Files
sm-seo-plugin/registry_test.go
Jakub Zych 49d3879954 feat: links settings and outbound rules in resolve
- migration 202610280003: settings links_new_tab (default on),
  nofollow_domains and sponsored_domains; robots llms_enabled and
  llms_intro (used by the llms.txt commit)
- SEO settings Links tab: the new-tab switch and the two domain lists,
  one domain per line, lowercased, refused with settings.domain_invalid
  naming the line; writable through seo_update_settings
- the resolve answer of an ok page carries outbound {new_tab,
  nofollow_domains, sponsored_domains}; no existing key changes
2026-11-04 13:26:31 +01:00

387 lines
20 KiB
Go

package seo_test
import (
"encoding/json"
"net/http"
"strings"
"testing"
"git.golem15.com/golem15/sm-sitemap-plugin/contract"
"git.golem15.com/golem15/summercms/modules/signpost"
)
// pagesAPI is the Pages screen's admin API.
var pagesAPI = adminAPI("/golem15/seo/pages")
// createRegistryPage creates a registry page through the admin form and
// returns its id.
func (e *testEnv) createRegistryPage(t *testing.T, token string, body map[string]any) string {
t.Helper()
return createdID(t, e.expect(t, http.StatusCreated, http.MethodPost, pagesAPI, token, body).Body.Bytes())
}
// refusal posts body and expects a 422 whose answer contains want.
func (e *testEnv) refusal(t *testing.T, method, path, token string, body map[string]any, want string) {
t.Helper()
rec := e.call(t, method, path, token, body)
if rec.Code != http.StatusUnprocessableEntity || !strings.Contains(rec.Body.String(), want) {
t.Fatalf("%s %s %v = %d %s, want 422 with %q", method, path, body, rec.Code, rec.Body.String(), want)
}
}
// TestRegistry covers the page registry (D-33, UI-SPEC C21): a page made
// in the Pages screen is a seo-page record with the SEO tab, resolves ok,
// is listed in the sitemap file of its kind with hreflang alternates, the
// C21 refusals, the delete, the filter groups and the management API.
func TestRegistry(t *testing.T) {
env := bootEnv(t)
token := env.admin(t, true)
contact := env.createRegistryPage(t, token, map[string]any{"url": map[string]any{"en": "/kontakt", "pl": ""}, "title": map[string]any{"en": "Kontakt"}, "kind": "page"})
got := env.resolve(t, "/kontakt")
seoBody, _ := got["seo"].(map[string]any)
if got["action"] != "ok" || !strings.Contains(seoBody["title"].(string), "Kontakt") {
t.Fatalf("resolve /kontakt = %v, want ok with the page name in the title", got)
}
// The registry is the content type seo-page with the SEO tab.
reg, _ := env.app.Lookup[*signpost.Registry]()
if typ, ok := reg.ContentType("seo-page"); !ok || typ.Content == nil || typ.Content.Controller != "golem15.seo.pages" || typ.Content.Kind != "page" {
t.Fatalf("seo-page content type = %+v ok=%v", typ, ok)
}
rec, err := reg.Describe(t.Context(), "seo-page", contact, "")
if err != nil || rec == nil || len(rec.Locales) != 1 || rec.Locales[0].URL != "/kontakt" || rec.Locales[0].Title != "Kontakt" || rec.LastMod == nil {
t.Fatalf("describe = %+v err=%v, want one en locale /kontakt with a content date (no pl address)", rec, err)
}
findField(t, env.formSchema(t, token, "golem15/seo/pages"), "seo_title")
env.expect(t, http.StatusOK, http.MethodPut, pagesAPI+"/"+contact, token, map[string]any{"seo_title": map[string]any{"en": "Contact us"}})
if seoBody, _ := env.resolve(t, "/kontakt")["seo"].(map[string]any); seoBody["title"] != "Contact us" {
t.Fatalf("after the SEO tab save the title = %v, want Contact us", seoBody["title"])
}
// Two locales: two URLs and hreflang alternates.
