import spotipy
from spotipy.oauth2 import SpotifyClientCredentials
import requests
import time
import lyricsgenius
# --------- Toggles ----------
USE_LOCAL_DEEPSEEK = False # True = use local DeepSeek via Ollama, False = use OpenAI online
USE_GENIUS = True # True = use Genius for lyrics, False = skip lyrics
USE_LASTFM = True # True = use Last.fm for additional info, False = skip
SHOW_LLM_PROMPT_DEBUG = False # True = show the full prompt with all fields, False = skip printing it
# ---------- Spotify API credentials ----------
CLIENT_ID = '6a....9f'
CLIENT_SECRET = '56....7'
PLAYLIST_ID = '1mZELHyKrXNyLOumaDiQwh'
# ---------- Genius API credentials ----------
GENIUS_TOKEN = 'cu....25'
# ---------- Last.fm API credentials ----------
LASTFM_API_KEY = 'd1....8g'
# ---------- Local DeepSeek/Ollama ----------
OLLAMA_URL = 'http://localhost:11434/api/generate'
MODEL_LOCAL = 'deepseek-coder-v2'
# ---------- OpenAI Online (GPT-3.5/4/4o) ----------
OPENAI_API_URL = 'https://api.openai.com/v1/chat/completions'
OPENAI_API_KEY = 'sk....1B'
MODEL_ONLINE = 'gpt-4o'
# ---------- Prompt templates ----------
with open('prompt.txt', encoding='utf-8') as f:
PROMPT_TEMPLATE = f.read()
with open('system_prompt.txt', encoding='utf-8') as f:
SYSTEM_PROMPT = f.read()
sp = spotipy.Spotify(
auth_manager=SpotifyClientCredentials(
client_id=CLIENT_ID,
client_secret=CLIENT_SECRET
)
)
genius = lyricsgenius.Genius(GENIUS_TOKEN, skip_non_songs=True, verbose=False, remove_section_headers=True)
def clean_track_name(name):
import re
name = re.sub(r"\(.*?\)", "", name)
name = name.replace("- En Vivo", "")
name = name.strip()
return name
def get_lyrics(track_name, artist_name):
if not USE_GENIUS:
return ""
cleaned_name = clean_track_name(track_name)
song = genius.search_song(cleaned_name, artist_name)
if song and getattr(song, "lyrics", None):
return song.lyrics[:1000]
song = genius.search_song(cleaned_name)
if song and getattr(song, "lyrics", None):
return song.lyrics[:1000]
return ""
def get_lastfm_info(track_name, artist_name):
if not USE_LASTFM:
return ""
try:
url_track = (
f"https://ws.audioscrobbler.com/2.0/?method=track.getInfo"
f"&api_key={LASTFM_API_KEY}&artist={requests.utils.quote(artist_name)}"
f"&track={requests.utils.quote(track_name)}&format=json"
)
resp_track = requests.get(url_track, timeout=10)
info_track = resp_track.json().get("track", {})
tags = [t["name"] for t in info_track.get("toptags", {}).get("tag", [])]
wiki = info_track.get("wiki", {}).get("summary", "")
url_artist = (
f"https://ws.audioscrobbler.com/2.0/?method=artist.getInfo"
f"&api_key={LASTFM_API_KEY}&artist={requests.utils.quote(artist_name)}&format=json"
)
resp_artist = requests.get(url_artist, timeout=10)
info_artist = resp_artist.json().get("artist", {})
artist_tags = [t["name"] for t in info_artist.get("tags", {}).get("tag", [])]
bio = info_artist.get("bio", {}).get("summary", "")
extra = ""
if tags:
extra = f"Last.fm track tags: {', '.join(tags)}. "
if artist_tags:
extra = f"Last.fm artist tags: {', '.join(artist_tags)}. "
if wiki:
extra = f"Track summary: {wiki[:300]}... "
if bio:
extra = f"Artist bio: {bio[:300]}... "
return extra.strip()
except Exception:
return ""
def classify_track(
track_name, artist_name, lyrics, album_name, release_date, artist_genres, lastfm_info, temperature=0.0
