diff options
Diffstat (limited to 'jfr_playoff')
| -rw-r--r-- | jfr_playoff/db.py | 4 | ||||
| -rw-r--r-- | jfr_playoff/remote.py | 10 |
2 files changed, 10 insertions, 4 deletions
diff --git a/jfr_playoff/db.py b/jfr_playoff/db.py index 9d26a67..bca05d2 100644 --- a/jfr_playoff/db.py +++ b/jfr_playoff/db.py @@ -35,9 +35,7 @@ class PlayoffDB(object): return row except mysql.connector.Error as e: PlayoffLogger.get('db').error(str(e)) - raise IOError( - message=str(e), filename=db_name, - errno=e.errno, strerror=str(e)) + raise IOError(e.errno, str(e), db_name) def fetch_all(self, db_name, sql, params): import mysql.connector diff --git a/jfr_playoff/remote.py b/jfr_playoff/remote.py index a113c96..339390c 100644 --- a/jfr_playoff/remote.py +++ b/jfr_playoff/remote.py @@ -1,3 +1,5 @@ +import re + import requests from bs4 import BeautifulSoup as bs @@ -12,7 +14,13 @@ class RemoteUrl: PlayoffLogger.get('remote').info( 'fetching content for: %s', url) if url not in cls.url_cache: - cls.url_cache[url] = requests.get(url).text + request = requests.get(url) + encoding_match = re.search( + 'content=".*;( )?charset=(.*)"', + request.content, re.IGNORECASE) + if encoding_match: + request.encoding = encoding_match.group(2) + cls.url_cache[url] = request.text PlayoffLogger.get('remote').info( 'content for %s not in cache: retrieved %d bytes', url, len(cls.url_cache[url])) |