fixed doc strings to be compliant

.gitlab-ci.yml

@@ -20,7 +20,8 @@ pages:
   - DEBIAN_FRONTEND=noninteractive apt-get install -y build-essential gcc cmake python3-pip python3 python3-dev libeigen3-dev
   - pip3 install -U -r requirements.txt
   - python3 setup.py install
-  - sphinx-apidoc -f -o docs/source public
+  - sphinx-build -b html docs/source public
+
   artifacts:
     paths:
     - public

docs/source/conf.py

@@ -34,7 +34,6 @@ extensions = [
     'sphinx.ext.viewcode',
     'sphinx_rtd_theme',
     'sphinx.ext.napoleon',
-    'numpydoc',
 ]
 
 # Add any paths that contain templates here, relative to this directory.

tools/DEM/DEM.py

@@ -5,8 +5,8 @@ def DEM_step(locMMin, radiiMM, k=0.5):
     """
     Lightweight DEM with assumption of same size spheres as mechanical regularisation
 
-    Parameters:
-    -----------
+    Parameters
+    ----------
         locMMin : Nx3 2D numpy array of floats
             xyz positions of spheres in mm, with the origin being the middle of the detector
 
@@ -17,8 +17,8 @@ def DEM_step(locMMin, radiiMM, k=0.5):
             Stiffness and timestep wrapped into one
             Default = 0.1
 
-    Returns:
-    --------
+    Returns
+    -------
         locMMin : output positions
     """
     locMM = locMMin.copy()

tools/optimisePositions.py

@@ -18,8 +18,8 @@ def optimiseSensitivityFields(radioMM, xyzGuessesMM, radiiMM, iterationsMax=20,
     synthetic radiograph.
     This is done locally in a ROI around each particle, which avoids doing the projection for all pixels.
 
-    Parameters:
-    -----------
+    Parameters
+    ----------
         radioMM : 2D numpy array of floats
             Projection of distance crossed in spheres in MM
 
@@ -85,8 +85,8 @@ def optimiseSensitivityFields(radioMM, xyzGuessesMM, radiiMM, iterationsMax=20,
             Default = None
 
 
-    Returns:
-    --------
+    Returns
+    -------
         xzyMM : Corrected positions array
     """
     import radioSphere.projectSphere
@@ -338,8 +338,8 @@ def _optimiseSensitivityFieldsMultiProj(radioMM, xyzGuessesMM, radiiMM, transfor
     synthetic radiograph.
     This is done locally in a ROI around each particle, which avoids doing the projection for all pixels.
 
-    Parameters:
-    -----------
+    Parameters
+    ----------
         radioMM : 3D numpy array of floats
             N Projections of distance crossed in spheres in MM, where N is the number of different views
 
@@ -397,8 +397,8 @@ def _optimiseSensitivityFieldsMultiProj(radioMM, xyzGuessesMM, radiiMM, transfor
             Perform a soft DEM correction between all iterations?
             Default = False
 
-    Returns:
-    --------
+    Returns
+    -------
         xzyMM : Corrected positions array
     """
     import radioSphere.projectSphere

tools/projectSphere.py

@@ -9,8 +9,8 @@ def pointToDetectorPixelRange(posMM, sourceDetectorDistMM=100, pixelSizeMM=0.1,
     The main idea is that it will be used with `ROIaroundSphere` option for projectSphereMM() and in turn singleSphereToDetectorPixelRange()
     in order to only project the needed pixels.
 
-    Parameters:
-    -----------
+    Parameters
+    ----------
         posMM : 1x3 2D numpy array of floats
             xyz position of sphere in mm, with the origin being the middle of the detector
 
@@ -27,8 +27,8 @@ def pointToDetectorPixelRange(posMM, sourceDetectorDistMM=100, pixelSizeMM=0.1,
             Number of pixels rows, columns of detector
             Default = [512,512]
 
-    Returns:
-    --------
+    Returns
+    -------
         detectorPixel : tuple
             row, column (j,i) coordinate on detector as per figures/projectedCoords_v2.pdf
     """
@@ -58,8 +58,8 @@ def singleSphereToDetectorPixelRange(spherePositionMM, radiusMM, radiusMargin=0.
     The main idea is that it will be used with `ROIaroundSphere` option for projectSphereMM()
     in order to only project the needed pixels.
 
-    Parameters:
-    -----------
+    Parameters
+    ----------
         spherePositionMM : 1x3 2D numpy array of floats
             xyz position of sphere in mm, with the origin being the middle of the detector
 
@@ -95,8 +95,8 @@ def singleSphereToDetectorPixelRange(spherePositionMM, radiusMM, radiusMargin=0.
             XYZ transformation matrix to apply to coordinates
             Default = None
 
-    Returns:
-    --------
+    Returns
+    -------
         JIrange : range in rows, colums of detector concerned by this grain
     """
     assert((transformationCentreMM is None) == (transformationMatrix is None)), "projectSphere.singleSphereToDetectorPixelRange(): transformationCentreMM and transformationMatrix must both be set or unset"
@@ -146,8 +146,8 @@ def projectSphereMM(spheresPositionMM, radiiMM, ROIcentreMM=None, ROIradiusMM=No
     In order to allow projections from diffferent angles, an XYZ centre and a transformation matrix can be provided,
     which will be applied to the particle positions.
 
-    Parameters:
-    -----------
+    Parameters
+    ----------
         spheresPositionMM : Nx3 2D numpy array of floats
             xyz positions of spheres in mm, with the origin being the middle of the detector
 
@@ -191,8 +191,8 @@ def projectSphereMM(spheresPositionMM, radiiMM, ROIcentreMM=None, ROIradiusMM=No
             sigma of blur to pass to scipy.ndimage.gaussian_filter to
             blur the radiograph at the end of everything
 
-    Returns:
-    --------
+    Returns
+    -------
         projectionMM : 2D numpy array of floats
             Radiography containing the total crossed distance through the spheres distance in mm for each beam path.
             To turn this into a radiography, the distances should be put into a calibrated Beer-Lambert law
@@ -313,18 +313,18 @@ def projectSphereMM(spheresPositionMM, radiiMM, ROIcentreMM=None, ROIradiusMM=No
 def computeLinearBackground(radioMM, mask=None):
     """
     This function computes a plane-fit for background greylevels to help with the correction of the background
-    
 
-    Parameters:
-    -----------
+
+    Parameters
+    ----------
         radioMM : 2D numpy array of floats
             The image
 
         mask : 2D numpy array of bools, optional
             Masked zone to fit?
-    
-    Returns:
-    --------
+
+    Returns
+    -------
         background : 2D numpy array of floats
             Same size as radioMM
     """
@@ -348,7 +348,7 @@ def computeLinearBackground(radioMM, mask=None):
 
 def gl2mm(radio, calib=None):
     """
-    This function takes a greylevel radiograph (I/I0) and returns 
+    This function takes a greylevel radiograph (I/I0) and returns
     a radiograph in mm (L), representing the path length encountered.
 
     Parameters
@@ -366,7 +366,7 @@ def gl2mm(radio, calib=None):
 
 def mm2gl(radioMM, calib=None):
     """
-    This function takes a radiograph in mm (L) and returns 
+    This function takes a radiograph in mm (L) and returns
     a radiograph in greylevels (I/I0)
 
     Parameters
@@ -380,4 +380,3 @@ def mm2gl(radioMM, calib=None):
     if calib is not None:
         print("projectSphere.mm2gl() I need to be implemented")
     return numpy.exp(-radioMM)
-