):
prompt = PROMPT_TEMPLATE.format(
track_name=track_name,
artist_name=artist_name,
album_name=album_name,
release_date=release_date,
artist_genres=artist_genres,
lastfm_info=lastfm_info,
lyrics=lyrics
)
# --- Print FULL prompt for debugging and transparency if toggle is active ---
if SHOW_LLM_PROMPT_DEBUG:
print("\n================= LLM PROMPT =================")
print(f"Song: {track_name}")
print(f"Artist: {artist_name}")
print(f"Album: {album_name}")
print(f"Release Date: {release_date}")
print(f"Genres: {artist_genres}")
print(f"Last.fm Info: {lastfm_info}")
print(f"Lyrics: {lyrics[:400]}")
print("==============================================\n")
# ------------------------------------------------------
if USE_LOCAL_DEEPSEEK:
data = {
"model": MODEL_LOCAL,
"prompt": prompt,
"stream": False
}
try:
response = requests.post(OLLAMA_URL, json=data, timeout=120)
if response.ok:
answer = response.json().get('response', '').strip()
if answer in ["Cubana", "Línea"]:
return answer
return None
else:
return None
except Exception:
return None
else:
headers = {
"Authorization": f"Bearer {OPENAI_API_KEY}",
"Content-Type": "application/json"
}
data = {
"model": MODEL_ONLINE,
"messages": [
{"role": "system", "content": SYSTEM_PROMPT},
{"role": "user", "content": prompt}
],
"max_tokens": 5,
"temperature": temperature
}
try:
response = requests.post(OPENAI_API_URL, headers=headers, json=data, timeout=120)
if response.ok:
answer = response.json()['choices'][0]['message']['content'].strip()
if answer in ["Cubana", "Línea"]:
return answer
return None
else:
return None
except Exception:
return None
# -- Fetch tracks --
tracks = []
offset = 0
limit = 100
while True:
results = sp.playlist_items(PLAYLIST_ID, offset=offset, limit=limit)
items = results.get('items', [])
if not items:
break
tracks.extend(items)
offset = len(items)
for idx, item in enumerate(tracks):
track = item.get('track')
if not track:
continue
name = track.get('name', 'Unknown Title')
artist = ', '.join([a.get('name', 'Unknown Artist') for a in track.get('artists', [])])
album = track.get('album', {}).get('name', 'Unknown Album')
release_date = track.get('album', {}).get('release_date', 'Unknown Date')
print(f"\n[{idx 1}] {name} - {artist}")
# Last.fm info
#print(" Fetching Last.fm info ...")
lastfm_info = get_lastfm_info(name, artist)
# Lyrics
#print(" Fetching lyrics ...")
lyrics = get_lyrics(name, artist)
# Genres for first artist
artist_id = track.get('artists', [{}])[0].get('id', None)
artist_genres = []
if artist_id:
try:
artist_info = sp.artist(artist_id)
artist_genres = artist_info.get('genres', [])
except Exception:
artist_genres = []
genres_str = ', '.join(artist_genres)
# Klassifikation (zeigt Prompt im classify_track)
style = classify_track(
track_name=name,
artist_name=artist,
album_name=album,
release_date=release_date,
artist_genres=genres_str,
lastfm_info=lastfm_info,
lyrics=lyrics
)
if style not in ["Cubana", "Línea"]:
style = classify_track(
track_name=name,
artist_name=artist,
album_name=album,
release_date=release_date,
artist_genres=genres_str,
lastfm_info=lastfm_info,
lyrics=lyrics,
temperature=0.3
)
if style not in ["Cubana", "Línea"]:
style = "Cubana"
print(f" => {style}")
time.sleep(1.5)
analyzer.py
