The package is everything needed for sophisticated searches and includes: the basic join, a memory attachment for chaining, the lazyJ attachment, plaint text, and console printf. Not completely bug proof, but in production form. Enough to search groups of words against raw text, and cross reference to other groups of words with multiple refinements.
The memory attachment is an extremely fast array of sequence records with index jmping instread of crawling a string. So typical searches load up memory with the raw words, in soem groping, then go over that with fast, short search graphs. The mem attachment used for most applications.
Take text processing. We might consider a ready set of key word groups. A set of key words found in a50 different contexts, 40 words each. Easy to do in Dot5,Comma, lazyJ. One then, generally run the raw text through plaint text and extract its words, comparing them to hierarchical groups already in memory, in naive form for the join. Your search graph produces result graphs, one or more depending upn the implied match logic. We get extremely fast, multi level, up to five, six deep, classification logic.Standard lazyJ output, if you want.
The memory attachment is an extremely fast array of sequence records with index jmping instread of crawling a string. So typical searches load up memory with the raw words, in soem groping, then go over that with fast, short search graphs. The mem attachment used for most applications.
Take text processing. We might consider a ready set of key word groups. A set of key words found in a50 different contexts, 40 words each. Easy to do in Dot5,Comma, lazyJ. One then, generally run the raw text through plaint text and extract its words, comparing them to hierarchical groups already in memory, in naive form for the join. Your search graph produces result graphs, one or more depending upn the implied match logic. We get extremely fast, multi level, up to five, six deep, classification logic.Standard lazyJ output, if you want.
No comments:
Post a Comment