 kyle
|
660c85c7d2
add notes and save from two months ago, shelving project for a bit
|
1 gadu atpakaļ |
|
kyle
|
d2a742cd7a
Merge branch 'master' of http://git.thestrengthfoundry.com/kyle/radon-ble-rd200
|
2 gadi atpakaļ |
 Koyle Coil
|
d08ad1c7d7
Merge branch 'master' of http://git.thestrengthfoundry.com/kyle/radon-ble-rd200
|
2 gadi atpakaļ |
|
Koyle Coil
|
ebb6fca22c
added a script for detecting when rd200.py is failing
|
2 gadi atpakaļ |
|
kyle
|
d414fe37fd
gets the last reading from mySQL, kinda slow, but up-to-date
|
2 gadi atpakaļ |
|
kyle
|
3c81e3b1f4
added FIR_OFF, turned off most verbose logging for cache database code
|
2 gadi atpakaļ |
|
kyle
|
5654fdeee3
script for cron to update cache
|
2 gadi atpakaļ |
|
kyle
|
85dea3559d
nothing
|
2 gadi atpakaļ |
|
kyle
|
ff35f47bc1
tested query options with frontend, seems okay
|
2 gadi atpakaļ |
|
kyle
|
47952f8285
added a frontend interface so that frontend can start using cached data
|
2 gadi atpakaļ |
|
kyle
|
e5960b4f0b
all queries will have start and end times defined by default
|
2 gadi atpakaļ |
|
kyle
|
aca0e6e60f
initial import for a year is down to 13 minutes, partial updates to cache can be had in about 4 secaonds for all 420 datasets
|
2 gadi atpakaļ |
|
kyle
|
e6caf14a2e
ingore snow if melt is found
|
2 gadi atpakaļ |
|
kyle
|
73739a824f
initial
|
2 gadi atpakaļ |
|
kyle
|
9281f5b3f3
most tests in rdp-algo are working, still a bit slow, but now 13 minutes for one year and 1 minute for a couple months
|
2 gadi atpakaļ |
|
kyle
|
f44720c584
all data sets are generated, trying to improve runtime, currently takes half an hour on an 800MHz CPU
|
2 gadi atpakaļ |
|
kyle
|
179e6306e0
merge
|
2 gadi atpakaļ |
|
kyle
|
cf7697ee98
try to make the code a little clearer
|
2 gadi atpakaļ |
|
Koyle Coil
|
f8b60fc920
adds some notes to help troubleshoot issues, also attempt to fix the error recovery process in `bluetoothctl`
|
2 gadi atpakaļ |
|
Koyle Coil
|
22cff4d881
attempts to fix a problem in the bluez org software, avoids a call to code that I think does not work
|
2 gadi atpakaļ |
|
Koyle Coil
|
055d44d352
Merge branch 'master' of http://git.thestrengthfoundry.com/kyle/radon-ble-rd200
|
2 gadi atpakaļ |
|
kyle
|
f5c8ca2c87
after profiling PHP 5.3, found that keyed array is not too bad versus internally keyed, so cleaned up the code
|
2 gadi atpakaļ |
|
kyle
|
3dcaa832dd
see previous commit, this cleans up the A/B profiling that was used to determine lowest memory usage by PHP 5.3 (separate arrays for x and y values rather than an array of [x,y])
|
2 gadi atpakaļ |
|
kyle
|
87491c61e5
tested okay with PHP 5.3 with 128Megabytes of memory and 100K initial points
|
2 gadi atpakaļ |
|
kyle
|
cf2f80ac49
adds abstraction to sensor readings for charting usage in frontend, also adds some sensor properties that detail how often they sample as well as assumptions about their noise relative to some arbitrary amount of time
|
2 gadi atpakaļ |
|
kyle
|
0effc7ea0c
tested working code using chartjs to prove that rdp and filters work;\
|
2 gadi atpakaļ |
|
kyle
|
43fa3008c8
filters and caching for sampled data after post-processing using RDP and other downsampling compression algorithms
|
2 gadi atpakaļ |
|
kyle
|
fc49567a7c
Merge branch 'master' of http://git.thestrengthfoundry.com/kyle/radon-ble-rd200
|
2 gadi atpakaļ |
|
kyle
|
2cc4de1a6a
ramer-douglas-peucker algorithm for reducing the number of points used to represent a line
|
2 gadi atpakaļ |
|
Koyle Coil
|
9339125887
found a problem, did not find source
|
2 gadi atpakaļ |
