llmgw/internal/web/mirror_explore_test.go

package web

import (
	"context"
	"errors"
	"html"
	"net/http"
	"net/http/httptest"
	"net/url"
	"os"
	"strings"
	"testing"
	"time"

	"gitea.mynodes.xyz/ilker/llmgw/internal/gguf"
	"gitea.mynodes.xyz/ilker/llmgw/internal/hf"
)

func ggufModel(id, arch string, params int64, created time.Time) hf.Model {
	return hf.Model{ID: id, CreatedAt: created, LastModified: created,
		GGUF: &hf.GGUFInfo{Total: params, Architecture: arch, ContextLength: 32768}}
}

func TestTheExplorerOpensOnTrendingModelsTheNodesCanRun(t *testing.T) {
	// Opened from the menu, with nothing typed: what is current, in the
	// shape this cluster runs, without a search to start it.
	now := time.Now()
	reg := &fakeRegistry{results: []hf.Model{
		ggufModel("Qwen/Qwen3-8B-GGUF", "qwen3", 8.2e9, now.Add(-90*24*time.Hour)),
		ggufModel("odd/Kev-GGUF", "ggmlc", 5e8, now.Add(-2*24*time.Hour)),
	}}
	body := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror?tab=explore").Body.String()

	q := reg.query
	if q.Sort != hf.SortTrending || q.Task != "text-generation" || !q.GGUF || q.Search != "" {
		t.Errorf("asked the registry for %+v", q)
	}
	if !strings.Contains(body, "Qwen/Qwen3-8B-GGUF") || !strings.Contains(body, "8.2B") {
		t.Errorf("the trending model is missing or unsized: %s", body)
	}
	// The kev case: listed, it invites a fetch the nodes cannot use.
	if strings.Contains(body, "odd/Kev-GGUF") {
		t.Error("a model llama.cpp cannot load is offered")
	}
	if !strings.Contains(body, "1 more is hidden") {
		t.Error("the page should say something was hidden, and why")
	}
}

func TestUntickingRunsHereShowsEverythingWithAVerdict(t *testing.T) {
	reg := &fakeRegistry{results: []hf.Model{
		ggufModel("odd/Kev-GGUF", "ggmlc", 5e8, time.Now()),
	}}
	// x=1 is the submitted form; "here" absent is the box unticked.
	body := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror?x=1&view=downloads&task=any&size=9").Body.String()

	q := reg.query
	if q.GGUF || q.Sort != hf.SortDownloads || q.Task != "" || q.MaxParams != 9e9 {
		t.Errorf("asked the registry for %+v", q)
	}
	if !strings.Contains(body, "odd/Kev-GGUF") || !strings.Contains(body, "ggmlc ✗") {
		t.Errorf("the unrunnable model should be listed and marked: %s", body)
	}
}

func TestNewThisMonthLeavesOutOlderRepositories(t *testing.T) {
	now := time.Now()
	reg := &fakeRegistry{results: []hf.Model{
		ggufModel("a/Old-GGUF", "llama", 1e9, now.Add(-200*24*time.Hour)),
		ggufModel("a/Fresh-GGUF", "llama", 1e9, now.Add(-3*24*time.Hour)),
	}}
	body := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror?view=new").Body.String()
	if strings.Contains(body, "a/Old-GGUF") || !strings.Contains(body, "a/Fresh-GGUF") {
		t.Errorf("new this month shows the wrong repositories: %s", body)
	}
}

func TestOpeningARepositoryKeepsTheWayBack(t *testing.T) {
	reg := &fakeRegistry{results: []hf.Model{ggufModel("a/B-GGUF", "llama", 1e9, time.Now())}}
	mux := mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg})
	body := html.UnescapeString(mirrorGet(t, mux, "/admin/mirror?view=new&size=3").Body.String())
	if !strings.Contains(body, "size=3") || !strings.Contains(body, "repo=a%2FB-GGUF") {
		t.Fatalf("the row's link should keep the filters: %s", body)
	}

