package adapters

import (
	"context"
	"testing"

	"github.com/missionworkforce/missionmatch/internal/modules/ontology"
	"github.com/missionworkforce/missionmatch/internal/platform/extraction"
)

type stubText struct{ blobs []string }

func (s stubText) Texts(_ context.Context, yield func(string) error) error {
	for _, b := range s.blobs {
		if err := yield(b); err != nil {
			return err
		}
	}
	return nil
}

type stubExtractor struct{ perText map[string][]string }

func (s stubExtractor) ExtractResume(_ context.Context, _ string) (extraction.RawExtraction, error) {
	return extraction.RawExtraction{}, nil
}

func (s stubExtractor) ExtractRequirement(_ context.Context, text string) (extraction.RawExtraction, error) {
	return extraction.RawExtraction{Skills: s.perText[text]}, nil
}

func TestCorpusYieldsDedupedSkills(t *testing.T) {
	blob1 := "Seeking a SIGINT analyst with Splunk and AWS experience."
	blob2 := "Cloud engineer with AWS and Kubernetes."
	ex := stubExtractor{perText: map[string][]string{
		blob1: {"AWS", "Splunk"},
		blob2: {"AWS", "Kubernetes"},
	}}
	src := NewCorpusSourceWith(stubText{blobs: []string{blob1, blob2}}, ex)

	seen := map[string]int{}
	err := src.Fetch(context.Background(), func(r ontology.VocabRecord) error {
		if r.Kind != ontology.KindSkill {
			t.Fatalf("unexpected kind %q", r.Kind)
		}
		seen[r.Term]++
		return nil
	})
	if err != nil {
		t.Fatalf("fetch: %v", err)
	}
	if seen["AWS"] != 1 {
		t.Fatalf("AWS should be deduped to 1, got %d", seen["AWS"])
	}
	if seen["Splunk"] != 1 || seen["Kubernetes"] != 1 {
		t.Fatalf("expected Splunk and Kubernetes once each, got %v", seen)
	}
}