86-7709ef65b496")
print(tmpl_ann["description"])

# Path to the render (PNG or MP4, whichever exists)
render_path = ds.get_render_path(layout_id)
print(render_path)
# lica-data/images/3b919d2e-.../gsessHF2ev5r4ZgwPUh5.png

# Single metadata row as a dict
meta = ds.get_metadata(layout_id)
print(meta["category"])  # "Presentations"

Iterate over all items

__getitem__ returns a fully-loaded dict for each layout:

for item in ds:
    layout_id        = item["layout_id"]
    template_id      = item["template_id"]
    metadata         = item["metadata"]             # dict of CSV fields
    layout           = item["layout"]               # layout JSON (or None if not on disk)
    annotation       = item["annotation"]           # per-layout annotation (or None)
    template_ann     = item["template_annotation"]  # template-level annotation (or None)
    render_path      = item["render_path"]          # Path to PNG or MP4

Module-level convenience functions

from lica_dataset import (
    load_dataset,
    load_layouts_by_template,
    load_layouts_by_category,
    iter_template_groups,
)

# Load the dataset (shorthand)
ds = load_dataset("lica-data")

# All layout JSONs for a template, sorted by template_layout_index
layouts = load_layouts_by_template("lica-data", "3b919d2e-539f-4b2c-8d86-7709ef65b496")

# All layout JSONs for a category (only those on disk)
presentation_layouts = load_layouts_by_category("lica-data", "Presentations")

# Iterate over template groups
for template_id, group in iter_template_groups("lica-data"):
    print(template_id, len(group))

---

API reference

LicaDataset

Constructor

Filtering methods (return a new LicaDataset view)

Data access methods

Properties

Other

Module-level functions

License

This project is licensed under the Creative Commons Attribution 4.0 International License.