hyperlink_preview

Purpose

Hyperlink_preview allows getting data needed to display a small visual preview of a http link.
It searches deeper than only <meta property="og: tags. It also parses other tags in <head> or the html <body> if needed to have all data needed for the preview.
It also have an "asynchronous" data getter to avoid having to wait for the full analysis of the images (see below).

Demo

A small demo is provided.
Create an venv, install the package, and test (above code is for windows):

py -3 -m venv venv
venv\Scripts\activate.bat
pip install hyperlink_preview
python -m hyperlink_preview.demo_html https://en.wikipedia.org/wiki/Your_Name

It will:

• get the data to build the preview:

Data for preview:
    title: Your Name - Wikipedia
    type: website
    image: https://upload.wikimedia.org/wikipedia/en/0/0b/Your_Name_poster.png
    url: https://en.wikipedia.org/wiki/Your_Name
    description: Your Name (Japanese: 君の名は。,...
    site_name: en.wikipedia
    domain: en.wikipedia.org

• and open your web browser with an exemple of a preview:

Install

pip install hyperlink_preview

Usage

import hyperlink_preview as HLP

hlp = HLP.HyperLinkPreview(url="https://en.wikipedia.org/wiki/Your_Name")
if hlp.is_valid:
    preview_data = hlp.get_data()
    # Return a dict with keys: ['title', 'type', 'image', 'url', 'description', 'site_name']
    # Values are None or the value for building a preview.

Details

HyperLinkPreview searches for og tags.
If the target link does not provide them (or not all), HyperLinkPreview searches deeper to find suitable data.

About images and performance

If no image is provided, we search for all img tags in the html. Today GIF, PNG and JPG image formats are handled.
We take the sizes of all those images, and we give preference to the largest, and whose ratio is <3 and whose sides are > 50px.
For the sake of efficiency:

• read only bytes necessary to know the dimensions of the images (not the whole image)
• parallelized requests to all the images

However, if the target link contains a lot of pictures, it can take a while (one to several seconds) to do all the requests. A hyperlink preview may need to be displayed quickly (for instance: on mouse hover). In that case:

Get all data except image first, then image

import hyperlink_preview as HLP

hlp = HLP.HyperLinkPreview(url="https://en.wikipedia.org/wiki/Your_Name")
if hlp.is_valid:
    preview_data = hlp.get_data(wait_for_imgs=False)
    # returns as soon as the data are fetched, but don't wait to "parse" all images tags if needed.
    # it allows you to display a spinner as link preview image (or anything else to keep your user waiting).
    

# ... later you can get the remaining image data if needed:
if preview_data["image"] is None:
    preview_data = hlp.get_data(wait_for_imgs=True)