This repository has been archived on 2024-06-20. You can view files and clone it, but you cannot make any changes to it's state, such as pushing and creating new issues, pull requests or comments.
coffee.pygments/tests/examplefiles/pig/test.pig.output
2022-01-06 18:34:38 +01:00

702 lines
20 KiB
Text
Generated

"/**\n * This script is an example recommender (using made up data) showing how you might modify item-item links\n * by defining similar relations between items in a dataset and customizing the change in weighting.\n * This example creates metadata by using the genre field as the metadata_field. The items with\n * the same genre have it's weight cut in half in order to boost the signals of movies that do not have the same genre.\n * This technique requires a customization of the standard GetItemItemRecommendations macro\n */" Comment.Multiline
'\n' Text.Whitespace
'import' Keyword
' ' Text.Whitespace
"'recommenders.pig'" Literal.String
';' Punctuation
'\n\n\n\n' Text.Whitespace
'%default' Keyword
' ' Text.Whitespace
'INPUT_PATH_PURCHASES' Text
' ' Text.Whitespace
"'../data/retail/purchases.json'" Literal.String
'\n' Text.Whitespace
'%default' Keyword
' ' Text.Whitespace
'INPUT_PATH_WISHLIST' Text
' ' Text.Whitespace
"'../data/retail/wishlists.json'" Literal.String
'\n' Text.Whitespace
'%default' Keyword
' ' Text.Whitespace
'INPUT_PATH_INVENTORY' Text
' ' Text.Whitespace
"'../data/retail/inventory.json'" Literal.String
'\n' Text.Whitespace
'%default' Keyword
' ' Text.Whitespace
'OUTPUT_PATH' Text
' ' Text.Whitespace
"'../data/retail/out/modify_item_item'" Literal.String
'\n\n\n' Text.Whitespace
'/******** Custom GetItemItemRecommnedations *********/' Comment.Multiline
'\n' Text.Whitespace
'define' Keyword
' ' Text.Whitespace
'recsys__GetItemItemRecommendations_ModifyCustom' Name.Function
'(' Punctuation
'user_item_signals,' Text
' ' Text.Whitespace
'metadata' Text
')' Punctuation
' ' Text.Whitespace
'returns' Keyword
' ' Text.Whitespace
'item_item_recs' Text
' ' Text.Whitespace
'{' Punctuation
'\n\n ' Text.Whitespace
'-- Convert user_item_signals to an item_item_graph' Comment
'\n ' Text.Whitespace
'ii_links_raw,' Text
' ' Text.Whitespace
'item_weights' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'recsys__BuildItemItemGraph' Name.Function
'(' Punctuation
'\n ' Text.Whitespace
'$user_item_signals,' Text
'\n ' Text.Whitespace
'$LOGISTIC_PARAM,' Text
'\n ' Text.Whitespace
'$MIN_LINK_WEIGHT,' Text
'\n ' Text.Whitespace
'$MAX_LINKS_PER_USER' Text
'\n ' Text.Whitespace
')' Punctuation
';' Punctuation
'\n ' Text.Whitespace
'-- NOTE this function is added in order to combine metadata with item-item links' Comment
'\n ' Text.Whitespace
'-- See macro for more detailed explination' Comment
'\n ' Text.Whitespace
'ii_links_metadata' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'recsys__AddMetadataToItemItemLinks' Name.Function
'(' Punctuation
'\n ' Text.Whitespace
'ii_links_raw,' Text
'\n ' Text.Whitespace
'$metadata' Text
'\n ' Text.Whitespace
')' Punctuation
';' Punctuation
'\n\n ' Text.Whitespace
'/********* Custom Code starts here ********/' Comment.Multiline
'\n\n ' Text.Whitespace
'--The code here should adjust the weights based on an item-item link and the equality of metadata.' Comment
'\n ' Text.Whitespace
'-- In this case, if the metadata is the same, the weight is reduced. Otherwise the weight is left alone.' Comment
'\n ' Text.Whitespace
'ii_links_adjusted' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'foreach' Keyword
' ' Text.Whitespace
'ii_links_metadata' Text
' ' Text.Whitespace
'generate' Keyword
' ' Text.Whitespace
'item_A,' Text
' ' Text.Whitespace
'item_B,' Text
'\n ' Text.Whitespace
'-- the amount of weight adjusted is dependant on the domain of data and what is expected' Comment
'\n ' Text.Whitespace
