Report format (`easybi/v1`)¶

{
  "schema": "easybi/v1",
  "title": "Dashboard",
  "visuals": [
    {
      "title": "Revenue by region and channel",
      "model": "sales",
      "type": "Bar",
      "axis": [ { "field": "Region", "grain": "None", "labels": { "Île-de-France": "IDF" } } ],
      "legend": { "field": "Channel", "grain": "None" },
      "legendPosition": "Bottom",
      "values": [ { "measure": "Amount" }, { "measure": "Count", "seriesType": "Line", "secondaryAxis": true } ],
      "tooltip": { "showPercent": true, "values": [ { "measure": "Quantity" } ], "template": "{category}: {value}" },
      "filters": [
        { "field": "Product.Category", "operator": "In", "values": [ "Services" ] },
        { "field": "Date", "operator": "Last", "values": [ "12" ], "grain": "Month" }
      ]
    }
  ],
  "filters": [
    { "model": "sales", "filter": { "field": "Region", "operator": "In", "values": [ "North", null ] } }
  ]
}

A visual's type is the id of its visual type: one of EasyBI's (Bar, StackedBar, Line, Area, Scatter, Pie, Donut, Waterfall, Kpi, Gauge, Table, Matrix, Text, Image, Variable, read whatever their case) or one an application declares. A measure's seriesType draws its series with another cartesian type than the visual's.

null in values means a missing value. Besides In and NotIn, a date filter's operator can be Between (values: from and to as yyyy-MM-dd, inclusive, null for an open end), Current, Last or Previous (value: the number of periods; grain is their unit). A number dimension or, in a visual, a measure can be compared: GreaterThan, GreaterThanOrEqual, LessThan, LessThanOrEqual (value: the number, invariant) or Range (values: minimum and maximum, inclusive, null for an open end). An AnyOf filter keeps the rows matching every filter of one of its alternatives (its field and values are empty): cross-filtering builds one when several selected elements differ on more than one dimension, { "field": "", "operator": "AnyOf", "values": [], "alternatives": [[...], [...]] }. EasyBIContext.Validate(report) reports the unknown models and fields of a loaded report, and the formulas that no longer match their model.

Calculated measures created by the end user are saved with the report (the property is omitted when there are none), and visuals reference them by name like a field:

"measures": [
  { "name": "calc_3f9a1c0d2b7e", "caption": "Margin rate", "model": "sales", "formula": "[Margin] / [Amount]", "format": "P1" },
  { "name": "Amount", "caption": "Average revenue", "model": "sales", "formula": "AVERAGE([Amount])", "format": "N0" }
]

A measure named like a measure of its model (the second one) redefines it in the report; an empty caption keeps the model's.

The report's variables are saved with their current values (the property is omitted when there are none). Values are invariant, like filter values:

"variables": [
  { "name": "N", "model": "sales", "formula": "10 BETWEEN 3 AND 25", "value": "12", "format": "N0" },
  { "name": "TopCustomers", "model": "sales", "formula": "SELECT TOP [@N] [Customer] ORDER BY [Amount] DESC" }
]

A variable is its formula, stored with field names (see the formulas of variables); value is the value the end user gave a parameter, omitted while it is its formula's. Visuals reference variables by name: a filter value "[@N]" (a set's values with In and NotIn), "topVariable": "N", a field { "field": "Customer", "set": "TopCustomers" } grouped in true (the set's values) and false (the others), [@N] in a formula, {N} in a title or a text, and { "type": "Variable", "variable": "N" } shows one on the dashboard.

Text and image visuals carry their content and no model:

{ "type": "Text", "text": { "text": "Sales dashboard", "fontSize": 24, "isBold": true, "alignment": "Left" } }
{ "type": "Image", "image": { "data": "iVBORw0KGgo...", "mediaType": "image/png", "fit": "Contain" } }

Dashboard file (`easybi-dashboard/v1`)¶

DashboardSerializer saves what a DashboardView shows: the unchanged report next to the WPF layout, which references visuals by id. Sizes that are not saved are greedy. It also opens a plain easybi/v1 report (visuals then get a default layout).

{
  "schema": "easybi-dashboard/v1",
  "report": { "schema": "easybi/v1", "visuals": [ ... ], "filters": [ ... ] },
  "layout": {
    "columns": 12,
    "rowHeight": 110,
    "bands": [
      {
        "mode": "Fit",
        "items": [
          { "visualId": "...", "rowSpan": 1 },
          { "split": "Row", "children": [ { "visualId": "...", "columnSpan": 6 }, { "visualId": "..." } ] }
        ]
      },
      { "mode": "Scroll", "items": [ { "visualId": "...", "width": 360 }, { "visualId": "..." } ] }
    ]
  }
}

The same file renders identically with any connector.

Report format (easybi/v1)¶

Dashboard file (easybi-dashboard/v1)¶

Report format (`easybi/v1`)¶

Dashboard file (`easybi-dashboard/v1`)¶