From 1ddf9be13ab16d4f170f592e1b1b454c3ce0877f Mon Sep 17 00:00:00 2001
From: Peter Howell <peter.howell@gmail.com>
Date: Fri, 29 Aug 2025 01:11:06 +0000
Subject: [PATCH] safe folders

---
 util.py | 52 +++++++++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 49 insertions(+), 3 deletions(-)
 mode change 100644 => 100755 util.py

diff --git a/util.py b/util.py
old mode 100644
new mode 100755
index 270e8b9..2e1fbf0
--- a/util.py
+++ b/util.py
@@ -1,16 +1,58 @@
+import os
+import builtins as _builtins
+import codecs as _codecs
 
+# --- Safe file I/O monkey patches ---
+# Ensure parent folders exist for write/append modes when using open/codecs.open
+_orig_open = _builtins.open
+_orig_codecs_open = _codecs.open
 
+def _ensure_parent_dir(path):
+    try:
+        if isinstance(path, (str, bytes, os.PathLike)):
+            d = os.path.dirname(os.fspath(path))
+            if d and not os.path.exists(d):
+                os.makedirs(d, exist_ok=True)
+    except Exception:
+        # Never block the open call due to directory check errors
+        pass
 
+def _open_with_dirs(file, mode='r', *args, **kwargs):
+    try:
+        if isinstance(file, (str, bytes, os.PathLike)) and any(m in mode for m in ('w','a','x','+')):
+            _ensure_parent_dir(file)
+    finally:
+        return _orig_open(file, mode, *args, **kwargs)
+
+def _codecs_open_with_dirs(filename, mode='r', encoding=None, errors='strict', buffering=1):
+    try:
+        if isinstance(filename, (str, bytes, os.PathLike)) and any(m in mode for m in ('w','a','x','+')):
+            _ensure_parent_dir(filename)
+    finally:
+        return _orig_codecs_open(filename, mode, encoding, errors, buffering)
+
+# Apply patches once
+_builtins.open = _open_with_dirs
+_codecs.open = _codecs_open_with_dirs
+
+# Patch pandas to_csv to auto-create parent folder if available
+try:
+    import pandas as _pd  # noqa: F401
+    _orig_to_csv = _pd.DataFrame.to_csv
+    def _to_csv_with_dirs(self, path_or_buf=None, *args, **kwargs):
+        if isinstance(path_or_buf, (str, bytes, os.PathLike)):
+            _ensure_parent_dir(path_or_buf)
+        return _orig_to_csv(self, path_or_buf, *args, **kwargs)
+    _pd.DataFrame.to_csv = _to_csv_with_dirs
+except Exception:
+    pass
 
 import re, csv
 from collections import defaultdict
-from bs4 import BeautifulSoup as bs
 import pytz, datetime, dateutil, json
 from datetime import timedelta
 from dateutil import tz
 
-import functools
-
 from functools import reduce
 
 # Teacher name format changed. Remove commas and switch first to last   
@@ -84,6 +126,8 @@ def extract_key_values(lst, x):
     return reduce(lambda acc, item: acc + [item[x]] if isinstance(item, dict) and x in item else acc, lst, [])
 
 def stripper(s):
+    from bs4 import BeautifulSoup as bs
+
     REMOVE_ATTRIBUTES = [
     'lang','language','onmouseover','onmouseout','script','style','font',
     'dir','face','size','color','style','class','width','height','hspace',
@@ -281,6 +325,8 @@ def clean_fn(s):
     return s
     
 def format_html(html):
+    from bs4 import BeautifulSoup as bs
+
     soup = bs(html, 'html.parser')
     return soup.prettify()