Troubleshooting

Troubleshooting

This page collects some possible errors you may encounter and trick how to fix them. If you have some questions about how to use this code, you are welcome to discuss with us.

If you have additional tips, please either report an issue or submit a PR with suggestions.

Installation problems

Cannot find the `julia` executable

Make sure you have Julia installed in your environment. Please download the latest stable version for your platform. If you are using a *nix system, the recommended way is to use Juliaup. If you do not want to install Juliaup or you are using other platforms that Julia supports, download the corresponding binaries. Then, create a symbolic link to the Julia executable. If the path is not in your $PATH environment variable, export it to your $PATH.

Some clusters, like Habanero, Comet, or Expanse, already have Julia installed as a module, you may just module load julia to use it. If not, either install by yourself or contact your administrator.

Have trouble installing `qha`

If you are seeing error message like this:

ERROR: LoadError: InitError: PyError (PyImport_ImportModule

The Python package qha could not be imported by pyimport. Usually this means
that you did not install qha in the Python version being used by PyCall.

PyCall is currently configured to use the Python version at:

/usr/bin/python3

and you should use whatever mechanism you usually use (apt-get, pip, conda,
etcetera) to install the Python package containing the qha module.

It is because on some operating systems, python is already installed, and Julia selects it as the default binary. But that python cannot install third-party Python packages. So qha cannot be automatically installed by Julia.

One solution is to re-configure PyCall to use a different Python version on your system: set ENV["PYTHON"] to the path of the python executable you want to use, run Pkg.build("PyCall"), and re-launch Julia. For example, in the Julia REPL, run

julia> using Pkg

julia> ENV["PYTHON"] = "... path of the python executable ..."
# ENV["PYTHON"] = raw"C:\Python37-x64\python.exe" # example for Windows, "raw" to not have to escape: "C:\\Python37-x64\\python.exe"
# ENV["PYTHON"] = "/usr/bin/python3.7"            # example for *nix

julia> Pkg.add("PyCall");

julia> Pkg.build("PyCall");

Please see this part for more detailed instructions.

Another solution is to configure PyCall to use a Julia-specific Python distribution via the Conda.jl package (which installs a private Anaconda Python distribution), which has the advantage that packages can be installed and kept up-to-date via Julia. As explained in the PyCall documentation, in Julia, run

julia> using Pkg
julia> ENV["PYTHON"] = ""  # empty string""
julia> Pkg.add("PyCall")   Resolving package versions...
    Updating `~/work/Express.jl/Express.jl/docs/Project.toml`
  [438e738f] + PyCall v1.96.1
  No Changes to `~/work/Express.jl/Express.jl/docs/Manifest.toml`
julia> Pkg.build("PyCall");    Building Conda ─→ `~/.julia/scratchspaces/44cfe95a-1eb2-52ea-b672-e2afdf69b78f/915ebe6f0e7302693bdd8eac985797dba1d25662/build.log`
    Building PyCall → `~/.julia/scratchspaces/44cfe95a-1eb2-52ea-b672-e2afdf69b78f/43d304ac6f0354755f1d60730ece8c499980f7ba/build.log`

Then re-launch Julia.

If qha still cannot be installed, go to the Python binary directory you specified (for the second solution, go to $JULIA_DEPOT_PATH/conda/3/bin) and run the following command

./pip install qha

Note

At least Python 3.6 and above is required to install qha. Please read its manual for more information.

Loading settings

Error parsing `YAML` files

If you encounter

ERROR: expected '<document start>' but found YAML.BlockMappingStartToken at nothing

ERROR: while scanning a simple key at line n, column 0: could not find expected ':' at line n+1, column 0

Check whether you have no space between the YAML key and its value like key:1 or key:some text, etc. To correct, change to key: 1, key: some text, etc. Otherwise, check other YAML syntax you may have broken.

Loading Express

Julia compiles/loads slow

First, we recommend you download the latest version of Julia. Usually, the newest version has the best performance.

