CrYOLO/1.6.1 modules

There are _cpu and _gpu flags available, which are selected based on
fancy modulefile code (Thanks Achim)

EM/crYOLO/files/variants

@@ -1,3 +1,5 @@
 crYOLO/1.2.3	unstable	cuda/9.0.176
 crYOLO/1.5.6_gpu	unstable	anaconda/2019.07
+crYOLO/1.6.1_gpu	unstable	anaconda/2019.07
+crYOLO/1.6.1_cpu	unstable	anaconda/2019.07
 

EM/crYOLO/modulefile

@@ -26,19 +26,28 @@ convolutional neural networks and utilizes the popular You Only Look Once
 set shelltype [module-info shelltype]
 switch -- $shelltype {
     "sh" {
+        puts stdout "source \"\${ANACONDA_PREFIX}\"/admintools/conda_setup.sh;\n"
     }
     default {
         puts stderr "Shells of type '$shelltype' are NOT supported!"
     }
 }
 
-#set useflag [string map {{:} {}} $USEFLAGS]
-#puts stderr "conda activate cryolo-${V_PKG}-${USEFLAGS%\n"
+# Determine useflags and construct conda env name
+set list [split $V "_"]
+set version  [lindex ${list} 0]
+set useflags [lrange ${list} 1 end]
+set flavour  [lindex $useflags 0]
+set env_to_activate [string tolower ${P}]-${version}-${flavour}
+
+#puts stderr "P/V: $P/$V"
+#puts stderr "USEFLAG: [lindex $useflags 0]"
+#puts stderr "$env_to_activate"
 
 switch [module-info mode] {
     "load" {
       # Relies on the anaconda module being loaded since we need to update the path
-      puts stdout "conda activate cryolo-1.5.6-gpu;\n"
+      puts stdout "conda activate $env_to_activate;\n"
     }
     "unload" -
     "remove" {

Programming/anaconda/2019.07/conda-env-defs/cryolo/README.md

@@ -2,7 +2,8 @@
 
 IMPORTANT! DON'T INSTALL WITH AURISTOR. Use an OpenAFS system (e.g. pmod6)
 
-    export TMPDIR=/opt/tmp/bliven_s
+    export TMPDIR=/opt/tmp/$USER
+    export XDG_CACHE_HOME=/opt/tmp/$USER/.cache
     CONDA_COPY_ALWAYS=1 conda env create --file cryolo.yml
     conda activate cryolo
 

Programming/anaconda/2019.07/conda-env-defs/cryolo/cryolo-cpu.yml

@@ -1,4 +1,4 @@
-name: cryolo-1.5.6-gpu
+name: cryolo-1.6.1-cpu
 channels:
   - defaults
 dependencies:
@@ -11,4 +11,4 @@ dependencies:
   - intel-openmp==2019.4
   - pip
   - pip:
-    - cryolo[gpu]==1.5.6
+    - cryolo[cpu]==1.6.1

Programming/anaconda/2019.07/conda-env-defs/cryolo/cryolo-gpu.yml

@@ -0,0 +1,14 @@
+name: cryolo-1.6.1-gpu
+channels:
+  - defaults
+dependencies:
+  - python=3.6
+  - pyqt=5
+  - cudnn=7.1.2
+  - numpy==1.14.5
+  - cython
+  - wxPython==4.0.4
+  - intel-openmp==2019.4
+  - pip
+  - pip:
+    - cryolo[gpu]==1.6.1

Programming/anaconda/files/admintools/conda_setup.sh

@@ -0,0 +1,14 @@
+# simulate `conda init` type hook
+# Expects CONDA_EXE and CONDA_PREFIX to be set
+__conda_setup="$("${CONDA_EXE:-conda}" 'shell.bash' 'hook' 2> /dev/null)"
+if [ $? -eq 0 ]; then
+    eval "$__conda_setup"
+else
+    if [ -f "${CONDA_PREFIX:?CONDA_PREFIX not set}/etc/profile.d/conda.sh" ]; then
+        . "${CONDA_PREFIX}/etc/profile.d/conda.sh"
+    else
+        export PATH="${CONDA_PREFIX}/bin:$PATH"
+    fi
+fi
+unset __conda_setup
+

Programming/anaconda/files/admintools/kernelwrapper.sh

@@ -0,0 +1,27 @@
+#!/bin/bash
+dbgecho() {
+    DBGLOG=/tmp/kernelwrapper-${USER}.log
+    if [[ $DEBUG_JUPYTER_WRAPPER == 1 ]]; then
+	echo "$1" >> $DBGLOG
+    fi
+}
+
+DATE=$(date +%Y-%m-%dT%H:%M)
+dbgecho "$DATE: $@" 
+
+# ensure that module commands are available
+dbgecho "sourcing profile.bash"
+source /opt/psi/config/profile.bash
+
+dbgecho "sourcing conda.sh"
+source /opt/psi/Programming/anaconda/2019.07/conda/etc/profile.d/conda.sh &>> $DBGLOG
+
+dbgecho "Running activate"
+conda activate $1
+
+# undefine module command since it causes error in the notebook when using
+# shell escapes
+unset -f module
+shift
+exec "$@"
+

Programming/anaconda/modulefile

@@ -57,7 +57,8 @@ switch [module-info mode] {
         # but we need to prevent conflicts with a conda from a different source.
 
         # puts stderr "DEBUG: Using conda from $P/$V\n"
-        puts stdout "source \"$PREFIX/conda/etc/profile.d/conda.sh\";\n"
+        #puts stdout "source \"$PREFIX/conda/etc/profile.d/conda.sh\";\n"
+        puts stdout "source $PREFIX/admintools/conda_setup.sh;\n"
 
         # Activate base? No: anaconda should just provide the conda tool
         #puts stdout "conda activate;\n"