HTML Serializer Parser

Based on html-to-json library https://pypi.org/project/html-to-json/ this library extends its functionality adding an additional layer for extra information like: query selector for every node, list of all query selectors, different return options, by list, by tree dictionary and/or by dict, if adds an specific property for every node

Quick Start

import json
from html2json.parser import ParserOptions, html2json


if __name__ == "__main__":
    # You can use an HTML file, raw HTML String or and endpoint
    FILE_DIR = "./PATH_TO_YOUR_FILES/index.html"
    output = html2json(
        input_path=open(FILE_DIR).read(),
        options=ParserOptions.parser_factory(
            store_as_list=True,
            store_as_dict=True,
            store_as_tree_dict=True
        ),
        raw_content=False,
    )
    # Will retrieve a dict with following keys
    # as_list, as_dict, as_tree_dict, query_selectors
    json_output = json.dumps(output, ensure_ascii=True, indent=2)
    with open("data.json", "w") as o:
        o.write(json_output)

Some code examples

Extracting query selectors

from html2json.parser import ParserOptions, html2json
VALID_SCRAPPING_SITE = "https://www.scrapethissite.com/pages/simple/"

def get_query_selectors():
    output = html2json(
        input_path=VALID_SCRAPPING_SITE,
        options=ParserOptions.parser_factory(False, False, False),
        raw_content=False,
    )
    print(output['query_selectors'])

get_query_selectors()

Let's create a block using the internal logic

If you wanna go deeper, maybe you want to build the logic by yourself replicating the functionality of html2jsonfunction

import requests
from bs4 import BeautifulSoup
from html2json.parser import ParserOptions, Html2JsonParser
VALID_SCRAPPING_SITE = "https://www.scrapethissite.com/pages/simple/"

def step_by_step_usage():
    # We get the HTML content using Requests
    html_content = requests.get(VALID_SCRAPPING_SITE).content
    # We instantiate a BeautifulSoup object using the content
    soup_instance = BeautifulSoup(html_content, 'html.parser')
    # We get a Html2JsonParser instance, prepared to return the
    # data in a List, injecting a BeautifulSoup instance
    parser = Html2JsonParser(
        soup_instance=soup_instance,
        **ParserOptions.parser_factory(
            store_as_list=True,
            store_as_dict=False,
            store_as_tree_dict=False
        ).as_dict()
    )
    # We have to tell the parser that we want to process the information
    # if you don't call this method, the content won't be processed
    parser.process_parser()

    # We extract the query selectors that we collected before
    query_selectors = parser.query_selectors
    # We extract the list of all the nodes that we collected before
    process_list = parser.as_list()

    # We return the data
    return query_selectors, process_list

# Using the Jupyter blocks we will show the 5 first nodes that we collected before
_, process_list = step_by_step_usage()
print(process_list[:5])

Let's use the library in a real case using Pandas

In this example we will extract the information with the library, load the dictionary into a Pandas dataframe, apply some filters, and then store the information in CSV format

import requests
from bs4 import BeautifulSoup
import pandas as pd
from html2json.parser import ParserOptions, Html2JsonParser
VALID_SCRAPPING_SITE = "https://www.scrapethissite.com/pages/simple/"

def get_pandas_info():
    # We get the HTML content using Requests
    html_content = requests.get(VALID_SCRAPPING_SITE).content
    # We instantiate a BeautifulSoup object using the content
    soup_instance = BeautifulSoup(html_content, 'html.parser')
    # We get a Html2JsonParser instance, prepared to return the
    # data in a List, injecting a BeautifulSoup instance
    parser = Html2JsonParser(
        soup_instance=soup_instance,
        **ParserOptions.parser_factory(
            store_as_list=True,
            store_as_dict=False,
            store_as_tree_dict=False
        ).as_dict()
    )
    # We have to tell the parser that we want to process the information
    # if you don't call this method, the content won't be processed
    parser.process_parser()

    # We extract the list of all the nodes that we collected before
    process_list = parser.as_list()

    # Based on the process list we build a Pandas DataFrame
    df = pd.DataFrame(process_list)

    # We know that, all the information that we want, is in an h3 tag
    # we will apply a filter just for having all the information that
    # we really want
    df = df[df['tag'] == 'h3']

    # We are going to reduce the columns that we really want
    # we obtain from the JSON: node_id, tag, children, attrs, query_selector
    # and content. we will store only node_id, tag and content
    df = df[['node_id', 'tag', 'content']]

    # We will return the generated CSV using Pandas method
    return df.to_csv()

csv_data = get_pandas_info()
print(csv_data)

Changelog

• 0.0.3
  • Modified internal logic for allowing dependency injection
  • If a BeautifulSoup object is injected html_content is not required
  • If a BeautifulSoup object is injected library won't analyze html_content because is None
  • If a BeautifulSoup object is not injected and html_content is not provided it will raise an Html2JsonEmptyBody exception

TODOS

• Improve Readme to be easier to understand
• Improve abstractions in order to be easier to modify specific steps
• Add docstrings
• Include more tests
• Avoid repeating node content extending bs4.element.Tag class