spellbook/jrnl-merge/jrnlFix.py

145 lines
4.8 KiB
Python
Raw Permalink Normal View History

2022-03-21 18:53:24 -07:00
# some code imported from https://github.com/ng110/TiddlPy/blob/master/TiddlPy/TiddlPy.py
import re
import shutil
from time import strftime, gmtime
import json
from bs4 import BeautifulSoup
from os import listdir
from os.path import isfile, join
def tid2dict(tiddler):
"""
tid2dict(tiddler)
Converts tiddler as a BeautifulSoup object into a Python dictionary
"""
tiddict = tiddler.attrs
tiddict[u'text'] = tiddler.text
return tiddict
def openTiddlers(file):
twsoup = BeautifulSoup(file, 'html.parser')
storearea = twsoup.find('div', id="storeArea")
tiddlers = storearea('div')
ori_matched_tiddlers = []
for tid in tiddlers:
# this is wrong, it only keeps the last/final tiddler it reads
ori_matched_tiddlers.append(tid2dict(tid))
return ori_matched_tiddlers
def wikiedit(wiki, tiddlers, deletelist, modi=u'python'):
# Try and figure out how this works so you can debug it
# so far it's pretty robust
"""
wiki is the wikifilename.
Deletes tiddlers in the deletelist and the tiddlers list, if they currently
exist, then writes the tiddlers list, adding/replacing the modifier
attribute with the specified or default value. Timestamps modified and/or
created dates in UTC.
Returns (writtenlist, deletedlist) - list of tiddlers successfully deleted
and written.
"""
encoding = 'utf-8'
savelist = []
writtenlist = []
deletedlist = []
created = {}
for tid in tiddlers:
savelist.append(tid[u'title'])
with open(wiki, 'br') as fhi, open('temp__.html', 'bw') as fho:
for line in fhi:
if re.search('^<div id="storeArea"', line.decode(encoding)):
fho.write(line)
break
fho.write(line)
else:
print('Store area not found')
raise SystemError
enddiv = False
for line in fhi:
if re.search(b'^<div', line):
enddiv = False
if len(re.findall(b'title=".*?"', line)) < 1:
print('error - no title in line: ', line)
raise SystemError
title = (re.findall(b'title=".*?"', line)
[0][7:-1]).decode(encoding)
if len(title) <= 0:
print('title not found in tiddler', line)
raise SystemError
if title not in deletelist and title not in savelist:
fho.write(line)
continue
deletedlist.append(title)
if re.search(b'created=".*?"', line):
createddate = re.findall(b'created=".*?"', line)[0][9:-1]
created[title] = createddate
for line2 in fhi:
if re.search(b'^</div>', line2):
enddiv = True
break
elif re.search(b'^</div>', line):
if enddiv:
break
else:
fho.write(line)
enddiv = True
else:
fho.write(line)
else:
print('end of Store area not found')
raise SystemExit
for tiddler in tiddlers:
tiddler['modifier'] = modi
tiddler['modified'] = strftime('%Y%m%d%H%M%S', gmtime())
if tiddler['title'] not in deletedlist:
tiddler['created'] = tiddler['modified']
elif tiddler['title'] in created:
tiddler['created'] = created[tiddler['title']].decode(encoding)
fho.write(b'<div')
for key in tiddler:
if key == 'text':
continue
fho.write(b' ' + bytes(key, encoding)
+ b'="' + bytes(tiddler[key], encoding) + b'"')
fho.write(
'>\n<pre>{}</pre>\n</div>\n'.format(tiddler['text']).encode(encoding))
writtenlist.append(tiddler['title'])
fho.write(line) # delayed write of </div> line at end of storearea
for line in fhi:
fho.write(line)
# end 'with'
shutil.copyfile('temp__.html', wiki)
return (writtenlist, deletedlist)
originalTiddlers = {}
conflictTiddlers = {}
with open('./jrnls/jrnl.html') as file:
originalTiddlers = openTiddlers(file)
syncConflicts = [f for f in listdir(
'./jrnls/') if isfile(join('./jrnls/', f)) and 'sync' in f]
writeList = []
if len(syncConflicts) > 0:
for conflict in syncConflicts:
json = openTiddlers(open(join('./jrnls/', conflict), 'r'))
for e in json:
if not e in originalTiddlers:
if not e['title'][0] == '$':
writeList.append(e)
wikiedit('./jrnls/jrnl.html', writeList, [])
else:
print("Nothing to do")