bpb-kit/home/bin/notes-collect-metadata

#!/usr/bin/env python3

import os
import re
import sys
import sqlite3
from panflute import *

def resolve_target(target, page):

    if re.search('https?://|file:|tel:|mailto:', target):
        return target

    # At this point, we should be fairly confident the link is a wiki page.
    # Normalize it by removing the .html extension that pandoc throws on there:
    if target.endswith('.html'):
        target = os.path.splitext(target)[0]

    # Check for an absolute path (within the wiki):
    if target.startswith('/'):
        return target.replace('/', '', 1);

    page_elements = page.split('/')
    # Get rid of the page name:
    page_elements.pop()

    path_elements = page_elements + target.split('/')

    resolved_path = []
    while len(path_elements) > 0:
        el = path_elements.pop()
        if el == '..' and len(path_elements) > 0:
            # Discard a directory:
            path_elements.pop()
        else:
            resolved_path.append(el)

    resolved_path.reverse()
    return '/'.join(resolved_path)

def extract_values(elem, doc):

    if isinstance(elem, Link):
        link_target = elem.url

        # Skip in-page anchors, for now:
        if link_target.startswith('#'):
            return;

        # Insert a row of data
        c.execute(
            "INSERT OR IGNORE INTO links VALUES (?, ?)",
            (
                pagename,
                resolve_target(link_target, pagename)
            )
        )

        conn.commit()

# Ensure we're in the wiki directory:
notes_dir = os.path.join(os.getenv('HOME'), 'notes')
vimwiki_dir = os.path.join(notes_dir, 'vimwiki')
os.chdir(notes_dir)

conn = sqlite3.connect('metadata.db')
c = conn.cursor()

for input_file in sys.argv[1:]:
    # Trim leading dir and .wiki:
    # XXX: This is such hacky garbage, jiminy:
    pagename = input_file.replace('./vimwiki/', '', 1)
    pagename = pagename.replace(vimwiki_dir + '/', '', 1)
    pagename = os.path.splitext(pagename)[0]

    with open(input_file) as page:
        doc = convert_text(
            page.read(),
            input_format='vimwiki',
            standalone=True
        )

    title = doc.get_metadata('title')
    date = doc.get_metadata('date')

    # Log the name and metadata of the page:
    c.execute("DELETE FROM pages WHERE page = ?", (pagename,))
    c.execute("INSERT INTO pages VALUES (?, ?, ?)", (pagename, title, date))

    # Clear any links from this page in case something's been deleted:
    c.execute("DELETE FROM links WHERE page = ?", (pagename,))
    doc.walk(extract_values)

conn.close()