Python : How to convert markdown formatted text to text

Question

I need to convert markdown text to plain text format to display summary in my website. I want the code in python.

Answer 1

Despite the fact that this is a very old question, I'd like to suggest a solution I came up with recently. This one neither uses BeautifulSoup nor has an overhead of converting to html and back.

The markdown module core class Markdown has a property output_formats which is not configurable but otherwise patchable like almost anything in python is. This property is a dict mapping output format name to a rendering function. By default it has two output formats, 'html' and 'xhtml' correspondingly. With a little help it may have a plaintext rendering function which is easy to write:

from markdown import Markdown
from io import StringIO


def unmark_element(element, stream=None):
    if stream is None:
        stream = StringIO()
    if element.text:
        stream.write(element.text)
    for sub in element:
        unmark_element(sub, stream)
    if element.tail:
        stream.write(element.tail)
    return stream.getvalue()


# patching Markdown
Markdown.output_formats["plain"] = unmark_element
__md = Markdown(output_format="plain")
__md.stripTopLevelTags = False


def unmark(text):
    return __md.convert(text)

unmark function takes markdown text as an input and returns all the markdown characters stripped out.

Answer 2

The Markdown and BeautifulSoup (now called beautifulsoup4) modules will help do what you describe.

Once you have converted the markdown to HTML, you can use a HTML parser to strip out the plain text.

Your code might look something like this:

from bs4 import BeautifulSoup
from markdown import markdown

html = markdown(some_html_string)
text = ''.join(BeautifulSoup(html).findAll(text=True))

Answer 3

This is similar to Jason's answer, but handles comments correctly.

import markdown # pip install markdown
from bs4 import BeautifulSoup # pip install beautifulsoup4

def md_to_text(md):
    html = markdown.markdown(md)
    soup = BeautifulSoup(html, features='html.parser')
    return soup.get_text()

def example():
    md = '**A** [B](http://example.com) <!-- C -->'
    text = md_to_text(md)
    print(text)
    # Output: A B

Answer 4

Commented and removed it because I finally think I see the rub here: It may be easier to convert your markdown text to HTML and remove HTML from the text. I'm not aware of anything to remove markdown from text effectively but there are many HTML to plain text solutions.

Answer 5

It's not necessarily a tremendously fast solution in my limited experience, but you might try the MarkdownCorpusReader from NLTK. It requires a directory full of markdown files and a regular expression for valid file IDs.

from nltk.corpus.reader.markdown import MarkdownCorpusReader
from nltk.tokenize.treebank import TreebankWordDetokenizer
# You might also need a punkt detokenizer for the English language.

filepath = './some/path/here' 
reader = MarkdownCorpusReader(filepath, r'[w\]*\.md')

def get_text(reader: MarkdownCorpusReader, fileid: str) -> str:
    tokens = reader.words(fileids=fileid)
    # You might also need a punkt detokenizer for the English language.
    return TreebankWordDetokenizer().detokenize(tokens)

Unfortunately there are variatons on markdown, so depending where it's coming from some of the formatting elements may still be present. I can't fully test this because I don't have example data to work on. You might also need a punkt detokenizer for English. I'm not intimately familiar with the default tokenization used here, but I presume it is nltk.tokenize.word_tokenize, which uses a combination of a treebank tokenizer + an english language punkt tokenizer.

I'll add that nlkt's markdown reader is built on markdown-it-py and mdit-plain, so presumably there exist tools within those modules as well to help with this.

Answer 6

As Neil suggested, nltk's parser is based on markdown-it and mdit-plain. It's quite easy to use those directly (no BeautifulSoup needed!).

pip install markdown-it-py mdit_plain

from markdown_it import MarkdownIt
from mdit_plain.renderer import RendererPlain

parser = MarkdownIt(renderer_cls=RendererPlain)

md_data = "# some markdown"
txt_data = parser.render(md_data)

Answer 7

I came here while searching for a way to perform s.c. GitLab Releases via API call. I hope this matches the use case of the original questioner.

I decoded markdown to plain text (including whitespaces in the form of \n etc.) in that way:

    with open("release_note.md", 'r') as file:
        release_note = file.read()
        description = bytes(release_note, 'utf-8')
    return description.decode("utf-8")