Adding new ratings and new user and re-training model

Nobody has claimed this yet.

Hello!

I am trying to add the new user with new ratings to the model and retrain it but I am not sure how to do this; I have tried adding the new user to the .csv files that I open, hoping that it will train again on those but I keep getting the embedding error. I would like to add couple of new ratings for a user that is also not in the data set, as in: new_user=[book_id:rating, book_id:rating...]Any help would be appreciated! Thank you!
Here is my model:
```
def get_recommendations(model, coo_mtrx, users_ids, books_id):
n_items = coo_mtrx.shape[1]

# TODO create known positives
# Books the model predicts they will like
scores = model.predict(users_ids, np.arange(n_items))
top_scores = np.argsort(-scores)[:5]

print ('Book recomendations:')
for x in top_scores.tolist():
print(books_id[str(x)])

def main():
#format_files()
data = book(min_score=2)
model = LightFM(loss='warp')
model.fit(data['matrix'], epochs=30, num_threads=2)
# model.fit_partial()
user = '1996'
get_recommendations(model, data['matrix'], user, data['books_id'])

if __name__ == "__main__" :
main()
```

Start at main and get_recommendations, then inspect the LightFM fit and fit_partial calls and the data['matrix'] passed to them. Done means documenting or implementing a clear workflow for adding a user with new ratings and obtaining recommendations without the reported embedding error.

Written by the indexing model from the issue text.