about := env.createRegistryPage(t, token, map[string]any{"url": map[string]any{"en": "/about", "pl": "/o-firmie"}, "title": map[string]any{"en": "About", "pl": "O firmie"}})
if got := env.resolve(t, "/o-firmie"); got["action"] != "ok" || got["locale"] != "pl" {
t.Fatalf("resolve /o-firmie = %v, want ok in pl", got)
}
show := env.expect(t, http.StatusOK, http.MethodGet, pagesAPI+"/"+about, token, nil).Body.String()
if !jsonContains(show, `"url":{"en":"/about","pl":"/o-firmie"}`) || !jsonContains(show, `"title":{"en":"About","pl":"O firmie"}`) {
t.Fatalf("show = %s, want the address and the name per locale", show)
}
offer := env.createRegistryPage(t, token, map[string]any{"url": map[string]any{"en": "/oferta"}, "title": map[string]any{"en": "Oferta"}, "kind": "offer"})
env.rebuildSitemap(t)
pages := env.sitemapFile(t, "sitemap-pages-1.xml")
for _, want := range []string{".test/kontakt</loc>", ".test/about</loc>", ".test/o-firmie</loc>", `hreflang="pl"`} {
if !strings.Contains(pages, want) {
t.Fatalf("sitemap-pages-1.xml lacks %s:\n%s", want, pages)
}
}
if strings.Contains(pages, "/oferta<") {
t.Fatalf("the offer is in the pages file:\n%s", pages)
}
if offers := env.sitemapFile(t, "sitemap-offers-1.xml"); !strings.Contains(offers, ".test/oferta</loc>") {
t.Fatalf("sitemap-offers-1.xml lacks the offer:\n%s", offers)
}
// The C21 refusals.
env.createPage(t, token, map[string]any{"title": "Lwia grzywa", "slug": "lwia-grzywa", "published": true})
env.refusal(t, http.MethodPost, pagesAPI, token, map[string]any{"url": map[string]any{"en": "/pages/lwia-grzywa"}}, "This address belongs to Page “Lwia grzywa”")
env.refusal(t, http.MethodPost, pagesAPI, token, map[string]any{"url": map[string]any{"en": "/kontakt"}}, "Another page already uses this address.")
env.refusal(t, http.MethodPost, pagesAPI, token, map[string]any{"url": map[string]any{"en": "/nowa", "pl": "/o-firmie"}}, "Another page already uses this address.")
env.refusal(t, http.MethodPost, pagesAPI, token, map[string]any{"url": map[string]any{"en": "kontakt"}}, "Start the address with / and leave out the domain.")
env.refusal(t, http.MethodPost, pagesAPI, token, map[string]any{"url": map[string]any{"en": "https://example.test/x"}}, "Start the address with / and leave out the domain.")
env.refusal(t, http.MethodPost, pagesAPI, token, map[string]any{"url": map[string]any{"en": "", "pl": "/tylko-pl"}}, "Enter the address in the default language.")
env.refusal(t, http.MethodPost, pagesAPI, token, map[string]any{"title": map[string]any{"en": "No address"}}, "Enter the address in the default language.")
env.expect(t, http.StatusOK, http.MethodPut, pagesAPI+"/"+about, token, map[string]any{"parent_id": contact})
env.refusal(t, http.MethodPut, pagesAPI+"/"+contact, token, map[string]any{"parent_id": about}, "A page cannot be placed under itself or one of its subpages.")
env.refusal(t, http.MethodPut, pagesAPI+"/"+contact, token, map[string]any{"parent_id": contact}, "A page cannot be placed under itself or one of its subpages.")
var count int64
if err := env.db.Raw(`SELECT COUNT(*) FROM golem15_seo_pages`).Scan(&count).Error; err != nil || count != 3 {
t.Fatalf("pages after the refusals = %d err=%v, want 3", count, err)
}
if rec, err := reg.Describe(t.Context(), "seo-page", about, ""); err != nil || len(rec.Parents) != 1 || rec.Parents[0].ID != contact {
t.Fatalf("about parents = %+v err=%v, want the contact page", rec, err)
}
// A changed address gets its 301, from the admin and from the API.
env.expect(t, http.StatusOK, http.MethodPut, pagesAPI+"/"+about, token, map[string]any{"url": map[string]any{"en": "/about-us", "pl": "/o-firmie"}})
if got := env.resolve(t, "/about"); got["action"] != "redirect" || got["location"] != siteURL+"/about-us" {
t.Fatalf("resolve of the old address = %v, want a redirect to /about-us", got)
}
env.expect(t, http.StatusOK, http.MethodPut, "/_seo/api/v1/pages/"+offer, token, map[string]any{"url": "/oferty"})
if got := env.resolve(t, "/oferta"); got["action"] != "redirect" || got["location"] != siteURL+"/oferty" {
t.Fatalf("resolve of the old API address = %v, want a redirect to /oferty", got)
}
// The filter's groups: the four kinds with their titles per locale.
