Check-Model.py

Check-Model.py is a Python3.x script that generates scriptfiles able to download, run and validate an instance of a model from the EBRAINS Model-Catalog.

EBRAINS Model-Catalog gives access to models developped and maintained by scientists.

A model can be implemented with multiple instances defined with a version number.
Check-Model.py downloads a single instance and runs the associated scriptfile to produce expected results.

---

How to use Check-Model.py

Check-Model.py uses hbp-validation-framework to use KnowledgeGraph v2:

• hbp-validation-client
• Documentation hbp-validation-client
• PyPi hbp-validation-framework

And uses the KG-Core to run with new KnowledgeGraph v3:

• kg-core
• kg-core Python Client

Python package to login to EBRAINS, get model's metadata and download an instance of this model.
hbp-validation-client is supposed to be installed.

Check-Model.py requieres some variables the user needs to have access:

• username, the username to login to EBRAINS,

• password, the password to login to EBRAINS,

• instance id, the instance identifier associated to the instance to download and run,

• workdir, the location where the scripts are and where the instance will be downloaded.
  No default value has been set at this moment, so the user needs to set this variable correctly.

---

How Check-Model.py operates

Check-Model.py is a script generator.
From instance id identifier, the model instance is retrieved and associated metadata are downloaded. Check-Model.py analyses these metadata and computes the instance's code location and download method. If the code is within an archive, the file is downloaded and extracted using Archiver.
Archiver has been chosen in order to normalize the extract instructions, but this method should be updated to use more native tools.
The extracted model instance is moved into a folder named with the instance identifier.

Finally, Check-Model.py calls the model instance's runscript, which is a script provided by the model custodian that runs the instance correctly.

To be clear, Check-Model.py generates a script WORKDIR/run_me.sh that

1. download instance's metadata,
2. computes code location,
3. download and extract instance's code into INSTANCE_ID subfolder,
4. cd to extracted code subfolder,
5. runs instance runscript

---

How to install/run Check-Model.py

There is no install process to install Check-Model.py at this time, but there are three dependencies:

1. Python-3.x,
2. PyPi hbp-validation-framework,
3. Archiver.

These must be first installed.

Example

A full example is proposed in check_model-kgv2.py that gives an example on how to use Instance objects using KnowledgeGraph v2.
Another full example would be proveded soon to illustrate KGV3 use.

It is highly recommended to check and validate the run_me.sh script file before running it.