'-- It is always best to adjust the weight by multiplying it by a factor rather than addition with a constant' Comment
'\n ' Text.Whitespace
'(' Punctuation
'metadata_B' Text
' ' Text.Whitespace
'=' Operator
'=' Operator
' ' Text.Whitespace
'metadata_A' Text
' ' Text.Whitespace
'?' Operator
' ' Text.Whitespace
'(' Punctuation
'weight' Text
' ' Text.Whitespace
'*' Text
' ' Text.Whitespace
'0.5' Literal.Number.Float
')' Punctuation
':' Text
' ' Text.Whitespace
'weight' Text
')' Punctuation
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'weight;' Text
'\n\n\n ' Text.Whitespace
'/******** Custom Code stops here *********/' Comment.Multiline
'\n\n ' Text.Whitespace
'-- remove negative numbers just incase' Comment
'\n ' Text.Whitespace
'ii_links_adjusted_filt' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'foreach' Keyword
' ' Text.Whitespace
'ii_links_adjusted' Text
' ' Text.Whitespace
'generate' Keyword
' ' Text.Whitespace
'item_A,' Text
' ' Text.Whitespace
'item_B,' Text
'\n ' Text.Whitespace
'(' Punctuation
'weight' Text
' ' Text.Whitespace
'<=' Operator
' ' Text.Whitespace
'0' Literal.Number.Integer
' ' Text.Whitespace
'?' Operator
' ' Text.Whitespace
'0' Literal.Number.Integer
':' Text
' ' Text.Whitespace
'weight' Text
')' Punctuation
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'weight;' Text
'\n ' Text.Whitespace
'-- Adjust the weights of the graph to improve recommendations.' Comment
'\n ' Text.Whitespace
'ii_links' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'recsys__AdjustItemItemGraphWeight' Name.Function
'(' Punctuation
'\n ' Text.Whitespace
'ii_links_adjusted_filt,' Text
'\n ' Text.Whitespace
'item_weights,' Text
'\n ' Text.Whitespace
'$BAYESIAN_PRIOR' Text
'\n ' Text.Whitespace
')' Punctuation
';' Punctuation
'\n\n ' Text.Whitespace
'-- Use the item-item graph to create item-item recommendations.' Comment
'\n ' Text.Whitespace
'$item_item_recs' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'recsys__BuildItemItemRecommendationsFromGraph' Name.Function
'(' Punctuation
'\n ' Text.Whitespace
'ii_links,' Text
'\n ' Text.Whitespace
'$NUM_RECS_PER_ITEM,' Text
'\n ' Text.Whitespace
'$NUM_RECS_PER_ITEM' Text
'\n ' Text.Whitespace
')' Punctuation
';' Punctuation
'\n' Text.Whitespace
'}' Punctuation
';' Punctuation
'\n\n\n' Text.Whitespace
'/******* Load Data **********/' Comment.Multiline
'\n\n' Text.Whitespace
'--Get purchase signals' Comment
'\n' Text.Whitespace
'purchase_input' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'load' Keyword
' ' Text.Whitespace
"'$INPUT_PATH_PURCHASES'" Literal.String
' ' Text.Whitespace
'using' Keyword
' ' Text.Whitespace
'org.apache.pig.piggybank.storage.JsonLoader' Text
'(' Punctuation
'\n ' Text.Whitespace
"'row_id: " Text
'int' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'movie_id' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'movie_name' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'user_id' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'purchase_price' Text
':' Text
' ' Text.Whitespace
'int' Keyword.Type
"');\n\n" Text
'--Get wishlist signals' Comment
'\n' Text.Whitespace
'wishlist_input' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'load' Keyword
' ' Text.Whitespace
"'$INPUT_PATH_WISHLIST'" Literal.String
' ' Text.Whitespace
'using' Keyword
' ' Text.Whitespace
'org.apache.pig.piggybank.storage.JsonLoader' Text
'(' Punctuation
'\n ' Text.Whitespace
"'row_id: " Text
'int' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'movie_id' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'movie_name' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'user_id' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
"');\n\n\n" Text
'/******* Convert Data to Signals **********/' Comment.Multiline
'\n\n' Text.Whitespace
