Add subruns information in DataKey of superruns to track metadata

[dachengx]

What is the problem / what does the code in this PR do

The subruns of a superrun is not checked when loading a saved superrun after redefining the superrun. You will find the example in the MWE.

Can you briefly describe how it works?

This PR does the following things:

1. Add an argument of DataKey: subruns, which is loaded from the metadata of superrun. Of course, if there is no metadata, an error will be raised.
2. Add Context.get_datakey for the context to get the decorated DataKey. Change all direct call of strax.DataKey in Context to Context.get_datakey, including Context.key_for.
3. Include the deterministic hash of subruns to the run_id in DataKey, so if everything is the same except subruns, the DataKey.__repr__ will still change.

The motivation that we do not include subruns in lineage is that we would like to keep the lineage hash of same data_type of subruns and superrun the same, like 0-records-j3nd2fjbiq and _superrun_test_uzafu3c2wk-records-j3nd2fjbiq.

Can you give a minimal working example (or illustrate with a figure)?

#!/usr/bin/env python
# coding: utf-8

import json
import datetime
import pytz
from bson import json_util

import numpy as np

import strax
from strax.testutils import Records, Peaks
import straxen


superrun_name = "_superrun_test"

# Prepare for a context
st = strax.Context(
    storage=[
        strax.DataDirectory(
            "./strax_data", provide_run_metadata=True, readonly=False, deep_scan=True
        )
    ],
    register=[Records, Peaks],
    config={"bonus_area": 42},
)
st.set_context_config({"use_per_run_defaults": False, "write_superruns": True})


def _write_run_doc(context, run_id, time, endtime):
    """Function which writes a dummy run document."""
    run_doc = {"name": run_id, "start": time, "end": endtime}
    with open(context.storage[0]._run_meta_path(str(run_id)), "w") as fp:
        json.dump(run_doc, fp, sort_keys=True, indent=4, default=json_util.default)


# Define run metadata of each subrun
offset_between_subruns = 10
now = datetime.datetime.now()
now.replace(tzinfo=pytz.utc)
subrun_ids = [str(r) for r in range(3)]
for run_id in subrun_ids:
    rr = st.get_array(run_id, "records")
    time = np.min(rr["time"])
    endtime = np.max(strax.endtime(rr))

    _write_run_doc(
        st,
        run_id,
        now + datetime.timedelta(0, int(time)),
        now + datetime.timedelta(0, int(endtime)),
    )

    st.set_config({"secret_time_offset": endtime + offset_between_subruns})  # untracked option
    assert st.is_stored(run_id, "records")

st.define_run(superrun_name, subrun_ids)
print(f"When subruns are: {list(st.run_metadata(superrun_name)['sub_run_spec'].keys())}:")
print(f"Old DataKey is {st.key_for(superrun_name, 'records')}")
st.make(superrun_name, "records")
assert st.is_stored(superrun_name, "records")

# But if you change the metadata definition...
st.define_run(superrun_name, subrun_ids[:-1])
print(f"When subruns are: {list(st.run_metadata(superrun_name)['sub_run_spec'].keys())}:")
print(f"New DataKey is {st.key_for(superrun_name, 'records')}")
assert not st.is_stored(superrun_name, "records")

The output will be

When subruns are: ['0', '1', '2']:
Old DataKey is _superrun_test_uzafu3c2wk-records-j3nd2fjbiq
When subruns are: ['0', '1']:
New DataKey is _superrun_test_zc44ingvpo-records-j3nd2fjbiq

Please include the following if applicable:

• Update the docstring(s)
• Update the documentation
• Tests to check the (new) code is working as desired.
• Does it solve one of the open issues on github?

Please make sure that all automated tests have passed before asking for a review (you can save the PR as a draft otherwise).

[dachengx]

The error in the GitHub action will be killed after updating the test of straxen.