filter, ok := env.app.Lookup[contract.Filter]()
if !ok {
t.Fatal("no contract.Filter published")
}
groups, err := filter.Groups(t.Context())
if err != nil || len(groups) != 4 {
t.Fatalf("groups = %+v err=%v", groups, err)
}
keys := []string{}
for _, g := range groups {
keys = append(keys, g.Key)
}
if strings.Join(keys, ",") != "pages,offers,services,landing-pages" || groups[3].Titles["en"] != "Landing page" || groups[0].Titles["pl"] != "Strona" {
t.Fatalf("groups = %+v", groups)
}
// Deleting a page deletes its SEO entries; it leaves the sitemaps.
env.expect(t, http.StatusOK, http.MethodDelete, pagesAPI+"/"+contact, token, nil)
var entries int64
if err := env.db.Raw(`SELECT COUNT(*) FROM golem15_seo_entries WHERE model_type = ? AND record_id = ?`, `Golem15\SEO\Models\Page`, contact).Scan(&entries).Error; err != nil || entries != 0 {
t.Fatalf("entries after delete = %d err=%v, want 0", entries, err)
}
env.rebuildSitemap(t)
if pages := env.sitemapFile(t, "sitemap-pages-1.xml"); strings.Contains(pages, ".test/kontakt</loc>") {
t.Fatalf("the deleted page is still listed:\n%s", pages)
}
if got := env.resolve(t, "/kontakt"); got["action"] != "not_found" {
t.Fatalf("resolve of the deleted page = %v, want not_found", got)
}
// The management API (the MCP tools' routes).
api := "/_seo/api/v1/pages"
created := env.expect(t, http.StatusCreated, http.MethodPost, api, token, map[string]any{"url": "/uslugi", "title": "Usługi", "kind": "service"}).Body.String()
var answer struct {
Data struct {
ID uint `json:"id"`
Origin string `json:"origin"`
URLs map[string]string `json:"urls"`
Kind string `json:"kind"`
} `json:"data"`
}
if err := json.Unmarshal([]byte(created), &answer); err != nil || answer.Data.ID == 0 || answer.Data.Origin != "manual" || answer.Data.Kind != "service" || answer.Data.URLs["en"] != "/uslugi" {
t.Fatalf("create answer = %s err=%v", created, err)
}
id := jsonID(answer.Data.ID)
env.expect(t, http.StatusOK, http.MethodPut, api+"/"+id, token, map[string]any{"url": map[string]any{"pl": "/uslugi-pl"}, "note": "x"})
list := env.expect(t, http.StatusOK, http.MethodGet, api+"?kind=service", token, nil).Body.String()
if !jsonContains(list, `"urls":{"en":"/uslugi","pl":"/uslugi-pl"}`) || !jsonContains(list, `"total":1`) {
t.Fatalf("list = %s", list)
}
rec2 := env.call(t, http.MethodPost, api, token, map[string]any{"url": "/about-us"})
if rec2.Code != http.StatusUnprocessableEntity || !strings.Contains(rec2.Body.String(), "Another page already uses this address.") {
t.Fatalf("API duplicate = %d %s", rec2.Code, rec2.Body.String())
}
env.expect(t, http.StatusOK, http.MethodDelete, api+"/"+id, token, nil)
env.expect(t, http.StatusNotFound, http.MethodDelete, api+"/"+id, token, nil)
other := env.admin(t, false)
env.expect(t, http.StatusForbidden, http.MethodGet, api, other, nil)
env.expect(t, http.StatusUnauthorized, http.MethodGet, api, "", nil)
}
// jsonID is an id as path text.
func jsonID(id uint) string {
raw, _ := json.Marshal(id)
return string(raw)
}
// pageRow reads a registry page's stored values by its address.
