Clean up utils.py to use database

Uses of the original library_stats dictionary in utils.py were either removed or commented out, replaced by the saving of models into the database.

spotifyvis/models.py

@@ -1,8 +1,7 @@
 from django.db import models
 
 # id's are 22 in length in examples but set to 30 for buffer
-id_length=30
+MAX_ID = 30
-
 #  Artist {{{ # 
 
 class Artist(models.Model):
@@ -10,7 +9,7 @@ class Artist(models.Model):
         verbose_name = "Artist"
         verbose_name_plural = "Artists"
 
-    artist_id = models.CharField(primary_key=True, max_length=id_length)
+    artist_id = models.CharField(primary_key=True, max_length=MAX_ID)
     # unique since only storing one genre per artist right now
     name = models.CharField(unique=True, max_length=50)
     genre = models.CharField(max_length=20)
@@ -27,7 +26,7 @@ class User(models.Model):
         verbose_name = "User"
         verbose_name_plural = "Users"
 
-    user_id = models.CharField(primary_key=True, max_length=id_length) # the user's Spotify ID
+    user_id = models.CharField(primary_key=True, max_length=MAX_ID) # the user's Spotify ID
     #  username = models.CharField(max_length=30) # User's Spotify user name, if set
 
     def __str__(self):
@@ -43,7 +42,7 @@ class Track(models.Model):
         verbose_name = "Track"
         verbose_name_plural = "Tracks"
 
-    track_id = models.CharField(primary_key=True, max_length=id_length)
+    track_id = models.CharField(primary_key=True, max_length=MAX_ID)
     #  artist = models.ForeignKey(Artist, on_delete=models.CASCADE)
     artists = models.ManyToManyField(Artist, blank=True)
     year = models.PositiveSmallIntegerField()
@@ -79,3 +78,39 @@ class AudioFeatures(models.Model):
         return super(AudioFeatures, self).__str__()
 
 #  }}} AudioFeatures # 
+
+
+'''
+class UserAudioSummary(models.Model):
+    """
+    Stores the summary of a user's audio preferences
+    """ 
+    class Meta:
+        verbose_name = "AudioFeatures"
+        verbose_name_plural = "AudioFeatures"
+
+    user = models.OneToOneField(User, on_delete=models.CASCADE, primary_key=True,)
+    avg_acousticness = models.DecimalField(decimal_places=3, max_digits=3)
+    stdev_acousticness = models.DecimalField(decimal_places=3, max_digits=3)
+
+    avg_danceability = models.DecimalField(decimal_places=3, max_digits=3)
+    stdev_danceability = models.DecimalField(decimal_places=3, max_digits=3)
+
+    avg_energy = models.DecimalField(decimal_places=3, max_digits=3)
+    stdev_energy = models.DecimalField(decimal_places=3, max_digits=3)
+
+    avg_instrumentalness = models.DecimalField(decimal_places=3, max_digits=3)
+    stdev_instrumentalness = models.DecimalField(decimal_places=3, max_digits=3)
+
+    avg_loudness = models.DecimalField(decimal_places=3, max_digits=6)
+    stdev_loudness = models.DecimalField(decimal_places=3, max_digits=6)
+
+    avg_speechiness = models.DecimalField(decimal_places=3, max_digits=3)
+    stdev_speechiness = models.DecimalField(decimal_places=3, max_digits=3)
+
+    avg_tempo = models.DecimalField(decimal_places=3, max_digits=6)
+    stdev_tempo = models.DecimalField(decimal_places=3, max_digits=6)
+
+    avg_valence = models.DecimalField(decimal_places=3, max_digits=3)
+    stdev_valence = models.DecimalField(decimal_places=3, max_digits=3)
+'''

spotifyvis/static/spotifyvis/scripts/test_db.js

@@ -1,10 +1,4 @@
-console.log("{{ user_id }}");
+d3.json("{% url "get_artist_data", user_id %}").then(function(error, data) {
-artist_data = JSON.parse('{{ artist_data }}');
-artist_data.forEach(function(d) {
-	console.log(d.name, d.num_songs);
-});
-
-d3.json("{% url "get_artist_data" user_id %}", function(error, data) {
 	data.forEach(function(d) {
 		console.log(d.name, d.num_songs);
 	});

spotifyvis/templates/spotifyvis/test_db.html

@@ -20,8 +20,13 @@
           {% endfor %}
         </ul>
         <pre> {% filter force_escape %} {% debug %} {% endfilter %} </pre>
-      <script src="http://d3js.org/d3.v3.js"></script>
+      <script src="https://d3js.org/d3.v5.min.js"></script>
-      {% load static %}
+      <script>
-      <script src="{% static "spotifyvis/scripts/test_db.js" %}"></script>
+        d3.json("{% url "get_artist_data" user_id %}").then(function(data) {
+            data.forEach(function(d) {
+                console.log(d.name, d.num_songs);
+            });
+        });
+      </script>
     </body>
 </html>

spotifyvis/utils.py

@@ -1,5 +1,4 @@
 #  imports {{{ # 
-
 import requests
 import math
 import pprint
@@ -8,6 +7,7 @@ from .models import Artist, User, Track, AudioFeatures
 from django.db.models import Count
 from django.http import JsonResponse
 from django.core import serializers
+import json
 
