Script to back up dropbox paper

paperback

#!/usr/local/bin/python3

# Backup dropbox paper.
# This expects a ~/.auth/dropbox file that contains your dropbox bearer token.

import requests, os, json, re
from pprint import pprint
oj = os.path.join


class DBox(object):
    def __init__(self):
        self.base = "https://api.dropboxapi.com/2"
        self.key = open(os.path.expanduser('~/.auth/dropbox')).read().strip()
        self.headers = {'Authorization': f"Bearer {self.key}"}

    def post(self, path, **args):
        if path[0] != '/':
            path = '/'+path
        r = requests.post(self.base+path, json=args, headers=self.headers)
        if not r.ok:
            print(r.text)
            assert r.ok
        return r.json()

    def download(self, path, **args):
        headers = dict(self.headers)
        headers['Dropbox-API-Arg'] = json.dumps(args)
        r = requests.post(self.base+path, headers=headers)

        if not r.ok:
            print(r.headers)
            print(r.text)
            assert r.ok

        meta = json.loads(r.headers['Dropbox-Api-Result'])
        return meta, r.content


def sync(db,key):
    # REVIEW - maybe backup HTML too?
    base = os.path.expanduser('~/Dropbox/Backup/paper')
    meta,md = db.download('/paper/docs/download', doc_id=key, export_format='markdown')
    _,html = db.download('/paper/docs/download', doc_id=key, export_format='html')
    os.makedirs(oj(base,key,'media'), exist_ok=True)
    mf = oj(base,key,'meta.json')
    if os.path.exists(mf):
        prev = json.load(open(mf))
        print(key, meta['revision'], 'previous', prev['revision'], repr(meta['title']))
        if prev['revision'] == meta['revision']:
            return False
    print('write',key,meta.get('title'))
    open(oj(base,key,'content.md'),'wb').write(md)
    open(oj(base,key,'content.html'),'wb').write(html)
    json.dump(meta,open(mf,'w'),indent=True)
    saveinfo(db,key)
    return True

def saveinfo(db,key):
    print("save info",key)    
    info = db.post('/paper/docs/get_folder_info',doc_id=key)
    base = os.path.expanduser('~/Dropbox/Backup/paper')
    infofn = oj(base,key,'folder.json')
    json.dump(info,open(infofn,"w",encoding="utf8"), indent=True)


def getmedia(key):
    base = os.path.expanduser('~/Dropbox/Backup/paper')
    infofn = oj(base,key,'folder.json')
    if not os.path.exists(infofn):
        saveinfo(db,key)

    md = open(oj(base,key,'content.md'),encoding='utf8').read()
    for m in re.finditer(r'!\[.*?\]\((.*?)\)', md):
        url = m.group(1)
        bn = os.path.basename(url)
        bn = bn.split('?')[0]
        fn = oj(base,key,'media',bn)
        if not os.path.exists(fn):
            print(key, bn, url)
            if '://' not in url:
               print("bad url",url)
               continue
            r = requests.get(url)
            assert r.ok
            open(fn,'wb').write(r.content)




if __name__ == '__main__':
    import glob, subprocess

    # Backup dropbox paper

    db = DBox()
    # If you have a lot of docs, you'll need to deal with the continuation stuff here
    docs = db.post('/paper/docs/list',sort_by="modified", sort_order='descending')
    for key in docs['doc_ids']:
        if not sync(db,key):
            break

    base = os.path.expanduser('~/Dropbox/Backup/paper')
    for fn in glob.glob(os.path.expanduser('~/Dropbox/Backup/paper/*')):
        if os.path.isdir(fn):
            key = os.path.basename(fn)
            getmedia(key)
            if not os.path.exists(oj(base,key,'content.html')) and not os.path.exists(oj(base,key,'DELETED')):
                print("backfill html",key)
                try:
                    _,html = db.download('/paper/docs/download', doc_id=key, export_format='html')
                    open(oj(base,key,'content.html'),'wb').write(html)
                except:
                    open(oj(base,key,'DELETED'),'wb')
    
    os.chdir(os.path.expanduser("~/Dropbox/backup/paper"))
    subprocess.run("git add .".split())
    subprocess.run("git commit -am checkpoint".split())