-
Notifications
You must be signed in to change notification settings - Fork 72
CellProfiler Analyst Developer's version installation for Mac
brew install pkg-config
brew install mysql
Install Java - OpenJDK
Find out what version of java is currently being shipped by homebrew
$ brew update
$ brew info javaYou can get the latest version if prokaryote supports it with
$ brew install javaOtherwise, Java11 is the most up-to-date Java LTS (LongTermSupport) release and can be installed with
$ brew install java11(For installing other versions of Java with homebrew see AdoptOpenJDK )
If homebrew tells you to softlink and/or add Java to your path at the end of installation, be sure to do so. Otherwise, any versions you install will not be visible in the next step.
Get the exact name of your available Java versions
$ ls /Library/Java/JavaVirtualMachines/Set your JAVA_HOME environment variable by adding it to your .zshrc or .bashrc (depending on the Shell used).
- For example, if your Java installation is located at
/Library/Java/JavaVirtualMachines/openjdk-14.0.2.jdk/Contents/Homethen add the following to your.zshrc:export JAVA_HOME="/Library/Java/JavaVirtualMachines/openjdk-14.0.2.jdk/Contents/Home" - Don't forget to source your .zshrc file
$ source ~/.zshrcconda create --name cpa python=3.8
conda activate cpa
If/when using conda, you will need a "framework build" of python. The python.app package is such a build. Once installed, to use it, type pythonw wherever you would normally type python.
conda install python.app
If you would rather not use conda, you should make sure that you are using python 3.8.
You can install with homebrew:
brew install [email protected]
To use it, you will need type python3.8 <command> in the terminal. To use the specific version of pip associated with this python version, you will need to type pip3.8 <command> in your terminal.
Run git clone from whatever directory you want the CellProfiler-Analyst source code to live.
git clone https://github.com/CellProfiler/CellProfiler-Analyst.git
cd CellProfiler-Analyst
python -m pip install "numpy~=1.24.4"
python -m pip install -e .
python -m CellProfiler-Analyst
If you used conda, remember to use python.app, the framework build:
pythonw -m CellProfiler-Analyst
Note: We also offer a [vagrant solution] (https://github.com/CellProfiler/metabolism).
CPA developer's version used to require CP to be installed, but it should be independent with the javabridge, etc. now factored out (afaik). Nevertheless, these notes are using my Mac (10.9) which already had CP developer's version working (May 2014).
pip install javabridge
pip install python-bioformats
pip install verlib
pip install mock (not sure if this is strictly necessary, but it was also listed in requirements.txt here so I pip-installed it)
Run CPA with (still from within virtualenv)
python ./<YOUR CPA GITHUB DIR>/CellProfiler-Analyst.py
To see what Java versions are installed on your system, use the command below; as of CPA 2.2.1 Java8 (or OpenJDK8) is required.
/usr/libexec/java_home -V
Matching Java Virtual Machines (3):
11.0.2, x86_64: "OpenJDK 11.0.2" /Library/Java/JavaVirtualMachines/openjdk-11.0.2.jdk/Contents/Home
1.8.0_192, x86_64: "Java SE 8" /Library/Java/JavaVirtualMachines/jdk1.8.0_192.jdk/Contents/Home
1.8.0_51, x86_64: "Java SE 8" /Library/Java/JavaVirtualMachines/jdk1.8.0_51.jdk/Contents/Home
To set JAVA_HOME on your system to a Java8/JDK8 installation, get the version from the Java8 install on your machine (ie on mine, I will use 1.8.0_192) - note that the second step is copying the line into a file, at which point you can save and close it (with Ctl+O then Ctl+X)
$ nano ~/.bash_profile
export JAVA_HOME=$(/usr/libexec/java_home -v 1.8.0_192)
$ source ~/.bash_profile
$ echo $JAVA_HOME
After installation, you might want to check the page [https://github.com/CellProfiler/CellProfiler-Analyst/wiki/CellProfiler-Analyst-Developer-Tweaks](CellProfiler Analyst Developer Tweaks) for relevant optimizations and performance tweaks.