package knowledge_test

import (
	"context"
	"encoding/json"
	"math"
	"net/http"
	"net/http/httptest"
	"strings"
	"testing"

	"github.com/krow/krow-backend/go-api/internal/knowledge"
)

// The local embedder.
//
// Driven against a stub rather than a real Ollama, because what is being tested
// is this package's half of the contract: the request shape, the normalisation,
// and — most of all — what an operator is told when it does not work. The model
// itself is somebody else's code and testing it here would test the network.

func TestTheLocalEmbedderSendsWhatOllamaExpects(t *testing.T) {
	var got map[string]any
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		if r.URL.Path != "/api/embed" {
			t.Errorf("posted to %s, want /api/embed", r.URL.Path)
		}
		json.NewDecoder(r.Body).Decode(&got)
		json.NewEncoder(w).Encode(map[string]any{
			"embeddings": [][]float32{{3, 4}, {1, 0}},
		})
	}))
	defer srv.Close()

	e := knowledge.NewOllama(srv.URL, "nomic-embed-text", 2)
	out, err := e.Embed(context.Background(), []string{"a", "b"}, knowledge.KindDocument)
	if err != nil {
		t.Fatalf("embed: %v", err)
	}

	if got["model"] != "nomic-embed-text" {
		t.Errorf("model = %v", got["model"])
	}
	if inputs, ok := got["input"].([]any); !ok || len(inputs) != 2 {
		t.Errorf("input = %v; the batch should travel as a list", got["input"])
	}
	if len(out) != 2 {
		t.Fatalf("%d vectors, want 2", len(out))
	}
}

func TestTheLocalEmbedderNormalisesWhatItGetsBack(t *testing.T) {
	// The schema's similarity function is a plain dot product, which equals
	// cosine similarity ONLY for unit vectors. Ollama returns whatever the
	// model produced. Skipping this would not error — it would rank badly,
	// dominated by whichever chunks happened to have the largest magnitude,
	// which is the kind of wrong that never looks broken.
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
		json.NewEncoder(w).Encode(map[string]any{"embeddings": [][]float32{{3, 4}}})
	}))
	defer srv.Close()

	out, err := knowledge.NewOllama(srv.URL, "m", 2).
		Embed(context.Background(), []string{"x"}, knowledge.KindDocument)
	if err != nil {
		t.Fatalf("embed: %v", err)
	}

	var sum float64
	for _, v := range out[0] {
		sum += float64(v) * float64(v)
	}
	if math.Abs(math.Sqrt(sum)-1) > 1e-5 {
		t.Errorf("vector has length %.4f, want 1 — the dot product will not be cosine similarity",
			math.Sqrt(sum))
	}
}

func TestOllamaNotRunningSaysWhatToDo(t *testing.T) {
	// The single most likely failure, and the one where a bad message costs the
	// most time: an operator reading "connection refused" goes looking for a
	// network problem.
	e := knowledge.NewOllama("http://127.0.0.1:1", "nomic-embed-text", 768)
	_, err := e.Embed(context.Background(), []string{"x"}, knowledge.KindQuery)
	if err == nil {
		t.Fatal("embedding against nothing succeeded")
	}
	msg := err.Error()
	if !strings.Contains(msg, "ollama pull") {
		t.Errorf("the failure does not say how to fix it: %s", msg)
	}

	var kErr *knowledge.Error
	if !asErr(err, &kErr) || kErr.Code != knowledge.ErrEmbedUnavailable {
		t.Errorf("want %s, got %v", knowledge.ErrEmbedUnavailable, err)
	}
	// Retryable: a model that is starting up will answer in a moment.
	if !kErr.Retryable() {
		t.Error("an unreachable local model should be retryable")
	}
}

func TestAMissingModelIsADifferentProblemFromADeadServer(t *testing.T) {
	// Ollama up but never told to pull the model. Same symptom to a user, a
	// completely different fix — and one of them is not worth retrying.
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
		w.WriteHeader(http.StatusNotFound)
	}))
	defer srv.Close()

	_, err := knowledge.NewOllama(srv.URL, "nomic-embed-text", 768).
		Embed(context.Background(), []string{"x"}, knowledge.KindQuery)

	var kErr *knowledge.Error
	if !asErr(err, &kErr) {
		t.Fatalf("want a knowledge error, got %v", err)
	}
	if kErr.Code != knowledge.ErrNotConfigured {
		t.Errorf("a missing model reported %s; it is a configuration problem, not an outage", kErr.Code)
	}
	if kErr.Retryable() {
		t.Error("a model that was never pulled will not appear by retrying")
	}
	if !strings.Contains(kErr.Message, "ollama pull") {
		t.Errorf("the failure does not name the fix: %s", kErr.Message)
	}
}

func TestALocalCorpusIsNotConfusedWithAHostedOne(t *testing.T) {
	// Vectors from two models are not comparable, and the model name on the
	// chunk row is the only thing standing between that and confident nonsense.
	// A local `nomic-embed-text` and a hosted model of the same name must not
	// share an identity.
	local := knowledge.NewOllama("", "nomic-embed-text", 768)
	if !strings.HasPrefix(local.Model(), "ollama/") {
		t.Errorf("local model name is %q; it must be distinguishable from a hosted one", local.Model())
	}
	if local.Model() == knowledge.NewVoyage("k", "nomic-embed-text", 768).Model() {
		t.Error("a local and a hosted model with the same name share an identity")
	}
}