Sha256: 8edc9363df2ba292d8199cac0a84fc5d468a54103cf09100805cb692585c187e
Contents?: true
Size: 1.24 KB
Versions: 27
Compression:
Stored size: 1.24 KB
Contents
module EasyML
module Data
module Splits
class InMemorySplit < Split
attr_accessor :dataset
def initialize(options = {})
@data = {}
@dataset = options[:dataset]
end
# We don't backup in-memory splits to s3
def download; end
def upload; end
def files
[]
end
def save(segment, df)
@data[segment] = df
end
def read(segment, split_ys: false, target: nil, drop_cols: [], filter: nil, limit: nil, select: nil,
unique: nil)
return nil if @data.keys.none?
df = if segment.to_s == "all"
Polars.concat(EasyML::Dataset::SPLIT_ORDER.map { |segment| @data[segment] }.compact)
else
@data[segment]
end
return nil if df.nil?
df = df.filter(filter) if filter.present?
drop_cols &= df.columns
df = df.drop(drop_cols) unless drop_cols.empty?
df = df.unique if unique
split_features_targets(df, split_ys, target)
end
def cleanup
@data.clear
end
def split_at
@data.keys.empty? ? nil : Time.now
end
end
end
end
end
Version data entries
27 entries across 27 versions & 1 rubygems