'-- Start with choosing 1 as max weight for a signal.' Comment
'\n' Text.Whitespace
'purchase_signals' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'foreach' Keyword
' ' Text.Whitespace
'purchase_input' Text
' ' Text.Whitespace
'generate' Keyword
'\n ' Text.Whitespace
'user_id' Text
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'user,' Text
'\n ' Text.Whitespace
'movie_name' Text
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'item,' Text
'\n ' Text.Whitespace
'1.0' Literal.Number.Float
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'weight;' Text
'\n\n\n' Text.Whitespace
'-- Start with choosing 0.5 as weight for wishlist items because that is a weaker signal than' Comment
'\n' Text.Whitespace
'-- purchasing an item.' Comment
'\n' Text.Whitespace
'wishlist_signals' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'foreach' Keyword
' ' Text.Whitespace
'wishlist_input' Text
' ' Text.Whitespace
'generate' Keyword
'\n ' Text.Whitespace
'user_id' Text
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'user,' Text
'\n ' Text.Whitespace
'movie_name' Text
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'item,' Text
'\n ' Text.Whitespace
'0.5' Literal.Number.Float
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'weight;' Text
'\n\n' Text.Whitespace
'user_signals' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'union' Keyword
' ' Text.Whitespace
'purchase_signals,' Text
' ' Text.Whitespace
'wishlist_signals;' Text
'\n\n\n' Text.Whitespace
'/******** Changes for Modifying item-item links ******/' Comment.Multiline
'\n' Text.Whitespace
'inventory_input' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'load' Keyword
' ' Text.Whitespace
"'$INPUT_PATH_INVENTORY'" Literal.String
' ' Text.Whitespace
'using' Keyword
' ' Text.Whitespace
'org.apache.pig.piggybank.storage.JsonLoader' Text
'(' Punctuation
'\n ' Text.Whitespace
"'movie_title: " Text
'chararray' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'genres' Text
':' Text
' ' Text.Whitespace
'bag' Keyword
'{' Punctuation
'tuple' Keyword.Type
'(' Punctuation
'content' Text
':' Text
'chararray' Keyword.Type
')' Punctuation
'}' Punctuation
"');\n\n\n" Text
'metadata' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'foreach' Keyword
' ' Text.Whitespace
'inventory_input' Text
' ' Text.Whitespace
'generate' Keyword
'\n ' Text.Whitespace
'FLATTEN' Keyword
'(' Punctuation
'genres' Text
')' Punctuation
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'metadata_field,' Text
'\n ' Text.Whitespace
'movie_title' Text
' ' Text.Whitespace
'as' Keyword
' ' Text.Whitespace
'item;' Text
'\n' Text.Whitespace
'-- requires the macro to be written seperately' Comment
'\n ' Text.Whitespace
'--NOTE this macro is defined within this file for clarity' Comment
'\n' Text.Whitespace
'item_item_recs' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'recsys__GetItemItemRecommendations_ModifyCustom' Name.Function
'(' Punctuation
'user_signals,' Text
' ' Text.Whitespace
'metadata' Text
')' Punctuation
';' Punctuation
'\n' Text.Whitespace
'/******* No more changes ********/' Comment.Multiline
'\n\n\n' Text.Whitespace
'user_item_recs' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'recsys__GetUserItemRecommendations' Name.Function
'(' Punctuation
'user_signals,' Text
' ' Text.Whitespace
'item_item_recs' Text
')' Punctuation
';' Punctuation
'\n\n' Text.Whitespace
'--Completely unrelated code stuck in the middle' Comment
'\n' Text.Whitespace
'data' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'LOAD' Keyword
' ' Text.Whitespace
"'s3n://my-s3-bucket/path/to/responses'" Literal.String
'\n ' Text.Whitespace
'USING' Keyword
' ' Text.Whitespace
'org.apache.pig.piggybank.storage.JsonLoader' Text
'(' Punctuation
')' Punctuation
';' Punctuation
'\n' Text.Whitespace
'responses' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'FOREACH' Keyword
