find_devices() from polartools.utils

[prjemian]

In discussions with 4ID-Polar today, they showed a new utility to identify ophyd devices/signals that share a common string. This tool (now) looks in the baseline stream for items which match (full or partial) a string.

Example: find m1 in the most recent run:

In [61]: find_devices(-1, db, "m1")
Out[61]: ['m1']

In [62]: find_devices(-1, db, "m")
Out[62]: ['time', 'm1']

In [63]: find_devices(-1, db, "")
Out[63]: ['time', 'm1']

Consider adding this to apstools.utils.

Thanks @strempfer & @gfabbris !

from https://github.com/APS-4ID-POLAR/polartools/blob/main/polartools/load_data.py

[prjemian]

def find_devices(scan, db, name, query=None):
    # so, for instance, if you give the name "lakeshore"
    # it will return all the devices with that string.
    table = load_databroker(scan, db, stream="baseline", query=query)
    output = []
    for col in table.columns:
        if name in col:
            output.append(col)
    return output

def read_baseline(scan, db, devices, query=None):
    table = load_databroker(scan, db, stream="baseline", query=query)
    return table[devices].mean()

def read_devices(scan, db, name, query=None):
    # companion to find_devices() above
    devices = find_devices(scan, db, name, query=query)
    return read_baseline(scan, db, devices)

def load_databroker(scan_id, db, stream="primary", query=None, use_db_v1=True):
    _db = db_query(db, query) if query else db
    if use_db_v1:
        return _db.v1[scan_id].table(stream_name=stream)
    else:
        return getattr(_db.v2[scan_id], stream).read().to_dataframe()

[prjemian]

The names seem rather broad. For example, find_devices is limited to finding devices in a single scan. All of these are focused on a single scan. Planning to append _in_scan to function names.

[prjemian]

Since read_baseline() uses devices as a single device, the call from read_devices() does not make much sense to me. Better to use the databroker v2 interface and get all this as a DataFrame or xarray table. Not planning to implement read_devices() without a better explanation of what it might provide that is not already available.

[prjemian]

Need definition of db_query()

[prjemian]

Compare with lookup_position() (from polartools) for read_devices()

[prjemian]

def lookup_position(db, scan, search_string="", query=None):
    """
    Lookup positioner values in past scans.
    Parameters
    ----------
    db : databroker database
        Searcheable database
    scan : integer
        Scan numbers or uids.
    search_string : string
        Full or part of positioner name.
    query: dict
        Search parameters.
    Returns
    -------
    output: list
    """

    db_range = db_query(db, query=query) if query else db

    baseline = load_databroker(scan, db_range, "baseline", use_db_v1=True)
    if len(baseline["time"]) == 2:
        date1 = baseline["time"][1].strftime("%m/%d/%y %H:%M:%S")
        date2 = baseline["time"][2].strftime("%m/%d/%y %H:%M:%S")
        print("=".center(100, "="))
        print(f"{'Positioner':>50}{date1:>25}{date2:>25}")
        print("-".center(100, "-"))
        for key in baseline.keys():
            if search_string in key:
                if isinstance(baseline[key][1], list):
                    print(f"{key:>50}{baseline[key][1]}{baseline[key][2]}")
                else:
                    print(
                        f"{key:>50}{baseline[key][1]:>25}{baseline[key][2]:>25}"
                    )

    else:
        date1 = baseline["time"][1].strftime("%m/%d/%y %H:%M:%S")
        print("=".center(100, "="))
        print(f"{'Positioner':>50}{date1:>25}")
        print("-".center(100, "-"))
        for key in baseline.keys():
            if search_string in key:
                if isinstance(baseline[key][1], list):
                    print(f"{key:>50}{baseline[key][1]}")
                else:
                    print(f"{key:>50}{baseline[key][1]:>25}")

    print("-".center(100, "-"))

[prjemian]

The index numbers 1 .. 2, ... are "sequence numbers" because baseline is a pandas DataFrame. Example:

In [1]: import databroker
d
In [2]: cat = databroker.catalog["training"]

In [3]: run = cat.v1[-1]

In [5]: baseline = run.table(stream_name="baseline")

In [6]: baseline["time"]
Out[6]: 
seq_num
1   2021-06-02 18:36:29.552511215
2   2021-06-02 18:36:34.705037355
Name: time, dtype: datetime64[ns]

In [7]: type(_)
Out[7]: pandas.core.series.Series

In [8]: 

[CrI3]

Awesome!