Reconstruct a Notepad++ session using an existing session.xml file (without entries or with entries but outdated) and contents of the 'backup' folder with your drafts. named fixshit because I was very displeased when I lost my session and the backup was pretty old. need to set up a cron job fr.

fixshit.py

# run this in the Notepad++ profile directory
# changing these variables is left as an exercise to the reader
# please, do backups of the session.xml
# and of the Notepad++ profile directory too
# just remember that the backups folder might contain sensitive data cuz it has all your drafts in it

# also, install lxml library for this to work

backup_dir = 'backup'

backup_fp = "C:\\users\\User\\AppData\\Roaming\\Notepad++\\backup\\"

session_in = 'session_beforesort.xml'
session_out = 'session.xml'

import lxml
from lxml import etree
from copy import copy
import os

def_itm = [('firstVisibleLine', '0'),
            ('xOffset', '0'),
            ('scrollWidth', '1'),
            ('startPos', '0'),
            ('endPos', '0'),
            ('selMode', '0'),
            ('offset', '0'),
            ('wrapCount', '1'),
            ('lang', 'None (Normal Text)'),
            ('encoding', '-1'),
            ('userReadOnly', 'no'),
            ('filename', ''),
            ('backupFilePath', ''),
            ('originalFileLastModifTimestamp', '0'),
            ('originalFileLastModifTimestampHigh', '0'),
            ('tabColourId', '-1'),
            ('mapFirstVisibleDisplayLine', '-1'),
            ('mapFirstVisibleDocLine', '-1'),
            ('mapLastVisibleDocLine', '-1'),
            ('mapNbLine', '-1'),
            ('mapHigherPos', '-1'),
            ('mapWidth', '-1'),
            ('mapHeight', '-1'),
            ('mapKByteInDoc', '0'),
            ('mapWrapIndentMode', '-1'),
            ('mapIsWrap', 'no')
]

with open(session_in, 'rb') as f:
    c = f.read()

root = etree.XML(c)

# surely unneeded, just cba to rewrite
entries = [f for f in root[0][0]]

# preexisting_files
pexf = []

for entry in entries:
    fn = entry.get('filename')
    fp = entry.get('backupFilePath')
    if fp and '@' in fp:
        af = fp.rsplit('\\', 1)[-1]
        afp, aft = af.split('@')
    else:
        af = ''; afp = ''; aft = ''
    # idk if I need all that but I'm fucking using it
    pexf.append([fn, fp, af, afp, aft])

# o(n^2) woohoo
for filename in os.listdir(backup_dir):
    #print(filename)
    if '@' not in filename:
        raise(filename) #lol what
    fn, ft = filename.rsplit('@', 1)
    # this might have issues processing things like non-'new NNN' files found in backup dir!
    for i, el in enumerate(pexf):
        if el[0] != fn:
            continue
        print(filename, 'found in session file', el)
        elm = root[0][0][i]
        # I don't seem need this, can just replace everything no matter if it differs or not
        # mainly because it won't work if the profile folder path changes, and Notepad++ isn't smart enough to use relative paths for at least the 'new NNN' files weh
        #if el[2] != filename:
        #    print(filename, 'is newer than in session file', el)
        elm.set('backupFilePath', backup_fp + filename)
        #print(backup_fp+filename)
        break
    else:
        print(filename, 'not found in session file')
        el = etree.Element("File")
        for k, v in def_itm:
            el.set(k, v)
        el.set('backupFilePath', backup_fp + filename)
        el.set('filename', fn)
        root[0][0].append(el)

# for some reason, this might leave the 'new NNN' documents unsorted
# and that is infuriating to deal with cuz the sorting is essentially chronological
# but given the numbers in names, easy to fix
# this sorting procedure is written to preserve the order
# of any already-opened documents that are not 'new NNN'
# so that it is comfy to use

# so, we go through the tree again and sort everything

print(len(root[0][0]))

unsorted = []

for i, entry in enumerate(root[0][0]):
    fn = entry.get('filename')
    if fn.startswith('new '):
        num = int(fn.rsplit(' ', 1)[-1])
        unsorted.append([i, num, entry])

ununsorted = sorted(unsorted, key=lambda x: x[1])

for i, e in enumerate(ununsorted):
    _, _, entry = e
    # get the number of the spot that the entry could be taking
    pos = unsorted[i][0]
    # occupy that position
    # apparently, I can't do this: root[0][0][pos] = entry
    # instead, need to recreate the damn element
    el = etree.Element("File")
    for k, v in entry.items():
        el.set(k, v)
    root[0][0][pos] = el

print(len(root[0][0])) # sanity check

#print(etree.tostring(root))
with open(session_out, 'wb') as f:
    f.write(etree.tostring(root))