weird life

 yayyy resting time now with Star Trek watching yayyyy then some topology reading time maybe also. 

 hmm i decided to utilize existing approaches in dependency parsing.  to create the knowledge graph from reading to create the dependency graph based interpretation in neo4j using wikidata ontology and my new grammar ontology definition to outline these dependency graph in such new ontology definition. hmm . hmm. I think i decided to go with initially following methodology: i would create the dependency and span based graphs interpretations of paragraphs. hmm i would also try to create a topological interpretation for verbs in the dependency graph.  e.g. take verb.  could be considered as one topological space giving one constituent to another topological space unless they are not hausdorffly separated in the new topological space. or for instance go verb: connected region in a topological space and a sequence like iterating subset in the topological space.  so i would try to myself create topological ontological rdf definitions for all verbs and their additiona...

 hey today ai project is going mostly with investigating the ontology structure from data. and thinking of how to tie to transformer network and how to also tie to grammar/mathematical language ontology.  I think I would start somehow with some methods not being the very most efficient methodologies but to start. In the end I am not a linguist so my methodology might not be very expertise including.  Hmm transformers, i see as rather a tool to: find correlatedness between parts of sentence and sentences. I might not have translation task but might use translation data sets to encode neural network connectivity to find those Matrix factorization alike inner sentence and per sentences relations. Then I think I need to tie that to wikidata ontology and wordnet knowledge and I might need to also define that in also grammatical ontology and mathematical ontology. so I decided to decide on an ontology storage mechanism right now since seems as I would need to start crafting tho...

 life when having extra grand parents aliens whom communicates with changing Star Trek scripts :) Its amazing :)  I now don't have suspicion that previous Star Trek coincidences were also Alien extra grandparents modification of space time :)  wov:)  I had goose bumps last day when watching some Star Episode's last parts where there were time traveling aliens communicating with Enterprise Captain Picard.  I mean since similar is happening even not being any captain of a Galaxy ship :)  I had goose bumps on some end parts of the episode where time traveling aliens talked.  its totally absolutely amazing :)  wov its so much amazing to have friends whom have time machines and communicates with alteirng my space time's scifi tv shows specifically Star Trek tv show :) From also other incidents I know they are like extra grandparents :) Lots of cheers & love to aliens extra grandparents :) 💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖💖...

 should i cry or laugh at this point whilst I lost 1 hour to setup and debug etc since I forgot to code in java (I code in scala) and converted code to scala etc :DDD 1 hour or passed doing so to see that library actually reads correctly :DDD   then went to see spark code that the rows were read as :DDDD var df_result =df_parsed.select(col("value")(1).as("subject"), col("value")(2).as("object"), col("value")(3).as("predicate")) it should  be as: var df_result =df_parsed.select(col("value")(0).as("subject"), col("value")(1).as("object"), col("value")(2).as("predicate")) :DDDDD Its this line were coming from a version where there were another Id column :D ayyy lost 1 hour just due to this :DDDDDDDDDD

 ayy :D: i created debug environment and debugged to see the data is read as its expected and predicates are not empty then :D turned out the spark code has were also for having 1 more column and other columns were offset 1 column in code that thats why predicate columns were null :DDDD

 hmm I see that there is an error in my modifications to ttl reader because predicate columns is null  :D  nice thing about it it compiles in a hour but i think i can limit it to have compile only jena-arq module hmm.  that would be fine.   hmm where did i made error in code :D????? it does not read the predicate hmm I guess I would debug yepp.  i guess i would resolve this issue in a hour or so then finally can do some data checks to figure out how to handle this massive data to include only one language to remove 3.4 billion rows. i think i would get spot instances to run that task  to first partition the data again and then process up to some partition only initially. hmm so much data.  but seems as if i could resolve these, i might then tomorrow download paragraphs data to craft the word embeddings data afterwards to train transformer network. yayyy