How to use Newspaper3k library without downloading articles?

There is indeed an official way to solve this as mentioned here

Once you've loaded your html in the program you can use the set_html() method to set it to article.html

import newspaper
with open("file.html", 'rb') as fh:
    ht = fh.read()
article = newspaper.Article(url = ' ')
article.set_html(ht)
article.parse()

You can, it's just a bit hacky. As an example

import requests
from newspaper import Article

url = 'https://www.cnn.com/2019/06/19/india/chennai-water-crisis-intl-hnk/index.html'

# get sample html
r = requests.get(url)

# save to file
with open('file.html', 'wb') as fh:
    fh.write(r.content)

a = Article(url)

# set html manually
with open("file.html", 'rb') as fh:
    a.html = fh.read()

# need to set download_state to 2 for this to work
a.download_state = 2

a.parse()

# Now the article should be populated
a.text

# 'New Delhi (CNN) The floor...'

Where the download_state comes from the snippet in newspaper.article.py:

# /path/to/site-packages/newspaper/article.py
class ArticleDownloadState(object):
    NOT_STARTED = 0
    FAILED_RESPONSE = 1
    SUCCESS = 2

~snip~

# This is why you need to set that variable
class Article:
    def __init__(...):
        ~snip~
         # Keep state for downloads and parsing
        self.is_parsed = False
        self.download_state = ArticleDownloadState.NOT_STARTED
        self.download_exception_msg = None

    def parse(self):
        # will throw exception if download_state isn't 2
        self.throw_if_not_downloaded_verbose()

        self.doc = self.config.get_parser().fromstring(self.html)

As an alternative, you could override the class to act just the same with the parse function:

from newspaper import Article
import io

class localArticle(Article):
    def __init__(self, url, **kwargs):
        # set url to be file_name in __init__ if it's a file handle
        super().__init__(url if isinstance(url, str) else url.name, **kwargs)
        # set standalone _url attr so that parse will work as expected
        self._url = url

    def parse(self):

        # sets html and things for you
        if isinstance(self._url, str):
            with open(self._url, 'rb') as fh:
                self.html = fh.read()

        elif isinstance(self._url, (io.TextIOWrapper, io.BufferedReader)):
            self.html = self._url.read()

        else:
            raise TypeError(f"Expected file path or file-like object, got {self._url.__class__}")

        self.download_state = 2
        # now parse will continue on with the proper params set
        super(localArticle, self).parse()


a = localArticle('file.html') # pass your file name here
a.parse()

a.text[:10]
# 'New Delhi '

# or you can give it a file handle
with open("file.html", 'rb') as fh:
    a = localArticle(fh)
    a.parse()

a.text[:10]
# 'New Delhi '

I'm sure that you have solved this, but Newspaper has the capabilities to process locally stored HTML files.

from newspaper import Article

# Downloading the HTML for the article
url = 'http://fox13now.com/2013/12/30/new-year-new-laws-obamacare-pot-guns-and-drones/'
article = Article(url)
article.download()
article.parse()
with open('fox13no.html', 'w') as fileout:
   fileout.write(article.html)

# Read the locally stored HTML with Newspaper
with open("fox13no.html", 'r') as f:
   # note the URL string is empty
   article = Article('', language='en')
   article.download(input_html=f.read())
   article.parse()
   print(article.title) 
   New Year, new laws: Obamacare, pot, guns and drones