Commit c4ae59ca authored by codey's avatar codey
Browse files

allow for excluding cache on lyrics search (temp?) + addl genius fixes

parent fed53073

lyric_search/sources/aggregate.py

+1 −1
Original line number Diff line number Diff line
@@ -47,7 +47,7 @@ class Aggregate: 
        search_result: Optional[LyricsResult] = None

        for source in sources:

            if source.label.lower() in self.exclude_methods:

                if source.label.lower() == "cache" or not plain:

                if not plain:

                    logging.info("Exclude conditions rejected - source requested to exclude: %s, plain: %s",

                                 source.label, plain)                    

                else:

lyric_search/sources/genius.py

+12 −4
Original line number Diff line number Diff line
@@ -3,6 +3,7 @@ sys.path.insert(1,'..') 
import traceback

import logging

import time

import re

from typing import Optional

from aiohttp import ClientTimeout, ClientSession

from bs4 import BeautifulSoup, ResultSet # type: ignore
@@ -101,16 +102,23 @@ class Genius: 
                        

                        

                        html = BeautifulSoup(htm.unescape(scrape_text).replace('<br/>', '\n'), "html.parser")

                        header_tags: Optional[ResultSet] =  html.find_all(['h1', 'h2', 'h3', 'h4', 'h5', 'span'])

                        if header_tags:

                            for tag in header_tags:

                        

                        header_tags_genius: Optional[ResultSet] = html.find_all(class_=re.compile('.*Header.*'))

                        if header_tags_genius:

                            for tag in header_tags_genius:

                                tag.extract()

                        

                        divs: Optional[ResultSet] = html.find_all("div", {"data-lyrics-container": "true"})

                                                

                        if not divs:

                            return

                        

                        for div in divs:

                            header_tags: Optional[ResultSet] =  div.find_all(['h1', 'h2', 'h3', 'h4', 'h5'])

                            if header_tags:

                                for tag in header_tags:

                                    tag.extract()                              



                            returned_lyrics += div.get_text()

                        

                        returned_lyrics: str = self.datautils.scrub_lyrics(returned_lyrics)

utils/radio_util.py

+3 −3
Original line number Diff line number Diff line
@@ -186,15 +186,15 @@ class RadioUtil: 
            LIMITED TO ONE/SMALL SUBSET OF GENRES

            """

            

            # db_query = 'SELECT distinct(artist || " - " || song) AS artistdashsong, id, artist, song, album, genre, file_path, duration FROM tracks\

            #     WHERE artist LIKE "%tethered%" OR artist LIKE "%more of mysel%" ORDER BY artist DESC, album ASC, song ASC'

            db_query = 'SELECT distinct(artist || " - " || song) AS artistdashsong, id, artist, song, album, genre, file_path, duration FROM tracks\

                WHERE genre LIKE "%edm%" OR artist LIKE "%sullivan king%" OR artist LIKE "%kai wachi%" OR artist LIKE "%kayzo%" ORDER BY RANDOM()' #ORDER BY artist DESC, album ASC, song ASC'

                

            """

            LIMITED TO ONE/SOME ARTISTS...

            """

            

            # db_query = 'SELECT distinct(artist || " - " || song) AS artistdashsong, id, artist, song, album, genre, file_path, duration FROM tracks\

            #     WHERE (artist LIKE "%sullivan king%" OR artist LIKE "%kayzo%" OR artist like "%kai wachi%" OR genre LIKE "%edm%" OR genre LIKE "%electronicore%") AND (NOT song LIKE "%%stripped%%" AND NOT song LIKE "%(2022)%" AND NOT song LIKE "%(live%%" AND NOT song LIKE "%%acoustic%%" AND NOT song LIKE "%%instrumental%%" AND NOT song LIKE "%%remix%%" AND NOT song LIKE "%%reimagined%%" AND NOT song LIKE "%%alternative%%" AND NOT song LIKE "%%unzipped%%") GROUP BY artistdashsong ORDER BY RANDOM()'# ORDER BY album ASC, id ASC'

            #     WHERE (artist LIKE "%a scent like wolves%" OR artist LIKE "%bad wolves%" or artist LIKE "%oceans%" OR artist LIKE "%oh,%" OR artist LIKE "%august%" OR artist LIKE "%periphery%") AND (NOT song LIKE "%%stripped%%" AND NOT song LIKE "%(2022)%" AND NOT song LIKE "%(live%%" AND NOT song LIKE "%%acoustic%%" AND NOT song LIKE "%%instrumental%%" AND NOT song LIKE "%%remix%%" AND NOT song LIKE "%%reimagined%%" AND NOT song LIKE "%%alternative%%" AND NOT song LIKE "%%unzipped%%") GROUP BY artistdashsong ORDER BY RANDOM()'# ORDER BY album ASC, id ASC'

            

            async with sqlite3.connect(self.active_playlist_path,

                                       timeout=2) as db_conn: