dwenv

Launch applications with environment variables from multiple config files.

Example:

python dwenv/launcher.py --config maya_np.env --executable maya

• maya_np.env content:

  np_configs_dir/prod.envc
  np_configs_dir/maya.envc
  np_configs_dir/animschoolpicker.envc
  ...
  

• prod.envc content:

  PROD_NAME = Neti Pikola
  PROD_CODE = NP
  FRAMERATE = 25
  PROD_ROOT = /path/to/prod_root
  SOFTWARE_ROOT = ${PROD_ROOT}/software
  PYTHONPATH > ${SOFTWARE_ROOT}/prodlib
  ...
  

• maya.envc content:

  PATH > ${MAYA_LOCATION}/bin
  MAYA_VP2_DEVICE_OVERRIDE = VirtualDeviceGL
  ...
  

• animschoolpicker.envc content:

  MAYA_PLUG_IN_PATH.windows > ${SOFTWARE_ROOT}/animschoolpicker/windows/2018
  MAYA_PLUG_IN_PATH.linux > ${SOFTWARE_ROOT}/animschoolpicker/linux/2018
  ...
  

Commandline arguments:

--config (-c)       Environment config path
--executable (-x)   Executable path
--arguments (-a)    Executable arguments. Use quotes: -a="..."
--dry-run (-d)      Prints resulting environment without running anything

Configs syntax:

Each .envc line look like this:
ENV_NAME(.platform) [operator] value

• There is three operators for .envc files. Spaces around them are mandatory:

  • " = " will set or replace a variable. Replacing an existing variable will print a warning.
  • " > " appends string/path to existing variable with os path separator (:/;)
  • " < " prepends ...

• If a variable must be set for a specific operating system, add [.platform] to variable name:

  • "VAR.windows = x"
  • "VAR.linux = y"
  • "VAR.darwin = z"
  • ...

• All environment expressions will be expanded if they exists. Thus, the order of the environment configs matter.
  There is no other dependency trickery. Just list what you need.

• You can comment a line with "#".

As a python module:

import dwenv.env
env = dwenv.env.build_env(
    configs_paths=None,
    start_env='current',
    vars_to_remove=None,
    initial_vars=None,
    target_platform=None,
    start_env_backup_path=None,
    set_current_env=False,
    verbose=False)

Using launcher.py or build_env function will also set DWENV_CONFIG var to know which config was used to generate the environment. You can then later re-use this path together with the target_platform argument if you need to send a job on another platform.