Example: fruit images from urls

This example shows how a Trelliscope can be created to view images of fruit and their attributes.

We create a dataset of fruit with images and 4 meta attributes: [name, size, weight, color].

Fruit data

import pandas as pd

from trelliscope import Trelliscope


df = pd.DataFrame([
    {
        "name": "apple",
        "size": 1,
        "weight": 3,
        "color": "red",
        "img": "https://upload.wikimedia.org/wikipedia/commons/1/15/Red_Apple.jpg"
    },
    {
        "name": "banana",
        "size": 3,
        "weight": 2,
        "color": "yellow",
        "img": "https://upload.wikimedia.org/wikipedia/commons/4/44/Bananas_white_background_DS.jpg"
    },
    {
        "name": "pineapple",
        "size": 5,
        "weight": 6,
        "color": "brown",
        "img": "https://upload.wikimedia.org/wikipedia/commons/2/20/Ananas_01.JPG"
    }
])
df

	name	size	weight	color	img
0	apple	1	3	red	https://upload.wikimedia.org/wikipedia/commons...
1	banana	3	2	yellow	https://upload.wikimedia.org/wikipedia/commons...
2	pineapple	5	6	brown	https://upload.wikimedia.org/wikipedia/commons...

Create Trelliscope

Create a Trelliscope object by passing a dataframe and a display name.

Trelliscope will be able to infer what column contains the image, and which columns contain attributes.

display_name = "Fruit"
tr = Trelliscope(df, display_name).write_display()
tr

INFO:root:Using ['name'] to uniquely identify each row of the data.
INFO:root:Saving to /tmp/tmp5hn6n742/fruit
INFO:root:No layout definition supplied. Using Default.
INFO:root:No labels definition supplied. Using Default.
INFO:root:Trelliscope written to `/tmp/tmp5hn6n742/fruit`

A trelliscope display
* Name: Fruit
* Description: Fruit
* Tags: None
* Key columns: ['name']
---
* Path: /tmp/tmp5hn6n742
* Number of panels: 1
---
* Meta Info:

View output

The output can be shown by opening a new browser tab.

tr.view_trelliscope()  # Opens browser tab

A trelliscope display
* Name: Fruit
* Description: Fruit
* Tags: None
* Key columns: ['name']
---
* Path: /tmp/tmp5hn6n742
* Number of panels: 1
---
* Meta Info:

Or you can go to the output directory and open the index.html file. You can configure the output directory by passing Trelliscope(..., path="my-trelliscope/")