biosimulator-processes 0.0.14

No project description provided

biosimulator-processes

Core implementations of process-bigraph.composite.Process() aligning with BioSimulators simulator tools.

Installation

The easiest way to download this tool is via the Python Package Index. You may download core-processes with:

pip install biosimulator-processes

We recommend using an environment/package manager like Conda to install the dependencies required for your use.

Most of the direct UI content for this tooling will be in the form of a jupyter notebook.

Using biosimulator_processes.smoldyn_process.SmoldynProcess():

Mac Users PLEASE NOTE:

Due to the multi-lingual nature of Smoldyn, which is primarily developed in C++, the installation process for utilizing the SmoldynProcess process implementation requires separate handling. This is particularly relevant for macOS and Windows users, where setting up the Python bindings can be more complex.

For your convienience, we have created an installation shell script that will install the correct distribution of Smoldyn based on your Mac processor along with the codebase of this repo. To install Smoldyn and this repo on your Mac, please adhere to the following instructions:

1. Clone this repo from Github:

    git clone https://github.com/vivarium-collective/biosimulator-processes.git
   

2. Provide adminstrative access to the scripts directory within the cloned repo:

    cd biosimulator-processes 
    chmod +x scripts 
   

3. Look for the install-with-smoldyn-for-mac- shell script where corresponds to your machine's processor:

    ls scripts 
   

4. Run the appropriate shell script (for example, using mac silicon):

    scripts/install-with-smoldyn-for-mac-silicon.sh 
   

Quick Start Example:

Composing, running, and viewing the results of a composite simulation can be achieved in as little as 4 steps. In this example, we use the CopasiProcess implementation to compose a sbml-based simulation.

1. Define the composite instance according to the process_bigraph.Composite API and relative process implementation (in this case the CopasiProcess). Each instance of the Copasi process requires the specification of an SBML model file, which is specified in the inner key, 'config' :

     from process_bigraph import Composite, pf
   
     instance = {
          'copasi': {
              '_type': 'process',
              'address': 'local:copasi',
              'config': {
                  'model_file': 'biosimulator_processes/tests/model_files/Caravagna2010.xml'
              },
              'inputs': {
                  'floating_species': ['floating_species_store'],
                  'model_parameters': ['model_parameters_store'],
                  'time': ['time_store'],
              },
              'outputs': {
                  'floating_species': ['floating_species_store'],
                  'time': ['time_store'],
              }
          },
          'emitter': {
              '_type': 'step',
              'address': 'local:ram-emitter',
              'config': {
                  'ports': {
                      'inputs': {
                          'floating_species': 'tree[float]',
                          'time': 'float'
                      },
                      'output': {
                          'floating_species': 'tree[float]',
                          'time': 'float'
                      }
                  }
              },
              'inputs': {
                  'floating_species': ['floating_species_store'],
                  'time': ['time_store']
              }
          }
     }
   

   As you can see, each instance definition is expected to have the following key heirarchy:

     instance[
        <INSTANCE-NAME>['_type', 'address', 'config', 'inputs', 'outputs'], 
        ['emitter']['_type', 'address', 'config', 'inputs', 'outputs']
     ]
   

   Each instance requires at least one process and one emitter. Usually, there may be multiple processes and just one emitter, thereby sharing memory amongst the chained processes.

   Both <INSTANCE-NAME> and 'emitter' share the same inner keys. Here, pay close attention to how the 'address' is set for both the instance name and emitter.

2. Create a process_bigraph.Composite instance:

     workflow = Composite({
        'state': instance
     })
   

3. Run the composite instance which is configured by the instance that we defined:

     workflow.run(10)
   

4. Gather and pretty print results:

     results = workflow.gather_results()
     print(f'RESULTS: {pf(results)}')
   

A simplified view of the above script:

     from process_bigraph import Composite, pf

     >> instance = {
             'copasi': {
                 '_type': 'process',
                 'address': 'local:copasi',
                 'config': {
                     'model_file': 'biosimulator_processes/tests/model_files/Caravagna2010.xml'
                 },
                 'inputs': {
                     'floating_species': ['floating_species_store'],
                     'model_parameters': ['model_parameters_store'],
                     'time': ['time_store'],
                 },
                 'outputs': {
                     'floating_species': ['floating_species_store'],
                     'time': ['time_store'],
                 }
             },
             'emitter': {
                 '_type': 'step',
                 'address': 'local:ram-emitter',
                 'config': {
                     'ports': {
                         'inputs': {
                             'floating_species': 'tree[float]',
                             'time': 'float'
                         },
                         'output': {
                             'floating_species': 'tree[float]',
                             'time': 'float'
                         }
                     }
                 },
                 'inputs': {
                     'floating_species': ['floating_species_store'],
                     'time': ['time_store']
                 }
             }
        }

     >> workflow = Composite({
           'state': instance
        })

     >> workflow.run(10)
     >> results = workflow.gather_results()
     >> print(f'RESULTS: {pf(results)}')