cohost-post-json-to-markdown.py

#!/usr/bin/env python
"""
Create a reasonable Markdown export from a Cohost post JSON in your exported profile

Usage: cohost-post-json-to-markdown.py './project/bitto/posts/published/123456-example/post.json' '/home/bitto/cohost-exports'
Post will be written to '/home/bitto/cohost-exports/123456.markdown'
"""

import json
import os
import sys

__version = "0.1"
__author = "bitto"
__license = "MIT"

IMAGE_TEMPLATE = '![{1}]({0})'

AUDIO_TEMPLATE = '''
### {1} - {2}
<audio controls>
    <source src="{0}" />
</audio>
'''

def convert_attachment_row(block: json):
    attachments = block["attachments"]
    return [convert_attachment(attachment) for attachment in attachments]

def convert_attachment(block: json):
    attachment = block["attachment"]
    kind = attachment.get("kind")
    if kind == "image":
        src = attachment["fileURL"]
        alt = attachment["altText"]
        return IMAGE_TEMPLATE.format(src, alt)
    elif kind == "audio":
        src = attachment["fileURL"]
        artist = attachment["artist"]
        title = attachment["title"]
        return AUDIO_TEMPLATE.format(src, artist, title)
    else:
        raise Exception(f"Attachment type not supported: {kind}")

if __name__ == "__main__":
    if not len(sys.argv) == 3:
        sys.exit(f"Usage: {sys.argv[0]} [post_json] [out_dir]")

    post_json_path = sys.argv[1]
    out_dir = sys.argv[2]

    post_json = json.load(open(post_json_path))
    post_id = post_json["postId"]
    post_title = post_json["headline"]
    post_date = post_json["publishedAt"]

    if not os.path.exists(out_dir):
        os.makedirs(out_dir, exist_ok=True)
    out_path = os.path.join(out_dir, f"{post_id}.markdown")
    sys.stdout = open(out_path, "wt")

    # Jekyll/Hugo/etc header
    print("---")
    print(f"title: {post_title}")
    print(f"date: {post_date}")
    print("---\n")

    post_blocks = post_json["blocks"]
    for block in post_blocks:
        block_type = block.get("type")
        if not block_type:
            raise Exception(f"No block type found for block in post: {post_title}")
        elif block_type == "markdown":
            print(block["markdown"]["content"])
        elif block_type == "attachment":
            print(convert_attachment(block))
        elif block_type == "attachment-row":
            print(*convert_attachment_row(block), sep="\n")
        else:
            raise Exception(f"Block type not supported: {block_type}")
        print() # Line break