safety check to prevent small sub dataframe from overwriting full dataframe on disk

[kushalkolar]

Addresses #125

Since a DataFrame batch path is propogated to any "sub dataframes" created by fancy indexing or other means, this prevents the existing full dataframe on disk from being overwritten by the smaller sub dataframe.

For example:

from mesmerize_core import *

df = load_batch("/path/to/batch.pickle")

# create a sub dataframe
sub_df = df[df["algo"] == "cnmf"]

# this is dangerous! So this PR would prevent it.
sub_df.save_to_disk()

User can still override with the max_index_diff kwarg:

# forced save if number of rows doesn't differ by more than 5
sub_df.save_to_disk(max_index_diff=5)