OCTAVVS is a set of graphical tools for high-throughput preprocessing and analysis of vibrational spectroscopy data. Currently, the preprocessing is primarily geared towards images from infrared absorption spectroscopy with focal plane array detectors.
There are two/three separate tools in the current version:
preprocessing deals with atmospheric correction, resonant Mie scattering correction, baseline correction and normalization.
decomposition decomposes observed spectra into nonnegative concentrations and spectra using the MCR-ALS algorithm, performs K-means clustering on the MCR-ALS concentrations (or on the input data), and allows the user to manually annotate clusters and export the annotated data.
The decomposition tool supersedes two earlier tools:
mcr_als decomposes observed spectra into nonnegative concentrations and spectra using the MCR-ALS algorithm.
clustering performs K-means clustering on the concentrations inferred by MCR-ALS.
OCCTAVS needs a working Python 3 environment with various packages. The easiest way to get this is through the Conda package management system.
Download and install the Python 3.7 (or later) version of Miniconda. There are some platform-specific differences during installation:
-
On Windows: The default options are good (no need to have Conda modify your path).
After installation, start a Conda console (found in the Start menu). -
On Mac OS X: The default options are good. (But if you choose to not let Conda modify your path, you can specify the full path to the commands in the following steps.)
After installation, start Terminal. -
On Linux: Make the Miniconda .sh package executable and run it. When it asks about running conda init, it wants to modify your $PATH in .bashrc. This can be fine, but on some Linux distributions (notably OpenSUSE) it breaks KDE when logging in. A suggested workaround is to change $PATH manually when needed. An alias in .bashrc can be convenient:
alias startconda='export PATH=~/miniconda3/bin:"$PATH"'
From the console, install OCTAVVS and its dependencies: conda install -c ctroein octavvs
The easiest way to access the OCTAVVS tools is through desktop shortcuts
which are created by running the oct_make_icons script from the command prompt.
(We are not entirely sure if this works on all Mac OS X versions)
Regardless of whether you created icons, the scripts
oct_preprocessing and oct_decomposition
can be run straight from the console (Conda console or terminal window, as described above).
The location of the OCTAVVS scripts will depend on where you installed Conda.
Within the Conda directory, the three scripts will be in bin whereas
the actual Python code will be located in lib/python3.7/site-packages/octavvs.
Test data from two 64x64 images of Paxillus hyphae growing on lignin can be downloaded here (zip archive, 47 MB).
Information about the most recent release of OCTAVVS can be found on its PyPI page, as well as on its Anaconda Cloud page.
To upgrade to the most recent version, do conda update octavvs in the Conda console / Terminal.
Information about released versions can be found here.
The main project homepage is its GitHub page, where developers can access the OCTAVVS code and submit bug reports and patches etc.
Questions, bug reports and other feedback may be sent to corresponding author Carl Troein carl@thep.lu.se.
Users familiar with Python could also install OCTAVVS through pip as an alternative to Anaconda, as new releases will be made to PyPI in parallel with releases to Anaconda Cloud. Note that pyqt and opencv sometimes don't work when installed through pip, depending on your system etc.