Skip to main content

perse converts HTML content into structured JSON data

Project description

Perse

PyPI version

Perse

Perse converts HTML to JSON using a mix of traditional html parsing and LLM based data extraction.

Features

It's core features includes:

  • Identify important fields to extract from html
  • Building a JSON schemas that handles nested fields
  • Process html tokens and fill the JSON schema object

You can install Perse using pip:

pip install zf-perse
export PERSE_OPENAI_API_KEY="your-openai-api-key"

And run it from CLI:

perse --url https://google.com

Optimizations

It performs a few optimizations after fetching the html while preventing any accidental removal of important data.

These optimizations includes:

  • Removal of styling, scripting and svg tags
  • Collapsing Tags (e.g. divs) with only one child

Comparison

There are a few other libraries but none of them provide a solution for reliable data extraction from html.

HTML to JSON

html2json library is a simple html to json converter that doesn't handle nested fields, nor does it remove unnecessary tags.

When ran on exactly the same html, Perse provides a more structured and cleaner output and at least 50% less verbose output.

HTML to JSON Perse
rate_1.0 rate_1.0

HTML to Markdown

Reader-LM is a language model that converts html to markdown. It doesn't provide a json output catering only to the reader mode which is not suitabel for data extraction, analysis and automations.

Usage

Process HTML content and get a Dictionary

html_content = "<html>...</html>"
json_dict = perse(html_content)
print(json_dict)

Process HTML content and get a JSON string

html_content = "<html>...</html>"
json_string = perses(html_content)
print(json_string)

Exclude specific tags from the JSON output

html_content = "<html>...</html>"
json_dict = perse(html_content, exclude_tags={"script", "style"})
print(json_dict)

Clean up the HTML content for side usage

html_content = "<html>...</html>"
clean_soup = simmer(html_content) # or use simmers for a string output
print(clean_soup.prettify())

Examples

Google's Homepage

$ perse --url https://google.com

{
  "image": "/images/branding/googleg/1x/googleg_standard_color_128dp.png",
  "title": "Google",
  "search_form": {
    "action": "/search",
    "method": "GET",
    "autocomplete": "off",
    "query": "",
    "buttons": [
      {
        "button_1": {
          "label": "Google Search",
          "value": "Google Search"
        },
        "button_2": {
          "label": "I'm Feeling Lucky",
          "value": "I'm Feeling Lucky"
        }
      }
    ]
  }
}

Zeff Muks's Homepage

$ perse --url https://zeffmuks.com

{
  "title": "Zeff Muks",
  "description": "Antifragile Entropy Assassin \ud83e\udd77",
  "og_data": {
    "type": "website",
    "title": "Zeff Muks",
    "description": "Antifragile Entropy Assassin \ud83e\udd77",
    "url": "https://zeffmuks.com/",
    "image": "https://www.zeffmuks.com/images/ZeffMuks-1920.png",
    "site_name": "Zeff Muks"
  },
  "twitter_data": {
    "card": "summary_large_image",
    "site": "@zeffmuks",
    "title": "Zeff Muks",
    "description": "Antifragile Entropy Assassin \ud83e\udd77",
    "image": "https://www.zeffmuks.com/images/ZeffMuks-1920.png"
  },
  "user_section": {
    "header": {
      "profile_image_url": "/images/ZeffMuks-6912.png",
      "title": "Antifragile Entropy Assassin \ud83e\udd77",
      "signature": ""
    },
    "builds": [
      {
        "date": "08/30/2024",
        "name": "Cursor Git",
        "description": "Enhanced Git for Cursor AI Editor",
        "download_link": "https://zf-static.s3.us-west-1.amazonaws.com/cursor-git-0.1.12.vsix",
        "preview_image": "https://zf-static.s3.us-west-1.amazonaws.com/cursor-git-logo128.png",
        "alternative_link": ""
      },
      {
        "date": "08/18/2024",
        "name": "PyZF",
        "description": "Enhancements for Python",
        "download_link": "https://pypi.org/project/PyZF",
        "preview_image": "https://zf-static.s3.us-west-1.amazonaws.com/pyzf-logo128.png",
        "alternative_link": ""
      },
      {
        "date": "08/05/2024",
        "name": "Xanthus",
        "description": "X (formerly Twitter) Assistant",
        "download_link": "https://pypi.org/project/zf-xanthus",
        "preview_image": "https://zf-static.s3.us-west-1.amazonaws.com/xanthus-logo128.png",
        "alternative_link": ""
      },
      {
        "date": "07/24/2024",
        "name": "Jenga",
        "description": "Fast JSON5 Python Library",
        "download_link": "https://pypi.org/project/zf-jenga",
        "preview_image": "",
        "alternative_link": ""
      },
      {
        "date": "07/12/2024",
        "name": "Pegasus",
        "description": "Next Generation Tech Stack",
        "download_link": "https://zf-static.s3.us-west-1.amazonaws.com/pegasus.zip",
        "preview_image": "https://zf-static.s3.us-west-1.amazonaws.com/pegasus-logo128.png",
        "alternative_link": ""
      },
      ...
      {
        "date": "11/01/2023",
        "name": "Z",
        "description": "Next Generation Content Platform",
        "download_link": "https://x.com/zeffmuks/status/1718507463321010429",
        "preview_image": "https://zf-static.s3.us-west-1.amazonaws.com/z-logo128.png",
        "alternative_link": "https://alpha.thez.ai/try"
      }
    ]
  }
}

License

MIT License

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

zf_perse-1.0.0.tar.gz (11.6 kB view details)

Uploaded Source

Built Distribution

zf_perse-1.0.0-py3-none-any.whl (10.4 kB view details)

Uploaded Python 3

File details

Details for the file zf_perse-1.0.0.tar.gz.

File metadata

  • Download URL: zf_perse-1.0.0.tar.gz
  • Upload date:
  • Size: 11.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.9

File hashes

Hashes for zf_perse-1.0.0.tar.gz
Algorithm Hash digest
SHA256 c17d749de6328c72472659e71e9676a51c3ff998c6b39ddf3df5b69e6d7fdd35
MD5 5eccf48ecce8ff27b000cc81a7c5e575
BLAKE2b-256 983a332ac67a807e28378ab071c077406708f3e5da690c8adba060cf3fb3c674

See more details on using hashes here.

File details

Details for the file zf_perse-1.0.0-py3-none-any.whl.

File metadata

  • Download URL: zf_perse-1.0.0-py3-none-any.whl
  • Upload date:
  • Size: 10.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.9

File hashes

Hashes for zf_perse-1.0.0-py3-none-any.whl
Algorithm Hash digest
SHA256 2beb28cf2e87bc404e842befd90ae177c5fbacc3539bbbc5f2e7d9bbdb8d1899
MD5 af357a680e3a18ca0a7cb70220794d45
BLAKE2b-256 b99e43adf1dee2a44fab778be2cef86f40718251e470aa125da3dcfb94d04d3b

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page