mirror/maloja
1
0
Fork 0
mirror of https://github.com/krateng/maloja.git synced 2023-08-10 21:12:55 +03:00
Code Issues Projects Releases Wiki Activity

Refactored some scrobble parsing

Browse Source
This commit is contained in:
krateng 2022-05-07 21:54:51 +02:00
parent f3a04c79b1
commit ac498bde73
1 changed files with 25 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
maloja/cleanup.py
View File

@ -82,7 +82,7 @@ class CleanerAgent:
def parseArtists(self,a): def parseArtists(self,a):
if isinstance(a,list): if isinstance(a,list) or isinstance(a,tuple):
res = [self.parseArtists(art) for art in a] res = [self.parseArtists(art) for art in a]
return [a for group in res for a in group] return [a for group in res for a in group]
@ -156,38 +156,37 @@ class CleanerAgent:
# t = p(t).strip() # t = p(t).strip()
return t return t
def parseTitleForArtists(self,t): def parseTitleForArtists(self,title):
for d in self.delimiters_feat:
if re.match(r"(.*) [\(\[]" + d + " (.*?)[\)\]]",t,flags=re.IGNORECASE) is not None:
(title,artists) = self.parseTitleForArtists(re.sub(r"(.*) [\(\[]" + d + " (.*?)[\)\]]",r"\1",t,flags=re.IGNORECASE))
artists += self.parseArtists(re.sub(r"(.*) [\(\[]" + d + " (.*?)[\)\]].*",r"\2",t,flags=re.IGNORECASE))
return (title,artists)
if re.match(r"(.*) - " + d + " (.*)",t,flags=re.IGNORECASE) is not None:
(title,artists) = self.parseTitleForArtists(re.sub(r"(.*) - " + d + " (.*)",r"\1",t,flags=re.IGNORECASE))
artists += self.parseArtists(re.sub(r"(.*) - " + d + " (.*).*",r"\2",t,flags=re.IGNORECASE))
return (title,artists)
if re.match(r"(.*) " + d + " (.*)",t,flags=re.IGNORECASE) is not None:
(title,artists) = self.parseTitleForArtists(re.sub(r"(.*) " + d + " (.*)",r"\1",t,flags=re.IGNORECASE))
artists += self.parseArtists(re.sub(r"(.*) " + d + " (.*).*",r"\2",t,flags=re.IGNORECASE))
return (title,artists)
artists = [] artists = []
for delimiter in malojaconfig["DELIMITERS_FEAT"]:
for pattern in [
r" [\(\[]" + re.escape(delimiter) + " (.*?)[\)\]]",
r" - " + re.escape(delimiter) + " (.*)",
r" " + re.escape(delimiter) + " (.*)"
]:
matches = re.finditer(pattern,title,flags=re.IGNORECASE)
for match in matches:
title = match.re.sub('',match.string) # Remove matched part
artists += self.parseArtists(match.group(1)) # Parse matched artist string
if malojaconfig["PARSE_REMIX_ARTISTS"]: if malojaconfig["PARSE_REMIX_ARTISTS"]:
for filter in malojaconfig["FILTERS_REMIX"]: for filter in malojaconfig["FILTERS_REMIX"]:
# match remix in brackets for pattern in [
m = re.match(r".*[\(\[](.*)" + filter + "[\)\]]", t, flags=re.IGNORECASE) r" [\(\[](.*)" + re.escape(filter) + "[\)\]]", # match remix in brackets
if m: r" - (.*)" + re.escape(filter) # match remix split with "-"
artists += self.parseArtists(m.groups()[0]) ]:
match = re.search(pattern,title,flags=re.IGNORECASE)
if match:
# title stays the same
artists += self.parseArtists(match.group(1))
# match remix split with "-"
m = re.match(r".*-(.*)" + filter, t, flags=re.IGNORECASE)
if m:
artists += self.parseArtists(m.groups()[0])
for st in self.rules_artistintitle: for st in self.rules_artistintitle:
if st in t.lower(): artists += self.rules_artistintitle[st].split("␟") if st in title.lower(): artists += self.rules_artistintitle[st].split("␟")
return (t,artists) return (title,artists)