 #  }}} imports # 
 
@@ -32,7 +32,6 @@ def parse_library(headers, tracks, library_stats, user):
     payload = {'limit': str(limit)}
     # use two separate variables to track, because the average popularity also requires num_samples 
     num_samples = 0  # number of actual track samples
-    feature_data_points = 0  # number of feature data analyses (some tracks do not have analyses available)
 
     # iterate until hit requested num of tracks
     for _ in range(0, tracks, limit):
@@ -43,22 +42,27 @@ def parse_library(headers, tracks, library_stats, user):
         # TODO: refactor the for loop body into helper function
         # iterate through each track
         for track_dict in saved_tracks_response['items']:
-            num_samples += 1 
             # update artist info before track so that Track object can reference
             # Artist object
             track_artists = []
             for artist_dict in track_dict['track']['artists']:
-                increase_artist_count(headers, artist_dict['name'], 
+                artist_obj, artist_created = Artist.objects.get_or_create(
-                        artist_dict['id'], library_stats)
-                track_artists.append(Artist.objects.get_or_create(
                     artist_id=artist_dict['id'],
                     name=artist_dict['name'],
-                    )[0])
+                    )
+
+                update_artist_genre(headers, artist_obj) 
+                # get_or_create() returns a tuple (obj, created)
+                track_artists.append(artist_obj)
             
-            track_obj = save_track_obj(track_dict['track'], track_artists, user)
+            track_obj, track_created = save_track_obj(track_dict['track'], track_artists, user)
-            get_track_info(track_dict['track'], library_stats, num_samples)
+
-            audio_features_dict = get_audio_features(headers,
+            # if a new track is not created, the associated audio feature does not need to be created again
-                    track_dict['track']['id'], track_obj)
+            if track_created:
+                save_audio_features(headers, track_dict['track']['id'], track_obj)
+            """
+            TODO: Put this login in another function
+            # Audio analysis could be empty if not present in Spotify database
             if len(audio_features_dict) != 0:
                 # Track the number of audio analyses for calculating
                 # audio feature averages and standard deviations on the fly
@@ -66,62 +70,55 @@ def parse_library(headers, tracks, library_stats, user):
                 for feature, feature_data in audio_features_dict.items():
                     update_audio_feature_stats(feature, feature_data, 
                             feature_data_points, library_stats)
-
+            """
         # calculates num_songs with offset + songs retrieved
-        library_stats['num_songs'] = offset + len(saved_tracks_response['items'])
         offset += limit
-    calculate_genres_from_artists(headers, library_stats)
+    # calculate_genres_from_artists(headers, library_stats)
     #  pprint.pprint(library_stats)
 
 #  }}} parse_library # 
 
 #  save_track_obj {{{ # 
 
+
 def save_track_obj(track_dict, artists, user):
     """Make an entry in the database for this track if it doesn't exist already.
 
     :track_dict: dictionary from the API call containing track information.
     :artists: artists of the song, passed in as a list of Artist objects.
     :user: User object for which this Track is to be associated with.
-    :returns: The created/retrieved Track object.
+    :returns: (The created/retrieved Track object, created) 
 
     """
-    track_obj_query = Track.objects.filter(track_id__exact=track_dict['id'])
+    new_track, created = Track.objects.get_or_create(
-    if len(track_obj_query) == 0:
+        track_id=track_dict['id'],
-        new_track = Track.objects.create(
+        year=track_dict['album']['release_date'].split('-')[0],
-            track_id=track_dict['id'],
+        popularity=int(track_dict['popularity']),
-            year=track_dict['album']['release_date'].split('-')[0],
+        runtime=int(float(track_dict['duration_ms']) / 1000),
-            popularity=int(track_dict['popularity']),
+        name=track_dict['name'],
-            runtime=int(float(track_dict['duration_ms']) / 1000),
+    )
-            name=track_dict['name'],
-        )
-        #  print("pop/run: ", new_track.popularity, new_track.runtime)
 
-        # have to add artists and user after saving object since track needs to
+    # have to add artists and user after saving object since track needs to
-        # have ID before filling in m2m field
+    # have ID before filling in m2m field
+    if created:
         for artist in artists:
             new_track.artists.add(artist)
         new_track.users.add(user)
         new_track.save()
-        return new_track
+    return new_track, created
-    elif len(track_obj_query) == 1:
-        return track_obj_query[0]
 
 #  }}} save_track_obj # 
 
 #  get_audio_features {{{ # 
 
-def get_audio_features(headers, track_id, track):
+def save_audio_features(headers, track_id, track):
-    """Returns the audio features of a soundtrack
+    """Creates and saves a new AudioFeatures object
 
     Args:
         headers: headers containing the API token
         track_id: the id of the soundtrack, needed to query the Spotify API
-        track: Track object to associate with the AudioFeatures object
+        track: Track object to associate with the new AudioFeatures object
         
