No project description provided
Project description
🤖 Oxy® Parser
Parse HTMLs automatically by only describing the Pydantic model.
Oxy® Parser does the heavy lifting of parsing HTMLs for you. It uses Pydantic models to describe the structure of the HTML and then automatically parses the HTMLs into the Pydantic models.:
- Describe a Pydantic model of your expected HTML structure
- Pass the URL or HTML to Oxy® Parser together with the Pydantic model
- Oxy® Parser will parse the HTML and return the parsed data as Pydantic models
- Oxy® Parser will also cache the selectors for later re-use, so you don't need to call OpenAI's API every time you want to parse the same HTML
Supported cache backends:
- Memory
- File
- Redis
See the flowchart for detailed view
Installation
pip install oxyparser
Supported LLMs
This project uses LiteLLM please refer to the documentation to see the supported LLMs: https://docs.litellm.ai/docs/providers
Usage
You will need to setup an .env file with the following variables:
LLM_API_KEY: Can be either OpenAI key, Claude key or any other LLM provider key. (See LiteLLM docs above for full info on supported LLMs)
LLM_MODEL: The model you want to use. (See LiteLLM docs above for full info on supported LLMs)
Scraper keys: they're optional, but if not provided, you will need to scrape and pass the HTMLs to Oxy® Parser yourself. However, we highly suggest to use Oxylabs scraper since it will remove the hassle of getting HTMLs and will also provide you with a lot of other benefits like rotating IPs, handling captchas, bypassing blocks.
Click here to sign up for Oxylabs free trial: https://dashboard.oxylabs.io/en/registration?productToBuy=SCRAPI_WEB
OXYLABS_SCRAPER_USER: Your Oxylabs scraper user (optional)
OXYLABS_SCRAPER_PASSWORD: Your Oxylabs scraper password (optional)
LLM_API_KEY=your_openai_api_key
LLM_MODEL=gpt-3.5-turbo
OXYLABS_SCRAPER_USER=your_oxylabs_scraper_user # optional
OXYLABS_SCRAPER_PASSWORD=your_oxylabs_scraper_pass # optional
Then you can use the following code to parse the website into structured data:
For full examples see the examples directory.
from pydantic import BaseModel
from oxyparser.oxyparser import OxyParser
class JobItem(BaseModel):
title: str
recruiter_name: str
location: str
description: str
# this page might expire
# if it does, please replace it with a new one
# https://career.oxylabs.io
# also if you're a python dev and looking for job, hit us up!
URL: str = "https://career.oxylabs.io/job/813b9ac5/python-developer-mid-senior/"
async def main() -> None:
parser = OxyParser()
job_item = await parser.parse(URL, JobItem)
print(job_item)
if __name__ == "__main__":
import asyncio
asyncio.run(main())
If you have a HTML string instead of a URL, you can pass it to the parser as well like so:
parser = OxyParser()
html = "<html><body>" "<h1>John</h1>" "<h2>Smith</h2>" "<p>Svitrigailos st.</p>" "<span>2 years old</span></body></html>"
url = "https://example.com" # url is needed to cache selectors
parsed_item = await parser.parse(url=url, model=JobItem, html=html)
print(job_item)
Known Issues
There are know nuances where sometimes the extracted xpath fails to extract some data. For example where description is very long or nested by many elements. In such cases for now we recommend manually editing the selectors in cache. These shouldn't be an often case though. Create an issue if you encounter any other issues.
Contributing
We welcome all contributions. To contribute - clone the repo, create a new branch, make your changes and create a pull request.
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for oxyparser-0.1.0-py3-none-any.whl
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 923e78a17d28a71eb44b9e84c990c42504ef882edf81648387e3c5de40c924bf |
|
| MD5 | bb9d01c05616de26ec587b1aed807cd9 |
|
| BLAKE2b-256 | 0f7b7fd0924e1703aec14b1258b3cceaecace4fd47062c8c3a16eb336ef4edd3 |
