Cooking Events from Recipes Version 1.0 14/3/2015 =============== This bundle contains the cooking recipes and events extracted from them, as used in the event ordering project, conducted in the University of Edinburgh by Omri Abend, Shay Cohen and Mark Steedman. Users of this dataset are kindly asked to cite the following publication: Lexical Event Ordering with an Edge-Factored Model Omri Abend, Shay B. Cohen and Mark Steedman, NAACL 2015. Files included: --------------- 1. Those ending with "mmf": original recipe files, as downloaded from "http://www.ffts.com/recipes.htm". See a list of the original downloaded files below. Each recipe is headed by a title and some meta-data, which are ignored in preprocessing. Only place holders for the file names are provided. In order to get the files, you would need to actually download the ZIP files below and unzip them. Contact the authors if you are having trouble downloading the files. 2. Those ending with "mmf.preprocessed": corresponding preprocessed files. The metadata is removed, recipes are delimited by "END_RECIPE". Preprocessing also adds the pronoun "you" before some cooking-related infinitive verbs, as it somewhat improves the fairly low performance of the parser we used on imperatives. 3. Those ending with "mmf.events.parses" : corresponding parsed recipes. Each sentence is separated by a blank line. Recipes are separated by a line containing "=========". Each parse starts with a phrase-structure parse, followed by the corresponding Stanford Dependencies. 4. Those ending with "mmf.events" : corresponding events and linkage relations extracted from the recipes. Each line is either an event or a linkage relation. Events look like this: EVENT (arg1 [dependency type of arg1]#arg2 [dependency type of arg2] ... )() For example: EVENT Preheat({oven} [dobj]#to 250 {degrees} [prep_to])() EVENT roasting(for 45 {minutes} [prep_for])(SECONDARY_continue) The argument typing is according to Stanford Dependencies. Headwords of arguments are in curly brackets. Secondary verbs (e.g., let, continue) are given in the format SECONDARY_*, where * corresponds to the secondary verb's lemma. Linkages look like this: LINKAGE ( ) Where event1 and event2 are in a similar format to the event lines described above. For example: LINKAGE chill({you} [nsubj]#at least {1} [iobj]#{hour} [dobj]#before {serving} [prepc_before])() serving()() (IN before) 4. train_dev_test_partition.sh is a script that partitions the event and linkage files into the development/train/test split used in the NAACL paper. The train/dev/test files are provided. Downloaded recipe files ----------------------- The files should be prefixed by "http://www.ffts.com/". 10000kz.zip 10000.zip 1000kz.zip 1000.zip 11000kz.zip 11000.zip 12000kz.zip 12000.zip 13000kz.zip 13000.zip 14000kz.zip 14000.zip 15000kz.zip 15000.zip 16000kz.zip 16000.zip 17000kz.zip 17000.zip 18000kz.zip 18000.zip 19000kz.zip 19000.zip 1davcuf.zip 20000kz.zip 20000.zip 2000kz.zip 2000.zip 21000kz.zip 21000.zip 22000kz.zip 22000.zip 23000kz.zip 23000.zip 24000kz.zip 24000.zip 25000kz.zip 25000.zip 26000kz.zip 26000.zip 27000kz.zip 27000.zip 28000kz.zip 28000.zip 29000kz.zip 29000.zip 30000kz.zip 30000.zip 3000kz.zip 3000.zip 31000kz.zip 31000.zip 32000kz.zip 32000.zip 32965.zip 33000kz.zip 33961kz.zip 4000kz.zip 4000.zip 5000kz.zip 5000.zip 6000kz.zip 6000.zip 7000kz.zip 7000.zip 8000kz.zip 8000.zip 9000kz.zip 9000.zip allrecip.zip amish.zip apetizer.zip apple butter.zip asparagu.zip atkins.zip barley.zip bbqsauces.zip biscotti.zip breadmaker.zip bredmake1.zip bredmake.zip brisket.zip brownric.zip cakes02.zip calzone.zip canadian.zip caribou.zip cburg2.zip ccakes.zip ChampChili03.zip ChampChili.zip cheddar.zip cheese.zip chickenb.zip chickwng.zip chilis.zip choclate.zip chocolates.zip cook0001.zip cookbook.zip crabberr.zip crockpot.zip diab722.zip diabetic.zip dips.zip eggnog.zip english.zip ethiopia.zip ffdressn.zip ffdrssn.zip filetmig.zip fruits.zip garlic.zip german.zip gift_mix.zip goose.zip health01.zip health02.zip health03.zip health04.zip healthy.zip holiday.zip kids.zip lemons.zip lentil.zip lg32965.zip lg33961.zip Liqueurs.zip londontn.zip lotasoup.zip lowcarbexport.zip lowfat.zip MammasRecipes.zip mexican.zip mincemea.zip misc2600.zip mm0222-1.zip mm0222-2.zip mm1000a.zip mm1000b.zip mm1000c.zip mm1000d.zip mm1000e.zip mm1000f.zip mm1000g.zip mm1000h.zip mm1000i.zip mm1000j.zip mm1000k.zip mm2155re.zip Mm3500_01.zip Mm3500_02.zip Mm3500_03.zip Mm3500_04.zip Mm3500_05.zip Mm3500_06.zip Mm3500_07.zip Mm3500_08.zip Mm3500_09.zip Mm3500_10.zip Mm3500_11.zip Mm3500_12.zip Mm3500_13.zip Mm3500_14.zip Mm3500_15.zip Mm3500_16.zip Mm3500_17.zip Mm3500_18.zip mmbread.zip mmcanuck.zip mmchic.zip mmcyber5.zip mmdrinks.zip mmfilip.zip mm_gc_ny.zip mmgermn1.zip mmgermn2.zip mm_greek.zip mmgsotw1.zip mmice.zip mmmicwv1.zip mmmicwv2.zip mmpasta.zip mmsoup2.zip mmsoup.zip mm_vegan.zip mm-vegan.zip mmvegy.zip mm_welsh.zip mufftxt.zip mushside.zip mustard.zip noodles.zip Over 600 Canning Recipes.zip pchef1.zip plumpudd.zip porkchop.zip porkgr.zip porkroas.zip porktend.zip pump101.zip punch.zip raisinbe.zip rawfood1.zip ribroast.zip spaghett.zip Spanish recipes.zip squash.zip stuffing.zip swordfis.zip the bubba gump.zip trifle.zip turkey.zip usdafood.zip vegan1.zip vegan2.zip vegan6.zip vegsal.zip wildgame.zip wildrice.zip WW_recipes1.zip WW_recipes2.zip WW.zip xmas.zip The data downloaded belongs to Loginetics, Inc.