parsera 0.1.8

Lightweight library for scraping web-sites with LLMs

📦 Parsera

Lightweight Python library for scraping websites with LLMs. You can test it on Parsera website.

Why Parsera?

Because it's simple and lightweight, with minimal token use which boosts speed and reduces expenses.

Table of Contents

• Installation
• Documentation
• Basic usage
• Running with Jupyter Notebook
• Running with CLI
• Running in Docker

Installation

pip install parsera
playwright install

Documentation

Check out documentation to learn more about other features, like running custom models and playwright scripts.

Basic usage

If you want to use OpenAI, remember to set up OPENAI_API_KEY env variable. You can do this from python with:

import os

os.environ["OPENAI_API_KEY"] = "YOUR_OPENAI_API_KEY_HERE"

Next, you can run a basic version that uses gpt-4o-mini

from parsera import Parsera

url = "https://news.ycombinator.com/"
elements = {
    "Title": "News title",
    "Points": "Number of points",
    "Comments": "Number of comments",
}

scraper = Parsera()
result = scraper.run(url=url, elements=elements)

result variable will contain a json with a list of records:

[
   {
      "Title":"Hacking the largest airline and hotel rewards platform (2023)",
      "Points":"104",
      "Comments":"24"
   },
    ...
]

There is also arun async method available:

result = await scrapper.arun(url=url, elements=elements)

Running with Jupyter Notebook:

Either place this code at the beginning of your notebook:

import nest_asyncio
nest_asyncio.apply()

Or instead of calling run method use async arun.

Running with CLI

Before you run Parsera as command line tool don't forget to put your OPENAI_API_KEY to env variables or .env file

Usage

You can configure elements to parse using JSON string or FILE. Optionally, you can provide FILE to write output.

python -m parsera.main URL {--scheme '{"title":"h1"}' | --file FILENAME} [--output FILENAME]

Running in Docker

In case of issues with your local environment you can run Parsera with Docker, see documentation.