logseq_helpers/zimwiki_txt_to_logseq_md.py

import os

def rewrite_file(filepath):
    with open(filepath, 'r') as f:
        base = filepath[:-4]
        base = base.replace("_", " ").strip()
        newfilepath = base + ".md"
        with open(newfilepath, 'w') as w:
            in_block = False
            frontmatter = False
            for i, line in enumerate(f):
                if ( (i == 0 and line == "Content-Type: text/x-zim-wiki\n")
                  or (i == 1 and line == "Wiki-Format: zim 0.26\n")
                  or (i == 2 and line == "Creation-Date: Not found\n")
                  or (i == 3 and line == "Modification-Date: Not found\n")
                  or (i == 4 and line == "\n")
                ):
                    continue
                if (i == 5):
                    # Set to always true when we have a type to write below
                    # frontmatter = True
                    title_pieces = line.split("======")
                    if len(title_pieces) == 3:
                        title = title_pieces[1].strip()
                        # If the filename is the same as the title, do not add
                        # the custom property.  (If either this custom property
                        # or the 'actual' title is touched after import, Logseq
                        # will change the filename to that and the custom title
                        # becomes an out-of-date inactive property.)  We set it
                        # if different because the initial import does use it!
                        if (title == base):
                            continue
                        w.write("title:: " + title + "\n")
                        if (frontmatter == False):
                            w.write("\n")
                if (frontmatter == True):
                    # Types of postable-posted, email-draft, email-sent set when running this for each subdirectory
                    w.write("type:: postable-draft\n")
                    w.write("\n")
                    frontmatter = False
                    continue
                if (i > 5 and line == "\"\"\"\n"):
                    if (in_block == False):
                        # Start our self-styled quotation block with a bullet.
                        line = "- " + line
                        in_block = True
                    else:
                        in_block = False
                        # Close our self-styled quotation block without a bullet.
                elif (i > 5 and in_block == False):
                    test = line.strip()
                    if (test == ""):
                        continue
                    prefix = ""
                    temp = line.lstrip()
                    if (temp[0:2] == "* " or temp[0:2] == "- "):
                        temp = temp[2:]
                        prefix = prefix + "\t"
                    elif (line[0:1] == "\t" or line[0:2] == "  "):
                        # If the source file has two indents without first
                        # having a single indent, Logseq does not blink, so we
                        # do two indents if we think it *might* be warranted
                        # without fear.
                        prefix = prefix + "\t"
                    if (line[0:2] == "\t\t" or line[0:4] == "    "):
                        prefix = prefix + "\t"
                    line = prefix + "- " + temp

                w.write(line)



for filepath in os.listdir():
    if filepath.endswith(".txt"):
        rewrite_file(filepath)