Improved the matching/filtering algorithm
Now, only songs that match at least 70% of the query are added. Also, short and common words (a, on, the, and) are ignored.
This commit is contained in:
Markil3
125
play
125
play
@@ -6,24 +6,42 @@ import subprocess
|
||||
import os
|
||||
import sys
|
||||
import random
|
||||
import re
|
||||
from os.path import exists
|
||||
from bisect import bisect_left, bisect_right
|
||||
import sqlite3
|
||||
|
||||
MATCH_RATE = 0.7
|
||||
songVals = {}
|
||||
|
||||
'''
|
||||
A sorting method for songs. It works by pushing file names with more matches up higher.
|
||||
'''
|
||||
def getValueFunction(terms):
|
||||
def valOfResult(song):
|
||||
val = 0
|
||||
for arg in terms:
|
||||
for part in song[2:]:
|
||||
if arg in part.lower():
|
||||
val -= 1
|
||||
# Double every point received by a playlist
|
||||
if song[0] == True:
|
||||
val -= 1
|
||||
#print("%s got %d hits" % (song, val))
|
||||
return val
|
||||
sortval = 0
|
||||
if song in songVals:
|
||||
val, sortval = songVals[song[1]][0]
|
||||
else:
|
||||
#print(song[1])
|
||||
#print('\t', end='')
|
||||
for arg in terms:
|
||||
i = 1
|
||||
for part in song[2:]:
|
||||
part = re.sub('[^A-Za-z ]', '', part).lower()
|
||||
if arg in part:
|
||||
val += 1
|
||||
# This helps make title matches more valuable
|
||||
sortval += 3 - 1
|
||||
#print("%d %s - %s" % (val, arg, part), end="\t")
|
||||
i += 1
|
||||
songVals[song[1]] = (val, sortval)
|
||||
#print()
|
||||
# Double every point received by a playlist
|
||||
if song[0] == True:
|
||||
val *= 2
|
||||
return (val, sortval)
|
||||
return valOfResult
|
||||
|
||||
'''
|
||||
@@ -70,6 +88,8 @@ def mergeWords(argList):
|
||||
return argList
|
||||
|
||||
def extractOptions(args):
|
||||
# Common words that are in a lot of song titles that we should avoid
|
||||
commonWords = ["the", "off", "for", "from", "with", "and"]
|
||||
options = {}
|
||||
i = 0
|
||||
while i < len(args):
|
||||
@@ -77,13 +97,34 @@ def extractOptions(args):
|
||||
if args[i].startswith("-") and args[i] != "-or" and args[i] != "-and" and args[i] != "--or" and args[i] != "--and":
|
||||
if args[i].startswith("--"):
|
||||
if "=" in args[i]:
|
||||
options[args[i][2:args[i].index("=")]] = args[i][args[i].index("=") + 1:]
|
||||
val = args[i][args[i].index("=") + 1:]
|
||||
if re.sub('\.', '', val).isnumeric():
|
||||
if '.' in val:
|
||||
val = float(val)
|
||||
else:
|
||||
val = int(val)
|
||||
options[args[i][2:args[i].index("=")]] = val
|
||||
else:
|
||||
options[args[i][2:]] = True
|
||||
# These two have special meanings
|
||||
else:
|
||||
lastChar = None
|
||||
numBuilder = ''
|
||||
for char in args[i][1:]:
|
||||
options[char] = True
|
||||
if char.isnumeric() or char == '.':
|
||||
numBuilder += char
|
||||
else:
|
||||
if numBuilder:
|
||||
if '.' in numBuilder:
|
||||
options[lastChar] = float(numBuilder)
|
||||
else:
|
||||
options[lastChar] = int(numBuilder)
|
||||
numBuilder = ''
|
||||
lastChar = char
|
||||
options[char] = True
|
||||
if lastChar and numBuilder:
|
||||
options[lastChar] = float(numBuilder)
|
||||
numBuilder = ''
|
||||
del args[i]
|
||||
elif args[i] == "playlist" or args[i] == "m3u":
|
||||
options["playlist"] = True
|
||||
@@ -92,6 +133,9 @@ def extractOptions(args):
|
||||
options["playlist"] = True
|
||||
args[i] = args[i][0:args[i].rindex(".m3u")]
|
||||
i += 1
|
||||
# Remove problematic words
|
||||
elif len(args[i]) < 3 or args[i].lower() in commonWords:
|
||||
del args[i]
|
||||
else:
|
||||
i += 1
|
||||
return options
|
||||
@@ -173,6 +217,17 @@ def appendDatabaseArguments(argList, args, options, playlist):
|
||||
|
||||
def searchDatabase(options):
|
||||
files = []
|
||||
addedUris = []
|
||||
vals = []
|
||||
minAcceptance = MATCH_RATE
|
||||
if "match" in options:
|
||||
minAcceptance = options["match"]
|
||||
elif "m" in options:
|
||||
