Added README.md

README.md

@@ -0,0 +1,22 @@
+# Music Meteorologist Data Science API
+
+This Flask app takes in the ID of a Spotify song from an HTTP POST request and outputs three recommended songs (specifically their IDs and cosine similarity scores).
+
+Once the Flask app has received a request, it calls the Spotify API to retrieve the characteristics of the Spotify song and converts those characteristics to a NumPy array. Then, the function all_similarities() is called, where a is the numpy array of the song characteristics and dfy is a component of our database of songs (to which this Spotify song will be compared, and from where we will receive our predictions).
+
+```
+def cosine_similarity(a, b):
+  dot_product = np.dot(a,b)
+  norm_a = np.linalg.norm(a)
+  norm_b = np.linalg.norm(b)
+  return dot_product / (norm_a * norm_b)
+  #return np.sqrt(np.sum((a - b) ** 2))
+def all_similarities(a, dfy):
+  similar_songs = []
+  for spotify_song, metadata in zip(array, dfy.values):
+    similarity = cosine_similarity(a, spotify_song)
+    similar_songs.append({'similarity': ''.join(str(similarity.tolist())), 'values': metadata[1]})
+  return similar_songs
+  ```
+
+

app.py

@@ -55,24 +55,21 @@ def all_similarities(a, dfy):
   similar_songs = []
   for spotify_song, metadata in zip(array, dfy.values):
     similarity = cosine_similarity(a, spotify_song)
-    similar_songs.append({'similarity': similarity, 'values': metadata[1]})
+    similar_songs.append({'similarity': ''.join(str(similarity.tolist())), 'values': metadata[1]})
   return similar_songs
 
 
 @app.route("/", methods=['GET', 'POST'])
 def default():
 
     content = request.get_json(silent=True)
-    #print("song", content)
     dataframe = pd.DataFrame.from_dict(json_normalize(content['audio_features']), orient='columns')
-    #print("dataframe", dataframe)
 
     song = dataframe.values
-    #print("content", song)
 
     #song = array[1549]
     similarities = all_similarities(song, dfy)
-    sorted_list = sorted(similarities, key=lambda i: i['similarity'], reverse=True)[1:3]
+    sorted_list = sorted(similarities, key=lambda i: i['similarity'], reverse=True)[:3]
     json_dict = {"songs": sorted_list}
     #data = json.dumps(json_dict)
     return jsonify(json_dict)