From c2ed4a39ac8573148dc62045c4562c52c04aa5c0 Mon Sep 17 00:00:00 2001 From: wengtad Date: Fri, 11 Jun 2021 00:08:10 +0800 Subject: [PATCH] fix clean_string error (#481) --- mealie/services/scraper/cleaner.py | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-) diff --git a/mealie/services/scraper/cleaner.py b/mealie/services/scraper/cleaner.py index 842113a18adc..401d12ca298b 100644 --- a/mealie/services/scraper/cleaner.py +++ b/mealie/services/scraper/cleaner.py @@ -36,13 +36,16 @@ def clean(recipe_data: dict, url=None) -> dict: def clean_string(text: str) -> str: - cleaned_text = html.unescape(text) - cleaned_text = re.sub("<[^<]+?>", "", cleaned_text) - cleaned_text = re.sub(" +", " ", cleaned_text) - cleaned_text = re.sub("

", "\n", cleaned_text) - cleaned_text = re.sub(r"\n\s*\n", "\n\n", cleaned_text) - cleaned_text = cleaned_text.replace("\xa0", " ").replace("\t", " ").strip() - return cleaned_text + if text == "" or text is None: + return "" + else: + cleaned_text = html.unescape(text) + cleaned_text = re.sub("<[^<]+?>", "", cleaned_text) + cleaned_text = re.sub(" +", " ", cleaned_text) + cleaned_text = re.sub("

", "\n", cleaned_text) + cleaned_text = re.sub(r"\n\s*\n", "\n\n", cleaned_text) + cleaned_text = cleaned_text.replace("\xa0", " ").replace("\t", " ").strip() + return cleaned_text def category(category: str):