Rewrite Simulation code

[Marc-Demoustier]

The current simulation code is hard to read and not very optimized. The previous simulation code is the baseline for all the simulations code we have in the lab.

If this code is improved we can easily imagine that every code that takes inspiration can be written faster and with a better code quality.

However, it would be even better to have only one version that is very optimized that can be imported in every language if needed.

[Marc-Demoustier]

The simulation 2D is now implemented with the new separations between functions (still need some updates for the domain and mesh) but I am facing an issue where the diffusion coefficient is off if I use the fit function from the previous version. It looks like some of the pixel values are pretty high, maybe it is linked to this.

It can be due to multiple things: (if it has been tested and we know it's not this, the box will be checked)

• [x] Wrong random poisson generator implementation
• Wrong emissions of photons
• [x] Wrong particle position updating

After this the simulation should be good

[Marc-Demoustier]

After analyzing the values of all variables it looks like that this is not where the issue is (see the end of this comment to see all variables values). The idea was to debug both code and see where we can observe some changes. And it looks like the same, both code are doing the same things with random numbers being in the correct range.

However, we can observe some differences in term of execution time: 01:05 (old code launched in IntelliJ IDE) vs 01:47 (new code launched in IntelliJ IDE). The issue is most likely due to too many iterations happening for some reason. Need to identify this issue.

Here are the logs: Old code:

this = {Imaging_FCS@2593} 
simSeed = 1
simNoParticles = 1000
simCPS = 10000
simTauBleach = 0.0
simPixelnum = 21
simExtFactor = 1.5
simNoTStep = 50000
simFrameTime = 0.001
simStepsPerFrame = 10
simD1 = 1.0E-12
simDoutDinRatio = 1.0
simD2 = 1.0E-13
simD3 = 1.0E-14
simF2 = 0.0
simF3 = 0.0
simKon = 300.0
simKoff = 700.0
simCameraOffset = 100
simCameraNoiseFactor = 3.0
simBleachRadius = 3.0E-6
simBleachFrame = 1000000
simDomainRadius = 3.0E-8
simDomainDensity = 30.0
simPin = 1.0
simPout = 0.6
simMeshworkSize = 1.0E-7
simHopProbability = 1.0
simPixelSizeRS = 2.4E-5
simMag = 100.0
simWavelength = 5.15E-7
simNA = 1.49
simSigma0 = 0.8
newSimSettings = {String[36]@3172} ["2D (free)", "Triplet off", "1", "1000", "10000", +31 more]
simTStep = 1.0E-4
simDarkF = 0.7
numOfSeeds = 18
simSeedArray = {int[18]@3403} [1, 2, 5, 10, 17, 26, 37, 50, 65, 82, 101, 122, 145, 170, 197, 226, 257, 290]
simPixelSize = 2.4000000000000003E-7
simPSFSize = 1.3825503355704699E-7
simGridSize = 5.040000000000001E-6
simMidPos = 2.5200000000000004E-6
simSizeLL = -7.560000000000001E-6
simSizeUL = 7.560000000000001E-6
simDetectorSize = 2.5200000000000004E-6
bleachFactor = 2.0
blinkFactor = {double[2]@3434} [0.9704455335485082, 0.9323938199059483]
gridLength = 1.5120000000000003E-5
gridMidPos = 7.560000000000001E-6
numberofdomains = 6859
domains = {double[6859][]@3453} 
subgridnum = 52
subgridsize = 2.964705882352942E-7
maxdomainperarea = 391
maxct = 0
domainsorted = {int[1][][]@3469} 
dsortmp = {int[52][][]@3473} 
dctr = {int[52][]@3477} 
num1 = 1000
num2 = 0
particles = {double[1000][]@3487} 
impSim = {ImagePlus@3498} "img["2D Simulation" (-2), 16-bit, 21x21x1x50000x1]"
cs = 18
rugxpos = {UniformGenerator@3540} 
rugypos = {UniformGenerator@3541} 
ruig = {UniformGenerator@3542} 
rugbf = {UniformGenerator@3543} 
rugpin = {UniformGenerator@3544} 
rugpout = {UniformGenerator@3545} 
rugphop = {UniformGenerator@3546} 
rugblink = {UniformGenerator@3547} 
rgg1 = {GaussianGenerator@3548} 
rgg2 = {GaussianGenerator@3549} 
rgg3 = {GaussianGenerator@3550} 
rggpsf = {GaussianGenerator@3551} 
rpgphoton = {PoissonGenerator@3552} 
rggnoise = {GaussianGenerator@3553} 
rggdom1 = {GaussianGenerator@3554} 
rggdom2 = {GaussianGenerator@3555} 
rggdom3 = {GaussianGenerator@3556} 
rggdrad = {GaussianGenerator@3557} 
n = 0
ipSim = {ShortProcessor@3614} "ip[width=21, height=21, bits=16, min=0.0, max=65535.0]"
s = 0
m = 20
dx = 1.2191758223123185E-8
dy = -1.3098260873548387E-9
index = 1
nop = 2
p = 0
cordx = -1.239981312765639E-6
cordy = 1.6351952106872658E-6
tpx = 5
tpy = 17
tmp = 97
simSettings = {String[36]@2629} [null, null, null, null, null, +31 more]
batchSim = false
nosimsettings = 36
imp = null
particle=[-1.2095871119180109E-6, 1.8302337335402536E-6, 1.0, 0.0, 1.0]

