Closed cpritcha closed 8 years ago
One integration test depends on the deployr server. I could create a more full interface to deployr and mock the return values so it will run on Travis
Is there a reason you removed AnalysisParameter and ParameterValues? These are needed to run the analyses against the DeployR server..
The AnalysisOutput
model has a field named parameter_values_json
which we can use to store the parameter values. The DataAnalysisScript
has a field named parameters_json
which we could use to store the parameters. I thought that we should either get rid of the AnalysisParameter
and ParameterValues
tables or the JSON fields because they seem to store the same data. I chose to get rid of the tables because some of the raw json parameter values do not always have the same fields (in the luxedemo example, one of the controls has a valueList
field and the other has a valueRange
field).
Great!
Instead of using Project.name, use Project.slug for the short name (in extract_archive)
Task miracle.core.tasks.extract_archive[2a151f33-7803-482a-a640-f93798b8299a] raised unexpected: EncodeError(PicklingError("Can't pickle <class 'miracle.core.ingest.ProjectFilePath'>: attribute lookup miracle.core.ingest.ProjectFilePath failed",),)
...
EncodeError: Can't pickle <class 'miracle.core.ingest.ProjectFilePath'>: attribute lookup miracle.core.ingest.ProjectFilePath failed
restore AnalysisParameter and ParameterValues, add the valueList / valueRange fields to AnalysisParameter as django.contrib.postgres.fields.ArrayField as you suggested