FastTabular.jl

			
			
			
			module
			
			 
			FastTabular
			
			

			

			
			using
			
			 

	
			FastAI
			

			
			
			using
			
			 

	
			FastAI
			:
			
			 
			# blocks
			
              
			Block
			,
			
			 
			WrapperBlock
			,
			
			 
			AbstractBlock
			,
			
			 
			OneHotTensor
			,
			
			 
			OneHotTensorMulti
			,
			
			 

	
			Label
			,
			
			
              

	
			LabelMulti
			,
			
			 
			wrapped
			,
			
			 

	
			Continuous
			,
			
			 
			getencodings
			,
			
			 
			getblocks
			,
			
			 
			encodetarget
			,
			
			
              
			encodeinput
			,
			
			

			# encodings
			
              
			Encoding
			,
			
			 
			StatefulEncoding
			,
			
			 

	
			OneHot
			,
			
			

			# visualization
			
              

	
			ShowText
			,
			
			

			# other
			
              

	
			Context
			,
			
			 

	
			Training
			,
			
			 

	
			Validation
			

			
			
			import
			
			 

	
			FastAI
			:
			
			 

	
			Datasets
			

			
			using
			
			 

	
			FastAI
			.

	
			Datasets
			

			

			# for tests
			

			
			
			using
			
			 

	
			FastAI
			:
			
			 
			testencoding
			

			

			# extending
			

			
			
			import
			
			 

	
			FastAI
			:
			
			
               

	
			Datasets
			,
			
			
               

	
			blockmodel
			,
			
			 

	
			blockbackbone
			,
			
			 

	
			blocklossfn
			,
			
			 

	
			encode
			,
			
			 

	
			decode
			,
			
			 

	
			checkblock
			,
			
			
               

	
			encodedblock
			,
			
			 

	
			decodedblock
			,
			
			 

	
			showblock!
			,
			
			 

	
			mockblock
			,
			
			 

	
			setup
			

			

			
			import
			
			 
			CSV
			

			
			import
			
			 

	
			DataAugmentation
			

			
			
			import
			
			 
			DataFrames
			:
			
			 
			DataFrame
			,
			
			 
			nrow
			

			
			
			import
			
			 

	
			MLUtils
			:
			
			 

	
			MLUtils
			,
			
			 

	
			eachobs
			,
			
			 

	
			getobs
			,
			
			 

	
			numobs
			

			
			import
			
			 

	
			Flux
			

			
			
			import
			
			 

	
			Flux
			:
			
			 
			Embedding
			,
			
			 
			Chain
			,
			
			 
			Dropout
			,
			
			 
			Dense
			,
			
			 
			Parallel
			,
			
			 
			BatchNorm
			

			
			import
			
			 
			PrettyTables
			

			
			
			import
			
			 
			ShowCases
			:
			
			 
			ShowCase
			

			
			import
			
			 
			Tables
			

			
			import
			
			 
			Statistics
			

			
			using
			
			 
			FilePathsBase
			

			

			
			using
			
			 
			InlineTest
			

			

			
			include
			(
			
			"
			container.jl
			"
			)
			

			

			# Blocks
			

			
			include
			(
			
			"
			blocks/tablerow.jl
			"
			)
			

			

			# Encodings
			

			
			include
			(
			
			"
			encodings/tabularpreprocessing.jl
			"
			)
			

			

			
			include
			(
			
			"
			models.jl
			"
			)
			

			

			
			const
			
			 
			_tasks
			 
			=
			 
			
			
			Dict
			{
			String
			,
			 
			Any
			}
			(
			)
			

			
			include
			(
			
			"
			tasks/classification.jl
			"
			)
			

			
			include
			(
			
			"
			tasks/regression.jl
			"
			)
			

			
			include
			(
			
			"
			recipes.jl
			"
			)
			

			

			
			function
			 
			
			__init__
			(
			)
			
			
    
			
			
			

	
			FastAI
			.
			

	
			Registries
			.
			
			registerrecipes
			(
			
			@
			__MODULE__
			,
			 

	
			RECIPES
			)
			
    
			
			
			foreach
			(
			
			values
			(
			_tasks
			)
			)
			 
			do
			
			 
			t
			
			
        
			
			if
			 
			
			!
			
			haskey
			(
			
			learningtasks
			(
			)
			,
			 
			
			t
			.
			
			id
			)
			
			
            
			
			push!
			(
			
			learningtasks
			(
			)
			,
			 
			t
			)
			
        
			end
			
    
			end
			

			end
			

			

			
			export
			 

	
			TableRow
			,
			 

	
			TabularPreprocessing
			,
			 

	
			TabularClassificationSingle
			,
			 

	
			TabularRegression
			,
			
       

	
			TableDataset
			

			

			end