pip install arakhne

from arakhne import Arakhne

# Load a corpus from a CSV
cassius_dio = Arakhne('greek').csv().load('dio_raw.csv')
# Perform on operation on the entire corpus
cassius_dio = cassius_dio.rm_lines()
# Save a corpus
cassius_dio.csv().save('dio_raw_backup.csv')

# Or load, transform, and save, all in one line
Arakhne('greek').csv().load('dio_raw.csv').rm_lines().csv().save('dio_new.csv')

# Removes all endline characters
cassius_dio = cassius_dio.rm_lines()
# Removes texts between editorial marks, e.g. []{}()<>, etc.
cassius_dio = cassius_dio.rm_edits()
# Removes characters not belonging to the specified language
cassius_dio = cassius_dio.rm_nonchars()
# Normalize certain Greek accent character encoding issues
cassius_dio = cassius_dio.normalize()
# Removes redundant whitespace (often created by the above operations)
cassius_dio = cassius_dio.rm_spaces()

# Or, chain them together to perform multiple operations
cassius_dio = cassius_dio.rm_lines().rm_edits().rm_nonchars().rm_spaces()

# Create a lemmatized version for indexing purposes
lemmatized_dio = cassius_dio.lemmatize()

# Create a filtered corpus containing only passages matching a Regex pattern
filtered_dio = stopped_dio.re_search('πολλοί')

# Pass a list to remove specified words
stopped_dio = cassius_dio.rm_stopwords(['ὁ', 'δὲ', 'ὅτι'])

# Using language default stopwords
stopped_dio = cassius_dio.rm_stopwords()

# Or using a specified plaintext list of stop phrases
stopped_dio = cassius_dio.set_stopwords('stopwords_greek.txt').rm_stopwords()

# Returns list of lists, each containing tuples of the word and part of speech
tagged_dio = scrubbed_dio.tag()

# Returns list of lists, each containing strings with entity names
entities_dio = scrubbed_dio.entities()