Reporting objects

This tutorial will show you how pipestat can report not just primitive types, but structured results as objects.

First create a pipestat.PipestatManager object with our example schema:

import pipestat
psm = pipestat.PipestatManager(
    record_identifier="sample1",
    results_file_path="pipestat_results.yaml",
    schema_path="https://schema.databio.org/pipestat/object_result.yaml",
)

Initialize FileBackend

Here, we're pointing to a remote schema_path. Let's take a look at the schema object. You can see a preview of it here at the schema server: https://schema.databio.org/?namespace=pipestat&schema=object_result. We can also see what it looks like in Python code:

psm.schema

ParsedSchema (refget)
 Project-level properties:
 - None
 Sample-level properties:
 - value : {'type': 'string', 'description': 'Value of the object referred to by the key'}
 - mydict : {'type': 'object', 'description': 'Can pipestat handle nested objects?'}
 Status properties:
 - None

This schema defines two sample-level variables: value is a string, and mydict is an object. Let's see how to report a mydict:

psm.report({"mydict": {"toplevel": {"value": 456}}}, force_overwrite=True)

["Reported records for 'sample1' in 'refget' :\n - mydict: {'toplevel': {'value': 456}}"]

And now we can retrieve those results (which is returned as a python dict):

psm.retrieve_one("sample1", "mydict")

{'toplevel': {'value': 456}}

psm.retrieve_one("sample1", "mydict")['toplevel']

{'value': 456}

psm.retrieve_one("sample1", "mydict")['toplevel']['value']