DeCarabas/cry
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Add "feed" as something we look for in an url

Browse source 
Helps us find cat and girl's feed automatically. Also just try "/feed"
maybe it works?
This commit is contained in:
John Doty 2024-09-13 10:30:03 -07:00
parent 3c278a09ba
commit 51049b4d30
1 changed files with 20 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

25
cry/feed.py
View file

@ -664,14 +664,14 @@ def classify_links(links, baseuri) -> typing.Tuple[list[str], list[str]]:
"""
baseuri = baseuri.lower()
local, remote = [], []
local, remote = set(), set()
for link in links:
if link.lower().startswith(baseuri):
local.append(link)
local.add(link)
else:
remote.append(link)
remote.add(link)
return local, remote
return list(local), list(remote)
def is_feed_link(link: str) -> bool:
@ -687,18 +687,27 @@ def is_feed_link(link: str) -> bool:
def is_XML_related_link(link: str) -> bool:
link = link.lower()
return "rss" in link or "rdf" in link or "xml" in link or "atom" in link
return (
"rss" in link
or "rdf" in link
or "xml" in link
or "atom" in link
or "feed" in link
)
async def check_feed(url: str) -> Feed | None:
"""Check to see if the given URL is a feed. If it is, return the feed,
otherwise return None.
"""
LOG.debug(f"Checking {url}: checking...")
meta = FeedMeta.from_url(url)
result, meta = await fetch_feed(meta)
if isinstance(result, Feed):
LOG.debug(f"Checking {url}: is a feed")
return result
LOG.debug(f"Checking {url}: is not a feed")
return None
@ -738,6 +747,11 @@ async def feed_search(uri: str) -> list[Feed]:
LOG.debug("No links, checking A tags...")
local_links, remote_links = classify_links(parser.a_links, meta.url)
for link in local_links:
LOG.debug(f" LOCAL {link}")
for link in remote_links:
LOG.debug(f" REMOTE {link}")
outfeeds = await check_links(filter(is_feed_link, local_links))
if len(outfeeds) > 0:
return outfeeds
@ -759,6 +773,7 @@ async def feed_search(uri: str) -> list[Feed]:
"rss.xml", # Dave Winer/Manila
"index.xml", # MT
"index.rss", # Slash
"feed", # catandgirl.com and sometimes others.
]
outfeeds = await check_links([urllib.parse.urljoin(meta.url, x) for x in suffixes])
return outfeeds