	reg.repo = hf.Repo{ID: "a/B-GGUF", Revision: "abc123def456"}
	body = html.UnescapeString(mirrorGet(t, mux, "/admin/mirror?view=new&size=3&repo=a/B-GGUF").Body.String())
	if !strings.Contains(body, "back to the list") || !strings.Contains(body, "view=new") {
		t.Errorf("the repository page should lead back to the same listing: %s", body)
	}
}

func TestAnAdapterIsNotListedAsAModel(t *testing.T) {
	reg := &fakeRegistry{results: []hf.Model{
		ggufModel("a/Some-LoRA-GGUF", "qwen2", 2e6, time.Now()),
		ggufModel("a/Tiny-GGUF", "qwen2", 135e6, time.Now()),
	}}
	body := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror?tab=explore").Body.String()
	if strings.Contains(body, "Some-LoRA") || !strings.Contains(body, "a/Tiny-GGUF") {
		t.Errorf("adapters should be left out and small models kept: %s", body)
	}
}

func TestARepositoryShowsItsCard(t *testing.T) {
	reg := &fakeRegistry{repo: hf.Repo{
		ID: "Qwen/X-GGUF", Revision: "abc123def456", License: "apache-2.0",
		BaseModel: []string{"Qwen/X"}, Card: "# X\n\nA **small** model. See [the license](LICENSE).",
		Files: []hf.File{{Path: "x.gguf", Bytes: 10}},
	}}
	body := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror?repo=Qwen/X-GGUF").Body.String()
	for _, want := range []string{"<strong>small</strong>", "license apache-2.0", "<code>Qwen/X</code>",
		`href="https://huggingface.co/Qwen/X-GGUF/blob/abc123def456/LICENSE"`,
		`href="https://huggingface.co/Qwen/X-GGUF"`} {
		if !strings.Contains(body, want) {
			t.Errorf("missing %s", want)
		}
	}
}

func TestARepositoryOpensAsAFragmentBeneathItsRow(t *testing.T) {
	reg := &fakeRegistry{repo: hf.Repo{
		ID: "Qwen/X-GGUF", Revision: "abc123def456", Card: "A **small** model.",
		Files: []hf.File{{Path: "x.gguf", Bytes: 10}},
	}}
	rec := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror/repo?repo=Qwen/X-GGUF")
	body := rec.Body.String()
	if rec.Code != 200 {
		t.Fatalf("status %d: %s", rec.Code, body)
	}
	// The card and the fetch form, and not a page around them.
	for _, want := range []string{"<strong>small</strong>", `action="/admin/mirror/fetch"`, `name="csrf"`} {
		if !strings.Contains(body, want) {
			t.Errorf("missing %s: %s", want, body)
		}
	}
	for _, unwanted := range []string{"<html", "back to the list", "<h3>"} {
		if strings.Contains(body, unwanted) {
			t.Errorf("the fragment carries %s", unwanted)
		}
	}
}

func TestAFragmentForARegistryThatFailsSaysSo(t *testing.T) {
	reg := &fakeRegistry{err: errors.New("reaching the model registry: timeout")}
	rec := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror/repo?repo=Qwen/X-GGUF")
	if rec.Code != 502 || !strings.Contains(rec.Body.String(), "timeout") {
		t.Errorf("status %d: %s", rec.Code, rec.Body.String())
	}
}

func TestAnImportanceMatrixIsNotOfferedAsAModel(t *testing.T) {
	reg := &fakeRegistry{repo: hf.Repo{
		ID: "a/X-GGUF", Revision: "abc123def456", Architecture: "qwen2",
		Files: []hf.File{{Path: "x-q4_k_m.gguf", Bytes: 10}, {Path: "imatrix-x.gguf", Bytes: 5}},
	}}
	body := mirrorGet(t, mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg}),
		"/admin/mirror/repo?repo=a/X-GGUF").Body.String()
	if !strings.Contains(body, "importance matrix") {
		t.Errorf("the imatrix should be marked as not a model: %s", body)
	}
	if strings.Count(body, `action="/admin/mirror/fetch"`) != 1 {
		t.Errorf("only the model should have a fetch form: %s", body)
	}
}

