logproc changes

.gitignore

@@ -10,3 +10,4 @@ fly2p/sample/*.*
 fly2p/sample/img/*.*
 *.pdf
 *.pyc
+*.tif

README_pypi.md

@@ -0,0 +1,30 @@
+<img src="https://user-images.githubusercontent.com/10639803/126242184-65bc84c4-3fa2-4034-be51-3e8c2e4d9f8c.png" align="middle" width="3000"/>
+
+# fly2p
+
+Tools for analyzing imaging data collected with [Vidrio Scanimage software](https://vidriotechnologies.com/scanimage/) or [micromanger](https://micro-manager.org/). Loading ScanImage data relies on [scanimageReader](https://pypi.org/project/scanimage-tiff-reader/), which can be installed via 'pip install scanimage-tiff-reader'. Other dependencies are tracked using poetry.
+
+### Organization
+The fly2p package contains the following submodules:
+* **preproc**: Some file-format specific functions that extract metadata and load the imaging data. imgPreproc.py defines a data object to hold metadata and imaging data as well as basic proporcessing functions.
+* **viz**: A collection of utility functions related to plotting flourescence traces and images.
+
+In addition, the **scripts** folder contains notebooks that illustrate how to use functions in this module based on example files in **sample** (sample files are not currently pushed to repo).
+
+### Installation
+I recommend using poetry to setup a custom conda environment. A helpful introduction can be found [here](https://ealizadeh.com/blog/guide-to-python-env-pkg-dependency-using-conda-poetry).
+
+0. Clone repo, navigate into folder
+1. If you don't already have poetry, [install poetry](https://python-poetry.org/docs/#installation). You may need to close command window and open a new one.
+2. Create conda environment:  
+ `conda create --name fly2p python=3.8`
+4. Activate environment:  
+ `conda activate fly2p`
+6. Make sure you are in the top folder of the cloned repo, then install dependencies:  
+ `poetry install`
+8. Setup the new environment as an ipython kernel:  
+    `conda install -c anaconda ipykernel`  
+    then  
+    `python -m ipykernel install --user --name=fly2p`
+
+Now you should be able to run the example notebooks in the **scripts** folder without problems.

fly2p/preproc/imgPreproc.py

@@ -26,9 +26,10 @@ class imagingTimeseries:
     refImage: xr.DataArray # image used for motion correction (MC)
     dffStack: xr.DataArray # image or stack, maximum intensity projection of DFF over time after MC
     F0stack: xr.DataArray # image or stack of F0 (baseline flourescences)
+    rawStack: xr.DataArray # raw image stack
 
     # roi data
-    roitype: str #polygons ("poly") or correlation-based ("corr")?
+    roitype: str #polygons ("poly") or correlation-based ("corr") or manual?
     roiMask: np.ndarray
     roiDFF: pd.DataFrame
 
@@ -48,6 +49,7 @@ def saveData(self, saveDir, saveName):
         self.refImage.to_netcdf(sep.join([savepath,'refImg.nc']), mode='w')
         self.dffStack.to_netcdf(sep.join([savepath,'dffStack.nc']), mode='w')
         self.F0stack.to_netcdf(sep.join([savepath,'F0stack.nc']), mode='w')
+        self.rawStack.to_netcdf(sep.join([savepath,'rawStack.nc']), mode='w')
 
         # save roi data
         np.save(sep.join([savepath,'roiMask']),self.roiMask)
@@ -126,7 +128,7 @@ def loadImagingTimeseries(path2imgdat):
     return imgTS
 
 ## CONVERT TO XARRAY
-def stack2xarray(stack, basicMetadat, data4D = True):
+def stack2xarray(stack, basicMetadat, data4D = True, clipping = False):
     volcoords = [i/basicMetadat['scanVolumeRate'] for i in range(stack.shape[0])]
     if data4D:
         slices = [i*basicMetadat['stackZStepSize'] for i in range(stack.shape[1])]
@@ -141,9 +143,15 @@ def stack2xarray(stack, basicMetadat, data4D = True):
         imgStack = xr.DataArray(stack, coords = [volcoords, xpx, ypx],
                             dims = ['volumes [s]', 'xpix [µm]', 'ypix [µm]'])
 
-    minval = np.min(imgStack)
-    if minval < 0: imgStack = imgStack - minval
-
+    # forcing all values to be above 0 by subtracting the minimum value is prone to biases from single pixel noise. A more ideal conversion to +ve values would clip negative values to 0
+    # see: https://docs.scanimage.org/Windows+Reference+Guide/Channels.html
+    if (clipping==False):
+        minval = np.min(imgStack) 
+        if minval < 0: imgStack = imgStack - minval
+    elif (clipping==True):
+        imgStack = xr.where(imgStack<0,0,imgStack)
+    else:
+        print('no +ve conversion applied when converting to xarray')
     return imgStack
 
 ## CONVERT TO XARRAY when no time dimension

pyproject.toml

@@ -1,7 +1,8 @@
 [tool.poetry]
 name = "fly2p"
-version = "0.1.0"
+version = "0.2.2"
 description = "Tools for analyzing 2p (calcium) imaging data collected in flies (or other insects)."
+readme = "README_pypi.md"
 authors = ["hjmh <hjmhaberkern@gmail.com>"]
 license = "MIT"
 
@@ -40,4 +41,4 @@ ipykernel = "^6.25.2"
 
 [build-system]
 requires = ["poetry-core>=1.0.0"]
-build-backend = "poetry.core.masonry.api"
+build-backend = "poetry.core.masonry.api"