Wrong subsetting for drop groups in sample annotation

[c-mertes]

When using 2 DROP_GROUP's where one is a substring of the other, the utils.subsetBy function will return false results.

drop/drop/utils.py

Lines 68 to 86 in 8df2d1e

	def subsetBy(df, column, values, exact_match=True):
	"""
	Subset by one or more values of different columns from data frame
	:param df: data frame
	:param column: column to subset by
	:param values: values to subset by
	:param exact_match: default True. when False match substrings. Important for subsetting drop groups
	:return: df subset by values and column
	"""
	if values is None:
	return df
	elif isinstance(values, str) and exact_match :
	return df[df[column] == values]
	elif not isinstance(values,str) and exact_match:
	return df[df[column].isin(values)]
	elif isinstance(values,str) and not exact_match:
	return df[df[column].str.contains(values)]
	else:
	return df[df[column].str.contains("|".join(values))]

It should always be an exact match and not a fuzzy one. Here we should rather use a regex to select the requested rows. As cell entries can be separated by , like in the DROP_GROUP column, we could use something like:

For single string search:

    elif isinstance(values, str) and exact_match :
        return df[df[column].str.contains("(^|,)" + values + "(,|$)")]

For multi string search:

    elif isinstance(values, str) and exact_match :
        return df[df[column].str.contains("(^|,)" + "(" + "|".join([values, "abc"]) + ")" + "(,|$)")]

@mumichae @nickhsmith do you know exactly where this function is used? Only in the SampleAnnotation class?