A visual database query language implemented in AVS

Dataflow

These genetic sequence database queries are implemented within the AVS (Version 5) scientific visualization software. The data flows through the networks from top to bottom.

Network 2
	This network shows an OR/AND query where (in a text-based query language) parentheses would be necessary to do the OR first: ( string == "pat1" OR string == "pat2" ) AND string == "pat3"

Network 3
	This network shows a simple Database `miner'. Here's an instance where data flow makes it work, because the lower search_db will not execute until all of the inputs are ready. In pseudo shell code it might be for str in [ string == "pat" ]   str == "pat" with an implicit |= between each interation of the for loop.

Network 1

This is a network that I was using for Nanogen. The read_sstr module introduces a list of search strings. It acts like the for loop in Network 2. Again the strings are OR'ed and the result of the leftmost branch of the network contains the records that matched at least one of the search strings from read_sstr. The right search branch performs a simple AND and then the two branches are OR'ed together with the merge_db. This result is in turn searched. In pseudo code: (str=="RS1" OR ... OR str=="RS2" OR str=="P1" AND str=="P2") AND str=="P3" The details were: the RS1 were the genus names of a bunch of interesting path bacteria, P1 is E.coli, P2 is the pattern \<[0O]157\>, the combination selecting on enterohemolytic E. coli, and the last search asks for all the of the features that describe `named' products. The search_db modules are set to pass along only new hits, and the select_field modules adds the organism name back. The extract_hits gets the actual character strings from the database, and the two group_output modules collect similar names into groups. The group_output on the left groups organisms by genus, the one the right groups named products. These three inputs (the two groups are optional) are fed into a table maker that creates a table that contains organisms as rows, products as colums. `Pages' of the table are viewed via a web browser.