/home/frank/temp/jfr-masters/code/frontends/SAX

README

allgraph.go
anomaly_check.go
basecomp.go
bitmap_parallel.go
bitmap_window.go
build-profile.go
check-prop.go
find_base.go
find_base_host.go
graph.go
multi-search-fig.go
multi-search.go
nsutil.go
profile_search.go
sax-test.go
saxutil.go
scomp.go
scomp_local.go
shist.go
subct.go
tsb_comp.go

Most of these Go programs implement SAX conversion and one or more methods
of anomaly detection. check-prop, build-profile, and multi-search are the
most complete versions; the other programs are predecessors of those. Many
of these programs use gnuplot to plot the results. You will need to have 
gnuplot installed. Many of the programs also use the convert program from 
ImageMagick.

The graph and allgraph programs retrieve data from a database and plot it
using gnuplot. They do not analyze the data.

I don't remember why I wrote shist. I apparently builds a histogram
of the values from a time series.

subct builds a subword histogram from a series of words and compares it to
the subword histogram for a single values series (all c's). subct is a
useful tool when you're trying to understand why an anomaly score behaves
the way it does.


gen_gamma.py
gen_impulse.py
gen_normal.py
gen_random.py
gen_unitstep.py
gendata.py
gendata2.py
genimp_noise.py
gensine.py
gensine2f.py
mktxt.py
nsconfig.py
nsutil.py
sine_good2f.py

These Python programs generated artificial series to be analyzed by the
Go programs. Their names are somewhat self explanatory.

goconf.conf
mktxt.conf
profconf.conf

I've sanitized all configuration files so that they do not contain database
login credentials. You'll need to provide your own database.

coreopsis.cs_sessioncount.txt
limestone.kaos_loadavg.txt
coreopsis.cs_sessioncount.prof

These are example data pulled from the database. You can use them as inputs
to the Go programs.

ex1
ex0
real0

These are some slightly more prepared examples that I used. You should
analyze them with the Go programs.