If you just want Julia to do a simple task and only once, you could start the Julia REPL with

julia --compile=min

to minimize compilation or

julia --optimize=0

to minimize optimizations, or just use both. Or you could make a system image and run with

julia --sysimage custom-image.so

See Fredrik Ekre's talk for details.

Running workflows

How to solve the error `come across ProcessFailedException`?

The reasons for this error are manifold. One common reason is that express did not find the executables of Quantum ESPRESSO or mpiexec (e.g., their parent directory is not in the PATH environment variable) or the template input files. If so, please specify the paths to these executables in the absolute path format:

cli:
  mpi:
    np: 64
    path: /usr/local/bin/gcc-10.2.0/openmpi/bin/mpiexec
  pw:
    path: /usr/local/bin/gcc-9.2.0/qe-6.5/bin/pw.x
  ph:
    path: /usr/local/bin/gcc-9.2.0/qe-6.5/bin/ph.x
  q2r:
    path: /usr/local/bin/gcc-9.2.0/qe-6.5/bin/q2r.x
  matdyn:
    path: /usr/local/bin/gcc-9.2.0/qe-6.5/bin/matdyn.x
template:
  scf: /home/qe/template.in
  dfpt: /home/qe/ph.in
  q2r: /home/qe/q2r.in
  disp: /home/qe/phdos.in

Besides, the paths to the directories in the template files could also be absolute to avoid unexpected errors:

&CONTROL
  verbosity = 'high'
  tstress = .true.
  tprnfor = .true.
  outdir = '/home/qe/temp'
  prefix = 'Ge'
  disk_io = 'high'
  pseudo_dir = '/home/qe/pseudo'
/

How to solve `EOFError: read end of file`?

If it happens when you run

wf = buildworkflow("workflow.yaml")
run!(wf)

Check the value of the key status in the save dictionary in workflow.yaml:

save:
  status: status.jls

Then delete the file status.jls and retry running run!(wf).

Why does `buildworkflow` return the same `Workflow` object if I am starting a new workflow?

This happens when running a workflow (say, a phonon workflow, for example) followed by a workflow (an equation of state workflow in this case). And after the following Julia commands:

using Express.PhononWorkflow.Recipes
using QuantumESPRESSOExpress

wf = buildworkflow("phonon.yaml")

you will find that wf still contains the jobs from the previous workflow, and they are probably all succeeded (because that is when you want to run the next workflow). The cause is likely in your configuration file (phonon.yaml). You probably have the following setting:

save:
  status: status.jls

where the status.jls already exists and is the serialized Workflow object saved from the previous (equation of state) workflow. Express.jl will first load this file if it exists. To fix this, change the value of the status key to another path.

How to solve the `LoadError: Only one AND undecorated format string is allowed`?

The probable reason for the error is that you changed the values of files.dirs.pattern, files.pattern.input, and files.pattern.output in the configuration file (their default values are "p=%.1f", "%s.in", and "%s.out") to some strings with more than one C-style format specifiers (such as "p=%d%d").

Miscellaneous errors

If you see the following error message

julia> w = buildworkflow("eos.toml");
ERROR: MethodError: Cannot `convert` an object of type Dict{String, Any} to an object of type AbInitioSoftwareBase.Commands.CommandConfig
Closest candidates are:
  convert(::Type{T}, ::Intervals.AnchoredInterval{P, T}) where {P, T} at ~/.julia/packages/Intervals/ua9cq/src/anchoredinterval.jl:181
  convert(::Type{T}, ::Intervals.Interval{T}) where T at ~/.julia/packages/Intervals/ua9cq/src/interval.jl:253
  convert(::Type{T}, ::P) where {T, P<:(Polynomials.AbstractPolynomial{T})} at ~/.julia/packages/Polynomials/WvTSC/src/common.jl:434
  ...

Then you probably forget loading a plugin package for Express.jl. For example, you should run

julia> using QuantumESPRESSOExpress

to fix this error.