' ' Text.Whitespace
'data' Text
' ' Text.Whitespace
'GENERATE' Keyword
' ' Text.Whitespace
'object' Text
'#' Operator
"'response'" Literal.String
' ' Text.Whitespace
'AS' Keyword
' ' Text.Whitespace
'response' Text
':' Text
' ' Text.Whitespace
'map' Keyword
'[' Punctuation
']' Punctuation
';' Punctuation
'\n' Text.Whitespace
'out' Text
' ' Text.Whitespace
'=' Operator
' ' Text.Whitespace
'FOREACH' Keyword
' ' Text.Whitespace
'responses' Text
'\n ' Text.Whitespace
'GENERATE' Keyword
' ' Text.Whitespace
'response' Text
'#' Operator
"'id'" Literal.String
' ' Text.Whitespace
'AS' Keyword
' ' Text.Whitespace
'id' Text
':' Text
' ' Text.Whitespace
'int' Keyword.Type
',' Operator
' ' Text.Whitespace
'response' Text
'#' Operator
"'thread'" Literal.String
' ' Text.Whitespace
'AS' Keyword
' ' Text.Whitespace
'thread' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
',' Operator
'\n ' Text.Whitespace
'response' Text
'#' Operator
"'comments'" Literal.String
' ' Text.Whitespace
'AS' Keyword
' ' Text.Whitespace
'comments' Text
':' Text
' ' Text.Whitespace
'{' Punctuation
't' Text
':' Text
' ' Text.Whitespace
'(' Punctuation
'comment' Text
':' Text
' ' Text.Whitespace
'chararray' Keyword.Type
')' Punctuation
'}' Punctuation
';' Punctuation
'\n' Text.Whitespace
'STORE' Keyword
' ' Text.Whitespace
'out' Text
' ' Text.Whitespace
'INTO' Keyword
' ' Text.Whitespace
"'s3n://path/to/output'" Literal.String
' ' Text.Whitespace
'USING' Keyword
' ' Text.Whitespace
'PigStorage' Name.Builtin
'(' Punctuation
"'|'" Literal.String
')' Punctuation
';' Punctuation
'\n\n\n' Text.Whitespace
'/******* Store recommendations **********/' Comment.Multiline
'\n\n' Text.Whitespace
'-- If your output folder exists already, hadoop will refuse to write data to it.' Comment
'\n\n' Text.Whitespace
'rmf' Keyword
' ' Text.Whitespace
'$OUTPUT_PATH/item_item_recs;' Text
'\n' Text.Whitespace
'rmf' Keyword
' ' Text.Whitespace
'$OUTPUT_PATH/user_item_recs;' Text
'\n\n' Text.Whitespace
'store' Keyword
' ' Text.Whitespace
'item_item_recs' Text
' ' Text.Whitespace
'into' Keyword
' ' Text.Whitespace
"'$OUTPUT_PATH/item_item_recs'" Literal.String
' ' Text.Whitespace
'using' Keyword
' ' Text.Whitespace
'PigStorage' Name.Builtin
'(' Punctuation
')' Punctuation
';' Punctuation
'\n' Text.Whitespace
'store' Keyword
' ' Text.Whitespace
'user_item_recs' Text
' ' Text.Whitespace
'into' Keyword
' ' Text.Whitespace
"'$OUTPUT_PATH/user_item_recs'" Literal.String
' ' Text.Whitespace
'using' Keyword
' ' Text.Whitespace
'PigStorage' Name.Builtin
'(' Punctuation
')' Punctuation
';' Punctuation
'\n\n' Text.Whitespace
'-- STORE the item_item_recs into dynamo' Comment
'\n' Text.Whitespace
'STORE' Keyword
' ' Text.Whitespace
'item_item_recs' Text
'\n ' Text.Whitespace
'INTO' Keyword
' ' Text.Whitespace
"'$OUTPUT_PATH/unused-ii-table-data'" Literal.String
'\n' Text.Whitespace
'USING' Keyword
' ' Text.Whitespace
'com.mortardata.pig.storage.DynamoDBStorage' Text
'(' Punctuation
"'$II_TABLE'" Literal.String
',' Operator
' ' Text.Whitespace
"'$AWS_ACCESS_KEY_ID'" Literal.String
',' Operator
' ' Text.Whitespace
"'$AWS_SECRET_ACCESS_KEY'" Literal.String
')' Punctuation
';' Punctuation
'\n\n' Text.Whitespace
'-- STORE the user_item_recs into dynamo' Comment
'\n' Text.Whitespace
'STORE' Keyword
' ' Text.Whitespace
'user_item_recs' Text
'\n ' Text.Whitespace
'INTO' Keyword
' ' Text.Whitespace
"'$OUTPUT_PATH/unused-ui-table-data'" Literal.String
'\n' Text.Whitespace
'USING' Keyword
' ' Text.Whitespace
'com.mortardata.pig.storage.DynamoDBStorage' Text
'(' Punctuation
"'$UI_TABLE'" Literal.String
',' Operator
' ' Text.Whitespace
"'$AWS_ACCESS_KEY_ID'" Literal.String
',' Operator
' ' Text.Whitespace
"'$AWS_SECRET_ACCESS_KEY'" Literal.String
')' Punctuation
';' Punctuation
'\n' Text.Whitespace