-    Returns:
-        A dictionary with the features as its keys, if audio feature data is missing for the track, 
-        an empty dictionary is returned.
     """
     
     response = requests.get("https://api.spotify.com/v1/audio-features/{}".format(track_id), headers = headers).json()
@@ -141,7 +138,6 @@ def get_audio_features(headers, track_id, track):
             setattr(audio_features_entry, key, val)
     audio_features_entry.save()
 
-    return features_dict
 
 #  }}} get_audio_features # 
 
@@ -300,30 +296,27 @@ def get_track_info(track_dict, library_stats, sample_size):
 
 #  }}} get_track_info # 
 
-#  calculate_genres_from_artists {{{ # 
+#  update_genres_from_artists {{{ # 
 
-def calculate_genres_from_artists(headers, library_stats):
+
-    """Tallies up genre counts based on artists in library_stats.
+def update_artist_genre(headers, artist_obj):
+    """Updates the top genre for an artist by querying the Spotify API
 
     :headers: For making the API call.
-    :library_stats: Dictionary containing the data mined from user's Spotify library
+    :artist_obj: the Artist object whose genre field will be updated 
 
     :returns: None
 
     """
-    for artist_entry in library_stats['artists'].values():
+    artist_response = requests.get('https://api.spotify.com/v1/artists/' + artist_obj.id, headers=headers).json()
-        artist_response = requests.get('https://api.spotify.com/v1/artists/' + artist_entry['id'], headers=headers).json()
+    # update genre for artist in database with top genre
-        # increase each genre count by artist count
+    artist_obj.update(genre=artist_response['genres'][0])
-        for genre in artist_response['genres']:
-            increase_nested_key('genres', genre, library_stats, artist_entry['count'])
-
-        # update genre for artist in database with top genre
-        Artist.objects.filter(artist_id=artist_entry['id']).update(genre=artist_response['genres'][0])
 
 #  }}} calculate_genres_from_artists # 
 
 #  process_library_stats {{{ # 
 
+
 def process_library_stats(library_stats):
     """Processes library_stats into format more suitable for D3 consumption
 
@@ -388,23 +381,3 @@ def get_genre_data(user):
     #  user_tracks = Track.objects.filter(users__exact=user)
     #  for track in user_tracks:
         #  print(track.name)
-
-
-def get_artist_data(user_id):
-    """Return artist data needed to create the graph for user.
-
-    :user_id: user ID for which to return the data for.
-    :returns: List of dicts containing counts for each artist.
-
-    """
-    # TODO: not actual artists for user
-    # PICK UP: figure out how to pass data to D3/frontend
-    print(user_id)
-    #  user = User.objects.get(user_id=user_id)
-    artist_counts = Artist.objects.annotate(num_songs=Count('track'))
-    processed_artist_data = [{'name': artist.name, 'num_songs': artist.num_songs} for artist in artist_counts]
-    #  for artist in artist_counts:
-        #  print(artist.name, artist.num_songs)
-    return JsonResponse(processed_artist_data, safe=False)
-    #  return serializers.serialize('json', processed_artist_data)
-    #  return processed_artist_data

spotifyvis/views.py

@@ -10,9 +10,9 @@ import pprint
 from datetime import datetime
 
 from django.shortcuts import render, redirect
-from django.http import HttpResponse, HttpResponseBadRequest
+from django.http import HttpResponse, HttpResponseBadRequest, JsonResponse
 from django.db.models import Count
-from .utils import parse_library, process_library_stats, get_artist_data
+from .utils import parse_library, process_library_stats 
 from .models import User, Track, AudioFeatures, Artist 
 
 #  }}} imports # 
@@ -44,6 +44,7 @@ def generate_random_string(length):
 
 #  token_expired {{{ # 
 
+
 def token_expired(token_obtained_at, valid_for):
     """Returns True if token expired, False if otherwise
 
@@ -143,6 +144,8 @@ def user_data(request):
     user_data_response = requests.get('https://api.spotify.com/v1/me', headers = headers).json()
     request.session['user_id'] = user_data_response['id'] # store the user_id so it may be used to create model
     #  request.session['user_name'] = user_data_response['display_name']
+
+    # get_or_create() returns a tuple (obj, created)
     user = User.objects.get_or_create(user_id=user_data_response['id'])[0]
 
     context = {
@@ -174,8 +177,20 @@ def user_data(request):
 def test_db(request):
     user_id = "polarbier"
     context = {
-        'artist_data': get_artist_data(user_id),
         'user_id': user_id,
     }
     #  get_artist_data(user)
     return render(request, 'spotifyvis/test_db.html', context)
+
+
+def get_artist_data(request, user_id):
+
+    # TODO: not actual artists for user
+    # PICK UP: figure out how to pass data to D3/frontend
+    print(user_id)
+    #  user = User.objects.get(user_id=user_id)
+    artist_counts = Artist.objects.annotate(num_songs=Count('track'))
+    processed_artist_data = [{'name': artist.name, 'num_songs': artist.num_songs} for artist in artist_counts]
+    #  for artist in artist_counts:
+        #  print(artist.name, artist.num_songs)
+    return JsonResponse(data=processed_artist_data, safe=False)