renishawWiRE Renishaw Raman spectroscopy parser in python

A python wrapper for read-only accessing the wdf Raman spectroscopy file format created by the WiRE software of Ranishaw Inc. Renishaw Inc owns copyright of the wdf file format.

Ideas for reverse-engineering the WDF format is inspired by:

• Renishaw File Reader by Alex Henderson (DOI:10.5281/zenodo.495477))
• Renishaw IO module in Gwyddion

Installation

Requirements:

• python>=3.6

• Numpy>=1.12.0

• Matplotlib>=2.1.0 (optional, if you want to plot the spectra in the examples)

• Pillow>=3.4.0 (optional, if you want to extract the white light image)

  Warning there are some API changes to Pillow after version 7.2.0 (thanks @markotoplak). Please update renishawWiRE to version >= 0.1.8 if you are experiencing problems with Tiff image extraction.

You can choose any of the following installation methods:

1. Versions hosted on PyPI: via pip

# Optionally on a virtualenv
# Add --user if you don't want to install as sys admin
pip install --upgrade renishawWiRE

If you need full plotting / image extraction support, consider specifying the extras to pip.

# Optionally on a virtualenv
pip install --upgrade "renishawWiRE[plot]"

2. HEAD version: via git + pip

To install the package without examples, run the following commands (installing extra matplotlib and Pillow if not present):

git clone https://github.com/alchem0x2A/py-wdf-reader.git
cd py-wdf-reader
pip install -e ".[plot]"

3. Via conda-forge

The version hosted on conda-forge will install all dependent packages including numpy matplotlib and Pillow. Note the slight difference of package name compared with PyPI.

conda install -c conda-forge renishaw_wire

Test installation

Simply try if importing works:

python -c "import renishawWiRE"

and test the executable binary:

wdf-export -h

Additionally, if you want to test the examples, download them from the binary release and overwrite the dummy files within examples/spectra_files/:

wget https://github.com/alchem0x2A/py-wdf-reader/releases/download/binary/spectra_files.zip 
unzip -o spectra_files.zip -d examples/ 
rm spectra_files.zip
# To avoid unexpected pushing to repo due to large file size
git update-index --skip-worktree examples/spectra_files.wdf

Basic Usage

Check the sample codes in examples/ folder for more details about what the package can do.

1. Simple script for exporting wdf spectra (version > 0.1.11)

The package also installs a simple wrapper script wdf-export for exporting spectra in the wdf file to plain-text formats. If the measurement contains mapping information, the optical image is also exported (as .svg file)

Usage

Simply use the base name of input file for the exported text file:

wdf-export path/to/wdf_file

or specifying the input and output files

wdf-export path/to/wdf_file -o path/to/output.csv

2. Python API

Get file information

renishawWiRE.WDFReader is the main entry point to get information of a WDF file.

# The following example shows how to get the info from a WDF file
# Check `examples/ex1_getinfo.py`
from renishawWiRE import WDFReader

#`filename` can be string, file obj or `pathlib.Path`
filename = "path/to/your/file.wdf"
reader = WDFReader(filename)
reader.print_info()

Get single point spectrum / spectra

When the spectrum is single-point (WDFReader.measurement_type == 1), WDFReader.xdata is the spectral points, and WDFReader.spectra is the accumulated spectrum.

# Example to read and plot single point spectrum
# Assume same file as in previous section
# Check `examples/ex2_sp_spectra.py`
import matplotlib.pyplot as plt
wavenumber = reader.xdata
spectra = reader.spectra
plt.plot(wavenumber, spectra)

An example is shown below:

Get depth series spectra

A depth series measures contains single point spectra with varied Z-depth. For this type WDFReader.measurement_type == 2. The code to get the spectra are the same as the one in the single point spectra measurement, instead that the WDFReade.spectra becomes a matrix with size of (count, point_per_spectrum). The WDFReader.zpos returns the values of z-scan points.

For details of Z-depth data processing, check this example

Get line scan from StreamLine™ / StreamHR Line™ measurements

For mapped measurements (line or grid scan), WDFReader.measurement_type == 3. The code to get the spectra are the same as the one in the single point spectra measurement, instead that the WDFReade.spectra becomes a matrix with size of (count, point_per_spectrum):

# Example to read line scane spectrum
# Check `examples/ex3_linscan.py`
filename = "path/to/line-scan.wdf"
reader = WDFReader(filename)
wn = reader.xdata
spectra = reader.spectra
print(wn.shape, spectra.shape)

An example of the line scane is shown below:

It is also possible to correlate the xy-coordinates with the spectra. For a mapping measurement, WDFReader.xpos and WDFReader.ypos will contain the point-wise x and y coordinates.

# Check examples/ex4_linxy.py for details
x = reader.xpos
y = reader.ypos
# Cartesian distance
d = (x ** 2 + y ** 2) ** (1 / 2)

Get grid mapping from StreamLine™ / StreamHR Line™ measurements

Finally let's extract the grid-spaced Raman data. For mapping data with spectra_w pixels in the x-direction and spectra_h in the y-direction, the matrix of spectra is shaped into (spectra_h, spectra_w, points_per_spectrum).

Make sure your xy-coordinates starts from the top left corner.

# For gridded data, x and y are on rectangle grids
# check examples/ex5_mapping.py for details
x = reader.xpos
y = reader.ypos
spectra = reader.spectra
# Use other packages to handle spectra
# write yourself the function or use a 3rd-party libray
mapped_data = some_treating_function(spectra, **params)
# plot mapped data using plt.imshow
plt.pcolor(mapped_data, extends=[0, x.max() - x.min(),
                                 y.max() - y.min(), 0])

An example of mapping data is shown below:

You can also work on the white-light image which automatically saved during a mapped scan. The jpeg-form image can be obtained by WDFReader.img as an io object, and some further informations about the dimensions etc. For this to work you need Pillow installed as third-party library:

• Get coordinates of white-light image

# There are two-sets of coordinates.
# `xpos` and `ypos` are the Stage XY-coordinate of the mapped area
# while `img_origins` and `img_dimensions` are size (μm) of white-light image
# See examples/ex6_mapping_img.py for details
import matplotlib.image as mpimg
img_x0, img_y0 = reader.img_origins
img_w, img_h = reader.img_dimensions
plt.imshow(reader.img, 
           extent=(img_x0, img_x0 + img_w,
                   img_y0 + img_h, img_y0))

An example of mapped area on white light image is shown below:

• Overlaying white-light image with mapped spectra

# `img_cropbox` is the pixel positions for cropping
# Requires PIL to operate
# See examples/ex7_overlay_mapping.py for details
img = PIL.Image.open(reader.img)
img1 = img.crop(box=reader.img_cropbox)
extent = ... # Same extent for both images
plt.imshow(img1, alpha=0.5, extent=extent) # White light image 
plt.imshow(spectra, alpha=0.5, extent=extent) # Mapped spectra

The following example shows the overlayed image of both fields. Some degree of misalignment can be observed.

TODOs

There are still several functionalities not implemented:

• Extract image info
• Verify image coordinate superposition
• Improve series measurement retrieval
• Testing on various version of Renishaw instruments
• Binary utilities

Bug reports

The codes are only tested on the Raman spectra files that generated from my personal measurements. If you encounter any peculiar behavior of the package please kindly open an issue with your report / suggestions. Thx!