Benjamin_Loison/waybackpy
1
0
Fork 0
Code Issues 1 Pull Requests Packages Projects Releases Wiki Activity

Update wrapper.py

Browse Source
This commit is contained in:
Akash Mahanty
2021-01-02 01:28:53 +05:30
committed by GitHub
parent 1623e2d940
commit 1f03b9146a
1 changed files with 13 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
waybackpy/wrapper.py
View File

@@ -65,7 +65,7 @@ class Url:
self.user_agent = user_agent self.user_agent = user_agent
self._url_check() # checks url validity on init. self._url_check() # checks url validity on init.
self._archive_url = None # URL of archive self._archive_url = None # URL of archive
self._timestamp = None # timestamp for last archive self.timestamp = None # timestamp for last archive
self._alive_url_list = [] self._alive_url_list = []
def __repr__(self): def __repr__(self):
@@ -81,13 +81,13 @@ class Url:
days=999999999, hours=23, minutes=59, seconds=59, microseconds=999999 days=999999999, hours=23, minutes=59, seconds=59, microseconds=999999
) )
if not self._timestamp: if not self.timestamp:
self._timestamp = self.timestamp self.timestamp = self._timestamp
if self._timestamp == datetime.max: if self.timestamp == datetime.max:
return td_max.days return td_max.days
diff = datetime.utcnow() - self._timestamp diff = datetime.utcnow() - self.timestamp
return diff.days return diff.days
def _url_check(self): def _url_check(self):
@@ -106,6 +106,10 @@ class Url:
@property @property
def archive_url(self): def archive_url(self):
"""Get URL of archive.""" """Get URL of archive."""
if self._archive_url:
return self._archive_url
data = self.JSON data = self.JSON
if not data["archived_snapshots"]: if not data["archived_snapshots"]:
@@ -119,7 +123,7 @@ class Url:
return archive_url return archive_url
@property @property
def timestamp(self): def _timestamp(self):
"""Get timestamp of last archive.""" """Get timestamp of last archive."""
data = self.JSON data = self.JSON
@@ -130,7 +134,7 @@ class Url:
ts = datetime.strptime( ts = datetime.strptime(
data["archived_snapshots"]["closest"]["timestamp"], "%Y%m%d%H%M%S" data["archived_snapshots"]["closest"]["timestamp"], "%Y%m%d%H%M%S"
) )
self._timestamp = ts self.timestamp = ts
return ts return ts
def _clean_url(self): def _clean_url(self):
@@ -143,7 +147,7 @@ class Url:
headers = {"User-Agent": "%s" % self.user_agent} headers = {"User-Agent": "%s" % self.user_agent}
response = _get_response(request_url, params=None, headers=headers) response = _get_response(request_url, params=None, headers=headers)
self._archive_url = "https://" + _archive_url_parser(response.headers) self._archive_url = "https://" + _archive_url_parser(response.headers)
self._timestamp = datetime.utcnow() self.timestamp = datetime.utcnow()
return self return self
def get(self, url="", user_agent="", encoding=""): def get(self, url="", user_agent="", encoding=""):
@@ -199,7 +203,7 @@ class Url:
) )
self._archive_url = archive_url self._archive_url = archive_url
self._timestamp = datetime.strptime( self.timestamp = datetime.strptime(
data["archived_snapshots"]["closest"]["timestamp"], "%Y%m%d%H%M%S" data["archived_snapshots"]["closest"]["timestamp"], "%Y%m%d%H%M%S"
) )