ooni / probe-cli / 9398667008

72%

Build Type

push

github

Committed by web-flow

Commit Message

feat: introduce richer input (#1615)

Most of the existing code is designed to move around lists of
`model.OOAPIURLInfo` and measuring such URLs.

The `model.OOAPIURLInfo` type is like:

```Go
// internal/model/ooapi.go

type OOAPIURLInfo struct {
	CategoryCode string
	CountryCode string
	URL string
}
```

This design originally suited Web Connectivity but it's not good enough
for richer input because it does not contain options.

With this diff, we move into the direction of richer input by replacing
`model.OOAPIURLInfo` lists with lists of:

```Go
// internal/model/experiment.go

type ExperimentTarget struct {
	Category() string
	Country() string
	Input() string
}
```

where `*model.OOAPIURLInfo` implements `model.ExperimentTarget` in a
trivial way and where, additionally:

1. the `InputLoader` is modified to load `ExperimentTarget`;

2. the `Experiment` is modify to measure an `ExperimentTarget`.

In addition to applying these changes, this diff also adapts the whole
tree to use `ExperimentTarget` in all places and adds a trivial
constructor to obtain `OOAPIURLInfo` when the category code and the
country code are unknown.

With this diff merged, implementing richer input for real is a matter of
implementing the following changes:

1. the `*registry.Factory` has a new func field, defined by each
experiment, that loads a list of `ExperimentTarget`;

2. we have a library for input loading containing the same code that we
currently use for the input loader;

3. the `InputLoader` is gone and instead we use the factory (or its
`*engine.experimentBuilder` wrapper) for input loading;

4. we modify the `ExperimentArgs` passed to the `ExperimentMeasurer` to
contain an additional field that is the `ExperimentTarget` we want to
measure;

5. each experiment that needs richer input type-casts from the
`ExperimentTarget` interface to the concrete type that the experiment
richer input should have and acc... (continued)

Run Details

70 of 104 new or added lines in 27 files covered. (67.31%)

2 existing lines in 2 files now uncovered.

27444 of 32952 relevant lines covered (83.28%)

53.43 hits per line