New code:

this = {Simulation2D@2631} 
 model = {SimulationModel@2633} 
  expSettingsModel = {ExpSettingsModel@2637} 
   binning = {Point@2772} "java.awt.Point[x=1,y=1]"
   CCF = {Dimension@2773} "java.awt.Dimension[width=0,height=0]"
   pixelSize = 24.0
   magnification = 100.0
   NA = 1.49
   sigma = 0.8
   emLambda = 515.0
   sigma2 = 0.8
   emLamdba2 = 600.0
   sigmaZ = 1000000.0
   sigmaZ2 = 1000000.0
   paramA = 240.0
   paramW = 276.51006711409394
   paramW2 = 322.1476510067114
   paramZ = 3.4563758389261746E8
   paramZ2 = 4.0268456375838923E8
   paramRx = 0.0
   paramRy = 0.0
  controller = {SimulationController@2763} 
   imageController = {ImageController@2769} 
   simulationView = {SimulationView@2770} 
   simulationModel = {SimulationModel@2633} 
   expSettingsModel = {ExpSettingsModel@2637} 
  is2D = true
  isDomain = false
  isMesh = false
  blinkFlag = false
  seed = 1
  numParticles = 1000
  CPS = 10000
  tauBleach = 0.0
  pixelNum = 21
  extFactor = 1.5
  numFrames = 50000
  frameTime = 0.001
  stepsPerFrame = 10
  D1 = 1.0E-12
  doutDinRatio = 1.0
  D2 = 1.0E-13
  D3 = 1.0E-14
  F2 = 0.0
  F3 = 0.0
  kon = 300.0
  koff = 700.0
  cameraOffset = 100
  cameraNoiseFactor = 3.0
  bleachRadius = 3.0E-6
  bleachFrame = 10000000
  domainRadius = 30.0
  domainDensity = 30.0
  pin = 1.0
  pout = 0.6
  meshWorkSize = 100.0
  hopProbability = 1.0
  progress = 0
  state = {SwingWorker$StateValue@2764} "STARTED"
  future = {SwingWorker$2@2765} 
  propertyChangeSupport = {SwingWorker$SwingWorkerPropertyChangeSupport@2766} 
  doProcess = null
  doNotifyProgressChange = null
  doSubmit = {SwingWorker$DoSubmitAccumulativeRunnable@2767} 
 settingsModel = {ExpSettingsModel@2637} 
 random = {RandomCustom@2632} 
 tStep = 1.0E-4
 darkF = 0.7
 pixelSize = 2.4000000000000003E-7
 PSFSize = 1.3825503355704699E-7
 midPos = 2.5200000000000004E-6
 sizeLowerLimit = -7.560000000000001E-6
 sizeUpperLimit = 7.560000000000001E-6
 bleachFactor = -1.0
 bleachFlag = false
 particleGroup1 = 1000
 particleGroup2 = 0
 blinkOnFactor = 0.9704455335485082
 blinkOffFactor = 0.9323938199059483
 domains = null
 particles = {Particle2D[1000]@2638} 
 impSim = {ImagePlus@2639} "img["2D Simulation" (-2), 16-bit, 21x21x1x50000x1]"
 width = 21
 height = 21
ipSim = {ShortProcessor@2629} "ip[width=21, height=21, bits=16, min=0.0, max=65535.0]"
particle = {Particle2D@2630} 
 x = -1.5523004382971206E-6
 y = -2.323074589977308E-6
 bleached = false
 on = true
 domainIndex = -1
 diffusionCoefficient = 1.0E-12
numPhotons = 2
i = 0
photonX = -1.4620406917571202E-6
photonY = -2.2238906310264783E-6
xPixel = 4
yPixel = 1
currentValue = 100.0

[Marc-Demoustier]

In the end, the issue was a small inattention mistake... I used a normal random generator instead of a gaussian for photon number generation

[Marc-Demoustier]

Everything in the simulation works, just need to implement the Batch button now, it shouldn't be a big challenge. However, there are still things to think about before merging this branch:

• [ ] The poisson generator is kinda slow, it may be worth it to spend some time trying to optimize it.
• [ ] Right now if we exit a mesh and we are not supposed to, we generate randomly new motions until the condition is valid. This follows the previous implementation but it is not very optimized. Instead we can make it bounce on the edge but will it still be valid? As every motion follow a specific range, if there is a bounce and the motion is not in the range then, will it make it harder to identify the diffusion coefficient.
• [ ] Pretty much the same issue if the particle try to exit a domain and is not supposed to. Should it bounce?