Skip to content

Instantly share code, notes, and snippets.

@hghwng

hghwng/opmlconv.py

Created August 7, 2016 12:21
Show Gist options
  • Save hghwng/324cc28b007a8f650ce3aac5df099ef8 to your computer and use it in GitHub Desktop.
Save hghwng/324cc28b007a8f650ce3aac5df099ef8 to your computer and use it in GitHub Desktop.
Download ZIP
Convert Dynalist flavored OPML to Org Mode
Raw
opmlconv.py
#!/usr/bin/env python
import bs4
def convert_element(lines, level=1):
result = ''
for line in lines:
if not isinstance(line, bs4.element.Tag) or \
line.name != 'outline':
continue
result += '*' * level + ' ' + line.attrs.get('text', '') + '\n'
if 'note' in line.attrs:
result += line.attrs['note'].replace('\r', '\n') + '\n'
result += convert_element(line.children, level + 1)
return result
def convert_file(path):
root = bs4.BeautifulSoup(open(path), "lxml")
return convert_element(root.select('html body opml')[0])
def main():
import sys
output_path = sys.argv[1][:-4] + 'org'
result = convert_file(sys.argv[1])
open(output_path, 'w').write(result)
if __name__ == '__main__':
main()
@d4h0
Copy link

d4h0 commented Jul 7, 2020

I have extended this script to convert inline Markdown formating (links, bold text, and so on) to Org Mode format.

The conversion is done by pandoc via pypandoc (which you both have to install, see here).

Warning: This script is extremely slow and takes ages to finish if you have a decent amount of content in your Dynalist files.

The reason is, that pypandoc spawns a new process for every invocation (i.e. it's not a binding via a C-API or so), and we have to call pypandoc for every headline and for every note (I tried several approaches to prevent this, but it wasn't possible easily).

I have used the following command with GNU parallel to run several instances of the script in parallel:

find . -name "*.opml" | parallel --bar --eta python convert.py {}

I'm using Linux, so I don't know if these dependencies are available on MacOS or Windows.

Here is the altered script:

#!/usr/bin/env python
import bs4
import pypandoc

def convert_element(lines, level=1):
    result = ''
    for line in lines:
        if not isinstance(line, bs4.element.Tag) or \
           line.name != 'outline':
            continue
        result += '*' * level + ' ' + md_to_org(line.attrs.get('text', '')) + '\n'
        if '_note' in line.attrs:
            result += md_to_org(line.attrs['_note'].replace('\r', '\n')) + '\n'
        result += convert_element(line.children, level + 1)
    return result

def md_to_org(md):
    return pypandoc.convert_text(md, "org", format="md")

def convert_file(path):
    content = open(path, "r").read()
    root = bs4.BeautifulSoup(content, "lxml")
    return convert_element(root.select('html body opml')[0])

def main():
    import sys
    print("Converting " + sys.argv[1])
    output_path = sys.argv[1][:-4] + 'org'
    result = convert_file(sys.argv[1])
    result = result.replace("\n\n*", "\n*")
    open(output_path, 'w').write(result)

if __name__ == '__main__':
    main()

@seanf
Copy link

seanf commented Sep 25, 2024

Thanks @hghwng and @d4h0!

I needed bs4 and lxml too, thus pip install pypandoc bs4 lxml. (Plus brew install pandoc parallel on macOS.)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment