schedule

scala/sbt jvm memory => solution: sbt -mem 4096
git it!!
newman gen_cinfo sorted incorrect?
murata
muratatri
trait KPartiteGraph writed
	func modularity
	func allposdq => c's local layer, local CID
	func mv_nd(layer, nid, cid) # newman compliant
	func calc_dq
	func updateE updateC
combiner: trait test
complete the code
combiner
make a entry point to run with args
completed: ufinder, kfinder, hfinder
split dS, dM
do the all dmlxy in kf
java run scala: java -cp scala-library.jar;. hetcom.hetmain
merger.scala
MERGEC! MERGEC! MERGEC!: PYVER CHECK IF ITS OK DO SCALA
to cache_lxy: size change: src to 0, dst to s+d. all edges passing through them are changed. all c in throught edges need recalc
PAR CALC: not doing it
test alldq
test mergeq
use dlxy - ds, dm
write dQ_merge
write FU instead of pack-unpack
try  1.not to use dtof to see if it's being faster (or in pyver use cProfile)
try  2.sort CE_cnt to calc dQlist?
optimize speed: round loop and merge

[next]
balance accuracy and speed:
	mod definition of M
	minimize method - mix merge and move:
		by loop
		inside loop
		move after merge
		half/quarter loop

report:
	bipartite
	real data visual
	bigger tripartite
	try merger version of hfinder? // parallel subgraph calc may solve to few cmus

package:
	multimain objects in one package
	java -cp hetcom.jar {package}.{object} ...(parameters)
	
be faster:
	not to calc whole newclnk, use cache_lxy to get src part, twice faster
	try  3.only calculate adjacent node
	do a profile?
	maybe mod trinetfinder