func TestAFetchFromTheListAnswersWithItsOutcomeAlone(t *testing.T) {
	// Started from a repository opened in the list, a fetch answers the
	// page's script with the outcome, not a whole new page that would close
	// the repository and lose the list's place.
	m := &fakeMirror{}
	mux := mirrorPage(t, m, true)
	send := func(form url.Values) *httptest.ResponseRecorder {
		r := httptest.NewRequest("POST", "/admin/mirror/fetch", strings.NewReader(form.Encode()))
		r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
		r.Header.Set("Accept", "application/json")
		r.AddCookie(&http.Cookie{Name: SessionCookie, Value: "good"})
		r.AddCookie(&http.Cookie{Name: csrfCookie, Value: "tok"})
		w := httptest.NewRecorder()
		mux.ServeHTTP(w, r)
		return w
	}
	w := send(url.Values{"csrf": {"tok"}, "source": {"http://up/x.gguf"}, "model": {"x"}, "version": {"v1"}, "force": {"1"}})
	if w.Code != 200 || !strings.Contains(w.Body.String(), `"message":"started downloading x/v1`) {
		t.Errorf("%d %s", w.Code, w.Body.String())
	}
	if strings.Contains(w.Body.String(), "<html") {
		t.Error("a whole page came back")
	}
	w = send(url.Values{"csrf": {"tok"}, "model": {"x"}})
	if w.Code != 422 || !strings.Contains(w.Body.String(), `"error"`) {
		t.Errorf("a refusal: %d %s", w.Code, w.Body.String())
	}
}

// noProbe answers every file as an unreadable one, for tests that are not
// about what the files are.
func TestMain(m *testing.M) {
	kindOf = func(context.Context, string) (gguf.Header, error) {
		return gguf.Header{}, errors.New("not probed in tests")
	}
	os.Exit(m.Run())
}

func TestAdaptersAreListedForAModelAndToldFromModelsByTheirFiles(t *testing.T) {
	prev := kindOf
	t.Cleanup(func() { kindOf = prev })
	kindOf = func(_ context.Context, url string) (gguf.Header, error) {
		if strings.Contains(url, "adapter") {
			return gguf.Header{Architecture: "qwen2", Type: "adapter"}, nil
		}
		return gguf.Header{Architecture: "qwen2"}, nil
	}
	reg := &fakeRegistry{
		results: []hf.Model{{ID: "ggml-org/LoRA-Qwen2.5-Coder-7B-Instruct-F16-GGUF", Downloads: 900}},
		repo: hf.Repo{ID: "ggml-org/LoRA-Qwen2.5-Coder-7B-Instruct-F16-GGUF", Revision: "776530df0000",
			BaseModel: []string{"Qwen/Qwen2.5-Coder-7B-Instruct"},
			Files:     []hf.File{{Path: "lora-adapter-f16.gguf", Bytes: 726 << 20}, {Path: "merged-Q4_K_M.gguf", Bytes: 4 << 30}}},
	}
	mux := mirrorPageWith(t, &fakeMirror{}, reg, true)

