def _recommend(cfg: dict):
top_n = cfg['top_n']
cfg_model, cfg_dataset = cfg['model'], cfg['dataset']
cfg_results = cfg['results']
_, col_files = cfg_dataset['cols']
x_df = utils \
.read_csv(cfg_dataset['path'], usecols=cfg_dataset['cols']) \
.pipe(utils.to_list_of_strings, col=col_files)
model_recommender = model.deserialize(cfg_model['path'])
y_df = model.recommend(model_recommender,
x_df,
cfg_dataset['cols'],
top_n=top_n)
if cfg_results.get('save', True):
utils.save_csv(cfg_results['out'], y_df)
def root(id=None):
count = request.query.count
if count != '':
count = int(count)
else:
count = 10
if id is None:
id = model_.id_to_article_id[0]
matrix_id = model_.article_id_to_id[id]
recommendations = model.recommend(model_, matrix_id, count)
# for id, distance in recommendations:
# print('{} => {}: {}'.format(model_.id_to_article_id[matrix_id],
# model_.id_to_article_id[id], distance))
items = [model_.id_to_article_id[id] for id, _ in recommendations]
return ','.join(items) + '\n'
def __init__(self):
self.run = recommend()
self.run.read_from_file('data/literature_data.txt')
self.result = list()
async def song_recommend(song_id: int):
#make recommendations
#by default, no. of recs is 10
recs = recommend(item_id = song_id)
return recs