Concepts / Terminologies

• Each “channel” produces a series of values.
• Query is made for a set of channels for a period of time.
• Query result is a set of channel data, indexed by the channel names.
• Each channel data consists of an array of values (i.e., series).
• If the length of the array is one, it could be represented as a single value instead of an array.
• Every value has an associated time-stamp (which does not have to be unique).
• A value can be a “scalar” (number, string or bool) or an “object” (histogram, table, tree, …).

Query Syntax and Reply Format

Query Format

http://ADDRESS/data/CHANNEL_LIST?OPTIONS

channel list is:

CH0,CH1,CH2...

options are:

length=LENGTH [default 3600]
to=TO [default 0 (now)]
resample=RESAMPLE [default 0 (auto)]
reducer=REDUCER [default "last"]

[TODO] Each channel can have options, which override the global options:

CH0{length=86400,resample=300},CH1,...

[TODO] and functors (see below for a full list of functors):

CH0{length=3600}->range[x](0,1000),CH1,...
CH0->histogram(100,0,100)

[TODO] As in the functor scheme, multiple channels can be combined using “data stack”:

CH0;CH1;->align()

Reply Packet Format

{
    CH0: {
        "start": START, "length": LENGTH,
        "t": [T0, T1, T2, ...], 
        "x": [X0, X1, X2, ...]
    },
    CH1: { 
    ...
}

where x[k] can be a scalar (simple time-series) or an object (object time-series: histogram, graph, table, tree or blob). t[k] is the time from START in seconds, or [TODO] for a series of time intervals, {"begin": tk_begin, "end": tk_end}.

Examples

Typical time-series of scalars

CH: {
    "start": START,  "length": LENGTH,
    "t": [T0, T1, T2, ...], 
    "x": [X0, X1, X2, ...]
}

Extensions

Reduced to a single value

If a time-series is reduced with last() etc., or if the data is time-neutral or “current”, the data consists of only one time point (which can be “neutral” or “current”):

CH: {
    "start": START,  "length": LENGTH,
    "t": T,
    "x": X
}

If the data is time-neutral, the value of T is 0. If the data time is “current”, the value of T will be the time of data query.

Bundling

[TODO] Instead of having multiple fields (such as y in addition to x), multiple time-series with the identical time points are indicated by the aligned property. This is used for:

• Multiple channels read at the same time
• Multiple channels alined by resampling
• Vectors, e.g., (x, y, z)
• Tuples, e.g., (mean, rms, min, max, n)
• Errors, e.g., (x, x_err)

[TODO] zip() and graph() transformers will make use of this.

Examples

Vectors:

CH0_X: {
    "start": START, "length": LENGTH,
    "aligned": ["CH0_Y", "CH0_Z"],
    "t": [T0, T1, T2, ...], 
    "x": [X0, X1, X2, ...],
}

Errors:

CH0: {
    "start": START, "length": LENGTH,
    "aligned": ["CH0_Error"],
    "t": [T0, T1, T2, ...], 
    "x": [X0, X1, X2, ...],
}

Tuples:

CH0_Mean: {
    "start": START, "length": LENGTH,
    "aligned": ["CH0_N", "CH0_RMS", "CH0_MIN", "CH0_MAX"]
    "t": [T0, T1, T2, ...], 
    "x": [X0, X1, X2, ...],
}

Value Types and JSON Representation

Scalar

X

• X can be a number, string, bool, or null.

Histogram

{ 
    "bins": { "min": MIN, "max": MAX }, 
    "counts": [ C0, C1, C2, ... ] 
}

• Having bins identifies the object as a Histogram.
• Number-of-bins is the length of the counts array.

2D Histogram

{ 
    "xbins": { "min": MIN, "max": MAX }, 
    "ybins": { "min": MIN, "max": MAX }, 
    "counts": [
        [ C00, C01, C02, ... ],
        [ C10, C11, C12, ... ],
        ...
    ]
}

• Having ybins identifies the object as a 2D Histogram.
• Number-of-bins is the length of the counts array.

Graph

{ 
    "labels": [ XLABEL, YLABEL, ZLABEL ],
    "x": [ x0, x1, x2, ... ],
    "y": [ y0, y1, y2, ... ],    
    "z": [ z0, z1, z2, ... ],
    "x_err": [ ex0, ex1, ex2, ... ],
    "y_err": [ ey0, ey1, ey2, ... ],
    "z_err": [ ez0, ez1, ez2, ... ]
}

• Having y identifies the object as a Graph.
• All the other fields are optional.
• If x does not exist, it will be filled with [0:len(y)].

Table

{
    "columns": [COLUMN0, COLUMN1, ...], 
    "table": [
        [ X00, X01, ...], 
        [ X10, X11, ...], 
        ...
    ],
    "attr": [
        [ A00, A01, ...], 
        [ A10, A11, ...], 
        ...
    ]
}

• Having table identifies the object as a Table.
• columns is optional.
• attr is optional. Axx are JSON objects.
  • Defined attributes are: color, background, href
• [TODO] row_attr and column_attr???

Tree

{
    "tree": {...},
    "attr": {...}
}

• Having tree identifies the object as a Tree.
• attr is optional. The tree structure of attr corresponds to tree, with values by JSON.
  • Defined attributes are: color, background, href

Blob

{
    "mime": MIME_TYPE,
    "id": ID,
    "meta": META_INFO
}

• Having mime identifies the object as a Blob.
• The id is required. Others are optional.