minAcceptance = options["m"]
|
||||
if minAcceptance > 10:
|
||||
minAcceptance /= 100
|
||||
elif minAcceptance > 1:
|
||||
minAcceptance /= 10
|
||||
|
||||
if "directory" in options:
|
||||
dbDir = Path(options["directory"])
|
||||
@@ -185,16 +240,45 @@ def searchDatabase(options):
|
||||
terms = appendDatabaseArguments(findArgs, sys.argv[1:], options, True)
|
||||
findArgs = " ".join(findArgs) + ";"
|
||||
results = db.execute(findArgs);
|
||||
|
||||
for row in results:
|
||||
files.append((True,) + row)
|
||||
matches = 0
|
||||
for term in terms:
|
||||
if term in re.sub('[^A-Za-z ]', '', row[1]).lower():
|
||||
matches += 1
|
||||
if matches >= minAcceptance * len(terms):
|
||||
# Make playlist matches more important
|
||||
matches *= 10
|
||||
songVals[row[0]] = (matches, matches)
|
||||
point = random.randint(bisect_left(vals, matches), bisect_right(vals, matches))
|
||||
vals.insert(point, matches)
|
||||
files.insert(point, (True,) + row)
|
||||
addedUris.append(row[0])
|
||||
db.close()
|
||||
db = sqlite3.connect(str(Path(dbDir, "lollypop.db")))
|
||||
findArgs = "SELECT tracks.uri, tracks.name, albums.name, artists.name FROM tracks INNER JOIN albums ON tracks.album_id = albums.id INNER JOIN track_artists ON tracks.id = track_artists.track_id INNER JOIN artists ON track_artists.artist_id = artists.id WHERE".split(" ")
|
||||
terms = appendDatabaseArguments(findArgs, sys.argv[1:], options, False)
|
||||
findArgs.append("GROUP")
|
||||
findArgs.append("BY")
|
||||
findArgs.append("tracks.uri")
|
||||
findArgs = " ".join(findArgs) + ";"
|
||||
#print(findArgs)
|
||||
results = db.execute(findArgs);
|
||||
|
||||
evaluate = getValueFunction(terms)
|
||||
|
||||
for row in results:
|
||||
files.append((False,) + row)
|
||||
# Skip any songs added as part of a playlist
|
||||
if row[0] in addedUris:
|
||||
continue
|
||||
#print(row[1])
|
||||
row = (False,) + row
|
||||
val = evaluate(row)
|
||||
if val[0] >= minAcceptance * len(terms):
|
||||
# Inserts it in order, but random to matches of an equal stature
|
||||
point = random.randint(bisect_left(vals, val[1]), bisect_right(vals, val[1]))
|
||||
vals.insert(point, val[1])
|
||||
files.insert(point, row)
|
||||
db.close()
|
||||
|
||||
return (files, terms)
|
||||
@@ -206,7 +290,6 @@ if __name__ == "__main__":
|
||||
print("VLC Media Player is not installed", file=sys.stderr)
|
||||
exit(1)
|
||||
|
||||
print(sys.argv)
|
||||
mergeWords(sys.argv)
|
||||
options = extractOptions(sys.argv)
|
||||
if "h" in options or "help" in options:
|
||||
@@ -217,9 +300,12 @@ OPTIONS:
|
||||
--directory=[dir] Define the database directory.
|
||||
-a The search must match every single term
|
||||
-v Displays the media player
|
||||
-l Loops through the found files until a new command is entered.
|
||||
-l Loops through the found files until a new command is
|
||||
entered.
|
||||
-s Only the first option will play.
|
||||
-h, --help Shows this message
|
||||
-m[rate], --match=<rate>All matches must have this percentage of matching
|
||||
keywords in order to be included. Defaults to 0.65.
|
||||
|
||||
TERMS:
|
||||
For the most part, you can just enter a single-word term, and it will be added
|
||||
@@ -246,14 +332,17 @@ m3u, playlist Normally, playlists are ignored during the file search.
|
||||
print("No files found", file=sys.stderr)
|
||||
exit(1)
|
||||
|
||||
random.shuffle(files)
|
||||
files.sort(key=getValueFunction(terms))
|
||||
files.reverse()
|
||||
|
||||
for i in range(len(files)):
|
||||
files[i] = files[i][1]
|
||||
if "s" in options:
|
||||
files = [files[0]]
|
||||
for song in files:
|
||||
print(song)
|
||||
if song in songVals:
|
||||
print("%02d - %s" % (songVals[song][0], song))
|
||||
else:
|
||||
print("Pl - %s" % (song))
|
||||
|
||||
if "v" in options:
|
||||
vlcArgs = ["vlc"]
|
||||
|
||||
Reference in New Issue
Block a user