diff options
| author | Bryan Newbold <bnewbold@archive.org> | 2021-10-26 12:22:38 -0700 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@archive.org> | 2021-10-26 12:22:38 -0700 |
| commit | 3cdf4af9be4c762ff2ed79a57b5ad30637909f1e (patch) | |
| tree | b7e7e27ff2032c99fd782b3ea40daf1d12f9164e /python/sandcrawler/html_metadata.py | |
| parent | f67d870ba4ca9cecd0b75f106335997c813e9df4 (diff) | |
| download | sandcrawler-3cdf4af9be4c762ff2ed79a57b5ad30637909f1e.tar.gz sandcrawler-3cdf4af9be4c762ff2ed79a57b5ad30637909f1e.zip | |
python: isort all imports
Diffstat (limited to 'python/sandcrawler/html_metadata.py')
| -rw-r--r-- | python/sandcrawler/html_metadata.py | 9 |
1 files changed, 4 insertions, 5 deletions
diff --git a/python/sandcrawler/html_metadata.py b/python/sandcrawler/html_metadata.py index 93c7269..c6725dc 100644 --- a/python/sandcrawler/html_metadata.py +++ b/python/sandcrawler/html_metadata.py @@ -1,17 +1,16 @@ -import sys import datetime -from typing import List, Optional, Any, Tuple, Dict +import sys import urllib.parse +from typing import Any, Dict, List, Optional, Tuple +import braveblock import dateparser -from selectolax.parser import HTMLParser import pydantic -import braveblock +from selectolax.parser import HTMLParser from sandcrawler.misc import url_fuzzy_equal - # this is a map of metadata keys to CSS selectors # sources for this list include: # - google scholar crawling notes (https://scholar.google.com/intl/ja/scholar/inclusion.html#indexing) |