Convert date format in file names and contents from Roam Research default style to Obsidian default style.

roam-obsidian-date-convert.py

import re
import os
from dateutil.parser import parse

path = '' #insert file path to your vault here

### Convert date format in file content

for root, dirs, files in os.walk(path):
    files = [f for f in files if re.match(r'.*\.md', f)] # only keep files end with `.md`
    
    #TODO: could better ignore all dirs with `.` like `.git`
    
    for f in files:
        fullpath = (os.path.join(root, f))
        with open(fullpath, 'r') as f: #opens each .md file 
            contents = f.read() #reads the contents
            #substitutes dates with the format [[April 20th, 2020]] for [[2020-04-20]]
            new_contents = re.sub(r'(?<=\[)[\w]+\s\d{1,2}\w{1,2},\s\d{4}(?=\])',
                                  lambda x: str(parse(x.group(0), ignoretz=True)).split(" ")[0], contents, flags=re.M)
        with open(fullpath, 'w') as f:
            f.write(new_contents) #writes the files with the new substitutions 
            
### Convert daily notes names

for root, dirs, files in os.walk(path):
    files = [f for f in files if re.match(r'[\w]+\s\d{1,2}\w{1,2},\s\d{4}\.md', f)]
    for f in files:
        fullpath = (os.path.join(root, f))
        new_fullpath = re.sub(r'[\w]+\s\d{1,2}\w{1,2},\s\d{4}',
                                  lambda x: str(parse(x.group(0), ignoretz=True)).split(" ")[0], fullpath, flags=re.M)
        os.rename(fullpath, new_fullpath)
            

Hi Joilence! Thanks for making this script!! I'd like to convert to ISO format without dashes ('20210220' instead of '2021-02-20'), can you help me with that? Best, Koen

You have to adapt these parts for that:

re.sub(r'(?<=\[)[\w]+\s\d{1,2}\w{1,2},\s\d{4}(?=\])',
                                  lambda x: str(parse(x.group(0), ignoretz=True)).split(" ")[0], contents, flags=re.M)

Worked very well - thank you very much

Hi! I'm getting the following error... using a non-english windows installation, if that matters. Any ideas of how to fix this? I am not very prolific in these things just yet - maybe there's some newbie error behind it.

C:\Users\furemajo>python roam-obsidian-date-convert.py
Traceback (most recent call last):
  File "C:\Users\furemajo\roam-obsidian-date-convert.py", line 17, in <module>
    contents = f.read() #reads the contents
  File "C:\Program Files\Python39\lib\encodings\cp1252.py", line 23, in decode
    return codecs.charmap_decode(input,self.errors,decoding_table)[0]
UnicodeDecodeError: 'charmap' codec can't decode byte 0x9d in position 2232: character maps to <undefined>

What did you try yet? @jeom123

-> https://lmgtfy.app/#gsc.tab=0&gsc.q=UnicodeDecodeError%3A%20'charmap'%20codec%20can't%20decode%20byte%200x9d%20
you should try the first result. Change line 16 to

with open(fullpath, 'r', encoding="utf8") as f:

What did you try yet? @jeom123

-> https://lmgtfy.app/#gsc.tab=0&gsc.q=UnicodeDecodeError%3A%20'charmap'%20codec%20can't%20decode%20byte%200x9d%20 you should try the first result. Change line 16 to

with open(fullpath, 'r', encoding="utf8") as f:

I tried a bunch of random stuff, but focused mainly on adding attributes to the thing that happens on line 17. When adding the encoding="utf8" on line 16, I get:

Traceback (most recent call last): File "C:\Users\furemajo\roam-obsidian-date-convert.py", line 17, in <module> contents = f.read() #reads the contents File "C:\Program Files\Python39\lib\codecs.py", line 322, in decode (result, consumed) = self._buffer_decode(data, self.errors, final) UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf6 in position 1558: invalid start byte

Looking at the files, a lot of dates seem to have changed inside the daily note files (but not the name of the files themselves). Also, all scandinavian characters (I write in Danish in the daily notes) have turned into weird symbols/question marks.

One guess is that Roam has screwed up something when I have written in scandinavian letters. When I open the original exports from Roam in visual studio code, all åäö are highlighted and when I hover them, it says: "The character U+00e4 "ä" is not a basic ASCII character". This is before conversion.

To format date add .strftime('%Y.%m.%d') after the parse

eg.
new_contents = re.sub(r'(?<=\[)[\w]+\s\d{1,2}\w{1,2},\s\d{4}(?=\])', lambda x: str(parse(x.group(0), ignoretz=True).strftime('%Y.%m.%d')).split(" ")[0], contents, flags=re.M)

I'm running this script within Obsidian using the "Execute Code" community plugin on a Mac, and I am getting:
name 're' is not defined
Seems like the import re line is not working?

Any ideas?

Hi Joilence,
I'm running this script within Obsidian using the "Execute Code" community plugin on Win11.
But I'm not getting the 'run' button.
Is there another way to use the script?
What might I be doing wrong?

Hi Joilence,
I'm a total newbie and I can't work out how to run the script. I've installed and activated the Execute Code plug-in. I've copy-posted your code into a note, but the "Run" button doesn't show up. Do I need to somehow indicate in the note that this is a code? Please help! :)

@max-fedoseev just in case you haven't found an answer elsewhere. To run code inside Obsidian you need to create a code block in your note like the following:

```run-python
import re
import os
from dateutil.parser import parse

....

```

Make sure the indentation is correct as you see above

What did you try yet? @jeom123

-> https://lmgtfy.app/#gsc.tab=0&gsc.q=UnicodeDecodeError%3A%20'charmap'%20codec%20can't%20decode%20byte%200x9d%20 you should try the first result. Change line 16 to

with open(fullpath, 'r', encoding="utf8") as f:

Foy my case I found that the markdown files that Roam exported were actually encoded in ANSI.
So on line 17, AND ON LINE 22, rather than

with open(fullpath, 'r', encoding="utf8") as f:

I wrote

with open(fullpath, 'r', encoding="ANSI") as f:

and everything worked fine.

Apparently there are some special characters that are different between the two. I only checked the characters in two files, but in one it was an apostrophe that it didn't like and in another it was a hyphen.