README.md

metadata

task_categories:
  - text-generation
  - feature-extraction
language:
  - he
tags:
  - cooking
  - art
  - recipes
  - schema.org
  - json-ld
pretty_name: Hebrew Recipes
size_categories:
  - 1K<n<10K

Dataset Card for Hebrew Recipes

A dataset of recipes scraped from the Israeli recipe websites.

Dataset Details

Dataset Description

The dataset contains recipes scraped from multiple Israeli recipe websites, including sugat.com and hashulchan.co.il. It includes structured JSON-LD data conforming to schema.org Recipe specifications, cleaned HTML from the printable recipe views, and various metadata for each recipe URL.

• Curated by: [@Wissotsky]
• Language: [Hebrew]

Dataset Sources

• Sugat Recipes
• Hashulchan Recipes

Dataset Structure

The data is stored in a single Parquet file (recipes.parquet) with the following columns:

• URL (string): The original URL of the recipe page.
• Title (string): The title of the webpage.
• JsonLd (string): The "Recipe" schema.org JSON-LD data, if found on the page. Stored as a JSON string.
• Html (string): The cleaned, printable HTML content of the recipe.
• Sitemap (string): The source sitemap filepath on disk (e.g., sitemaps/recipes-sitemap.xml).
• ScrapeTimestamp (int64): Unix timestamp indicating when the data for the specific URL was scraped.
• JsonLdPresent (bool): A flag that is true if a "Recipe" JSON-LD block was successfully found and extracted.
• HtmlRecipePresent (bool): A flag that is true if the printable HTML block (#print_area) was found.
• HttpStatusCode (int): The HTTP status code returned when scraping the URL (e.g., 200, 404).

Dataset Creation

Data Collection and Processing

A scraper written in golang using the gocolly library. It goes through recipe URLs found in the XML sitemaps, extracting the page title, the "Recipe" JSON-LD block, and the printable recipe HTML from each page. The HTML is cleaned to match the site's print view. The data is then saved into a Parquet file (recipes.parquet) with Zstandard compression.