featurizer.apply(docs=train_docs) fails on clearing

[HiromuHota]

Describe the bug
A clear and concise description of what the bug is.

Giving docs asfeaturizer.apply(docs=train_docs) instead of split as featurizer.apply(split=0) fails at fonduer.features.featurizer.Featurizer#clear

To Reproduce
Steps to reproduce the behavior:

1. Give docs asfeaturizer.apply(docs=train_docs) instead of split as featurizer.apply(split=0)

Expected behavior
A clear and concise description of what you expected to happen.

featurizer works when docs are given.

Error Logs/Screenshots
If applicable, add error logs or screenshots to help explain your problem.

[2019-05-02 11:46:10,363][INFO] fonduer.features.featurizer:190 - Clearing Features (split ALL)
Traceback (most recent call last):
  File "/Users/hiromu/miniconda3/envs/fonduer-mlflow/lib/python3.7/site-packages/sqlalchemy/engine/base.py", line 1244, in _execute_context
    cursor, statement, parameters, context
  File "/Users/hiromu/miniconda3/envs/fonduer-mlflow/lib/python3.7/site-packages/sqlalchemy/engine/default.py", line 552, in do_execute
    cursor.execute(statement, parameters)
psycopg2.errors.InvalidTextRepresentation: invalid input syntax for integer: "ALL"
LINE 5: WHERE candidate.split = 'ALL')

Environment (please complete the following information):
Reproducible on any env.

Additional context
Add any other context about the problem here.

[HiromuHota]

A workaround is to use clear=False as featurizer.apply(train_docs, clear=False).