package main import ( "encoding/json" "fmt" "strings" "testing" ) // A reporting-only partner keeps its entry, but is not a source of the file: // publisher: must not match it and the source category must not list it. func TestReportingOnlyIsNotASource(t *testing.T) { var d Document if err := json.Unmarshal([]byte(`{"publication": [ {"publisher": "The Intercept", "title": "Norway Used NSA Technology", "date_published": "2018-03-01", "reporters": ["Gallagher, Ryan"]}, {"publisher": "NRK", "title": "Antennene som samler inn data", "date_published": "2018-03-01", "reporters": ["Hopland, Standahl"], "reporting_only": true} ]}`), &d); err != nil { t.Fatal(err) } if got := fmt.Sprint(facetValues(&d, "PU")); got != "[The Intercept]" { t.Errorf("publisher values %s, want only [The Intercept]", got) } // The /browse/publisher/{outlet} index is built from the same values, so // the page and the facet agree: the publisher is a source, the partner // that only reported the story is not. The index is package state, so // put the corpus back for whatever runs next. prev := corpus().archive.Docs t.Cleanup(func() { corpus().archive.Docs = prev; corpus().buildCategoryDocs() }) corpus().archive.Docs = []Document{d} corpus().buildCategoryDocs() if got := len(corpus().categoryMatches("publisher", "The Intercept")); got != 1 { t.Errorf("the publishing outlet's page lists %d documents, want 1", got) } if got := len(corpus().categoryMatches("publisher", "NRK")); got != 0 { t.Errorf("the reporting-only partner's page lists %d documents, want 0", got) } // The entry itself, its byline and its article survive. if len(d.Publication) != 2 || d.Publication[1].Publisher != "NRK" || len(d.Publication[1].Journalists) != 1 { t.Errorf("the reporting-only entry lost data: %+v", d.Publication) } if !strings.Contains(fmt.Sprint(d.JournalistsDeduped()), "Hopland") { t.Error("a partner's reporters should still be credited") } }