	// From a held model's card: its repository is looked up, and the
	// adapters listed are those of the model it was made from.
	body := mirrorGet(t, mux, "/admin/mirror?tab=explore&view=adapters&from=bartowski/Qwen2.5-Coder-7B-Instruct-GGUF").Body.String()
	want := "base_model:adapter:Qwen/Qwen2.5-Coder-7B-Instruct"
	if len(reg.query.Filters) != 1 || reg.query.Filters[0] != want {
		t.Errorf("asked the registry for %v, expected %s", reg.query.Filters, want)
	}
	if !strings.Contains(body, "LoRA adapters made for <code>Qwen/Qwen2.5-Coder-7B-Instruct</code>") {
		t.Error("the page does not say whose adapters these are")
	}
	if strings.Contains(body, `aria-label="Task"`) {
		t.Error("task and size filters shown for adapters, which rarely carry either")
	}

	repo := mirrorGet(t, mux, "/admin/mirror?repo=ggml-org/LoRA-Qwen2.5-Coder-7B-Instruct-F16-GGUF").Body.String()
	if !strings.Contains(repo, "LoRA adapter · qwen2") {
		t.Error("the adapter file is not marked as one")
	}
	if !strings.Contains(repo, `<span class="opt">model · qwen2</span>`) {
		t.Error("the merged file is not marked as a model")
	}
}

func TestARepositoryWithNothingANodeLoadsIsNotListed(t *testing.T) {
	// Pickles only, or ONNX: a node refuses the first and no engine here
	// reads the second, so the row would only invite a fetch that fails.
	now := time.Now()
	pickles := hf.Model{ID: "someone/old-pytorch", CreatedAt: now, LastModified: now, Tags: []string{"pytorch"}}
	st := hf.Model{ID: "Qwen/Qwen3-8B", CreatedAt: now, LastModified: now, Tags: []string{"safetensors"}}
	reg := &fakeRegistry{results: []hf.Model{pickles, st, ggufModel("Qwen/Qwen3-8B-GGUF", "qwen3", 8.2e9, now)}}
	mux := mirrorPageLlama(t, MirrorDeps{Mirror: &fakeMirror{}, Registry: reg})

	body := mirrorGet(t, mux, "/admin/mirror?tab=explore").Body.String()
	if strings.Contains(body, "someone/old-pytorch") {
		t.Error("a repository with no GGUF and no safetensors is listed")
	}
	if !strings.Contains(body, "Qwen/Qwen3-8B<") && !strings.Contains(body, "Qwen/Qwen3-8B\"") || !strings.Contains(body, "Qwen/Qwen3-8B-GGUF") {
		t.Errorf("the loadable repositories should stay: %s", body)
	}
	if !strings.Contains(body, "neither GGUF nor safetensors") {
		t.Error("the page should say why one was hidden")
	}
	// Unticked, everything is shown.
	body = mirrorGet(t, mux, "/admin/mirror?x=1&view=downloads&task=any").Body.String()
	if !strings.Contains(body, "someone/old-pytorch") {
		t.Error("with the box unticked the repository should be listed")
	}
}

func TestARepositorysPickleFilesAreNotOffered(t *testing.T) {
	reg := &fakeRegistry{repo: hf.Repo{
		ID: "a/Both", Revision: "abc123def456",
		Files: []hf.File{{Path: "model.safetensors", Bytes: 10}, {Path: "pytorch_model.bin", Bytes: 10},
			{Path: "config.json", Bytes: 1}},
	}}
	mux := mirrorPageWith(t, &fakeMirror{}, reg, true)
	body := mirrorGet(t, mux, "/admin/mirror?repo=a/Both").Body.String()
	if strings.Contains(body, "pytorch_model.bin") {
		t.Error("the pickle file is offered")
	}
	if !strings.Contains(body, "model.safetensors") || !strings.Contains(body, "1 file is not shown") {
		t.Errorf("the safetensors should be offered and the pickle counted: %s", body)
	}

	reg.repo = hf.Repo{ID: "a/OnlyPickle", Revision: "abc123def456",
		Files: []hf.File{{Path: "model.ckpt", Bytes: 10}}}
	body = html.UnescapeString(mirrorGet(t, mux, "/admin/mirror?repo=a/OnlyPickle").Body.String())
	if strings.Contains(body, "/resolve/abc123def456/model.ckpt") || !strings.Contains(body, "only in a pickle format") {
		t.Errorf("a repository of pickles only should say so and offer nothing: %s", body)
	}
}