Documentation
¶
Overview ¶
Package semantic is the port through which the KPIs Cartograph records leave for, and later come back from, a semantic layer: the warehouse's own definitions of what each number is (TAXONOMY.md D57, docs/adr/0026). The engine builds a Layer from the record in the semantic layer's own terms (semantic models over warehouse models, measures, metrics) and an adapter writes it in one tool's syntax. The dbt adapter beside this package is the one implementation today; a distribution may add another.
Index ¶
Constants ¶
const ( Simple = "simple" Ratio = "ratio" Cumulative = "cumulative" Derived = "derived" )
Metric types, as the semantic layer names them.
Variables ¶
var ErrNoFormat = errors.New("no semantic layer exporter for that format")
ErrNoFormat is an export asked for in a syntax no exporter writes.
Functions ¶
This section is empty.
Types ¶
type Dimension ¶
type Dimension struct {
Name, Column string
}
Dimension is a categorical column a metric may be split by.
type Exporter ¶
type Exporter interface {
// Format names the syntax, such as dbt.
Format() string
Export(l Layer) ([]File, error)
}
Exporter writes a layer in one tool's syntax. The same layer always gives the same files.
type Importer ¶
Importer reads a layer back from one tool's files, so metrics defined in the warehouse can be recorded as KPIs. No adapter implements it yet.
type Layer ¶
Layer is a whole semantic layer: the models the data sources are and the metrics the KPIs are.
type Measure ¶
type Measure struct {
Name, Agg, Expr string
}
Measure is an aggregation over a model's rows.
type Metric ¶
type Metric struct {
// Name is the metric's name; KPI the KPI it is.
Name, KPI string
Label, Description string
Type string
// Measure is the measure of a simple or cumulative metric, by name.
Measure string
// Numerator and Denominator are, for a ratio, metrics by name.
Numerator, Denominator string
// Window is how far back a cumulative metric sums; empty for all time.
Window string
// Expr and Uses are a derived metric's expression and the metrics,
// by name, it uses.
Expr string
Uses []string
Filter string
}
Metric is one number, as the semantic layer defines it.
type Model ¶
type Model struct {
// Name is the semantic model's name; Source the data source it is.
Name, Source string
Description string
// Ref is the warehouse model it is built on.
Ref string
// Entity is what one row is, Key the column that names a row.
Entity, Key string
// TimeColumn is the time each row is counted at, Grain its finest.
TimeColumn, Grain string
Dimensions []Dimension
Measures []Measure
}
Model is a semantic model: a data source as it sits in the warehouse.
Source Files
Directories
¶
| Path | Synopsis |
|---|---|
|
Package conformance holds any semantic layer exporter to the port's promises: every KPI and data source in the layer is in what it writes, and the same layer always gives the same files.
|
Package conformance holds any semantic layer exporter to the port's promises: every KPI and data source in the layer is in what it writes, and the same layer always gives the same files. |
|
Package dbt writes a semantic layer as dbt's: one YAML file of semantic_models and metrics, in the shape dbt's semantic layer reads (dbt 1.6 and later), for an analytics engineer to put in a dbt project beside the models it refers to.
|
Package dbt writes a semantic layer as dbt's: one YAML file of semantic_models and metrics, in the shape dbt's semantic layer reads (dbt 1.6 and later), for an analytics engineer to put in a dbt project beside the models it refers to. |