Add data mapping.

[anthonytw]

Add data mapping / proxy datasets, similar to CCDEF, maybe adding additional flexibility.

/standard_clinical_numerics/.meta
{
    "columns": {},
    "proxies": {
        // Some datasets only have a single source, so we can use a direct mapping. This is
        // equivalent to what was proposed, e.g., pick your favorite signal for a given vital.
        // This can, of course, be used for all vitals if you wish. The major drawback is when
        // the source disappears you lose the vital, even if it's available from another source.
        "SpO2": {
            "type": "direct",
            "dataset": "/numerics/Pleth.SpO2"
        }
        // Some signals might have multiple sources, and maybe the right action is to choose based
        // on a priority list (if first is available, pick that, else, move down the list). This
        // allows use of a favored channel, but also falling back on other sources at times when
        // the preferred source is unavailable.
        "SysBP": {
            "type": "ranked",
            "datasets": [
                "/numerics/ART.Sys",
                "/numerics/CNAP.Sys",
                "/numerics/NBP.Sys"
            ]
        }
        // We might also just want to return all data points for a specific signal and let the user
        // do what they want with it, e.g., take the median.
        "HR": {
            "type": "all",
            "datasets": [
                "/numerics/ART.HR",
                "/numerics/ECG.HR",
                "/numerics/Pleth.HR"
            ]
        }
    }
}

And it could be used just like any other dataset:

// Returns only SpO2 from the Pleth, since it is a direct proxy. SpO2 available from other sources
// will be ignored.
spo2 = my_file['/standard_clinical_numerics'][start_time:end_time, 'SpO2']
// Returns the highest-priority data available for SysBP in the time range given, which could include
// data from multiple sources if a higher-priority source drops off or is added on.
sys = my_file['/standard_clinical_numerics'][start_time:end_time, 'SysBP']
// Returns all available data for HR in the time range given, because the proxy type is "all".
hr = my_file['/standard_clinical_numerics'][start_time:end_time, 'HR']