type pageRow struct {
ID uint `gorm:"column:id"`
URL string `gorm:"column:url"`
Title string `gorm:"column:title"`
Kind string `gorm:"column:kind"`
Origin string `gorm:"column:origin"`
Note string `gorm:"column:note"`
Stamp string `gorm:"column:stamp"`
}
func (e *testEnv) pageRow(t *testing.T, url string) pageRow {
t.Helper()
var rows []pageRow
if err := e.db.Raw(`SELECT id, url, title, kind, origin, note, COALESCE(reported_at::text, '') AS stamp FROM golem15_seo_pages WHERE url = ?`, url).Scan(&rows).Error; err != nil {
t.Fatal(err)
}
if len(rows) != 1 {
t.Fatalf("pages at %s = %d, want 1", url, len(rows))
}
return rows[0]
}
// reportRoutes posts a route report with the bearer token.
func (e *testEnv) reportRoutes(t *testing.T, token string, routes []string) map[string]any {
t.Helper()
rec := e.expect(t, http.StatusOK, http.MethodPost, "/_seo/api/v1/registry/routes", token, map[string]any{"routes": routes})
if rec.Header().Get("X-Robots-Tag") != "noindex" {
t.Fatalf("route report headers = %v, want X-Robots-Tag noindex", rec.Header())
}
var out map[string]any
if err := json.Unmarshal(rec.Body.Bytes(), &out); err != nil {
t.Fatal(err)
}
return out
}
// TestRouteReport covers the route report (D-33, P33-20, T-33-18): the
// token, the size cap, insert-only behaviour, reported_at, the missing
// badge and P9 warning, owned paths skipped, and that a later report never
// deletes a page or overwrites an editor's values.
func TestRouteReport(t *testing.T) {
off := bootEnv(t)
off.expect(t, http.StatusNotFound, http.MethodPost, "/_seo/api/v1/registry/routes", "anything", map[string]any{"routes": []string{"/"}})
// notfound_token is the fallback token.
fallback := bootEnv(t, "SUMMER_GOLEM15__SEO__NOTFOUND_TOKEN=nf-token")
fallback.reportRoutes(t, "nf-token", []string{"/x"})
env := bootEnv(t, "SUMMER_GOLEM15__SEO__ROUTE_REPORT_TOKEN=report-token", "SUMMER_GOLEM15__SEO__NOTFOUND_TOKEN=other")
admin := env.admin(t, true)
env.expect(t, http.StatusUnauthorized, http.MethodPost, "/_seo/api/v1/registry/routes", "wrong", map[string]any{"routes": []string{"/"}})
env.expect(t, http.StatusUnauthorized, http.MethodPost, "/_seo/api/v1/registry/routes", "other", map[string]any{"routes": []string{"/"}})
env.expect(t, http.StatusUnauthorized, http.MethodPost, "/_seo/api/v1/registry/routes", "", map[string]any{"routes": []string{"/"}})
many := make([]string, 2001)
for i := range many {
many[i] = "/p"
}
env.expect(t, http.StatusUnprocessableEntity, http.MethodPost, "/_seo/api/v1/registry/routes", "report-token", map[string]any{"routes": many})
env.createRegistryPage(t, admin, map[string]any{"url": map[string]any{"en": "/kontakt"}, "title": map[string]any{"en": "Kontakt"}})
env.createPage(t, admin, map[string]any{"title": "Blog post", "slug": "post", "published": true})
out := env.reportRoutes(t, "report-token", []string{"/", "/kontakt", "/o-nas", "/pages/post", "nope", "/o-nas"})
if out["inserted"] != float64(2) || out["updated"] != float64(1) || out["skipped"] != float64(2) {
t.Fatalf("first report = %v, want 2 inserted (/, /o-nas), 1 updated, 2 skipped", out)
}
about := env.pageRow(t, "/o-nas")
if about.Origin != "reported" || about.Kind != "page" || about.Title != "O nas" || about.Stamp == "" {
t.Fatalf("/o-nas = %+v, want origin reported, kind page, title O nas, reported", about)
}
if home := env.pageRow(t, "/"); home.Title != "Home" {
t.Fatalf("home = %+v, want the title Home", home)
}
contact := env.pageRow(t, "/kontakt")
if contact.Origin != "manual" || contact.Title != "Kontakt" || contact.Stamp == "" {
t.Fatalf("/kontakt = %+v, want the editor's page with reported_at", contact)
}
var owned int64
env.db.Raw(`SELECT COUNT(*) FROM golem15_seo_pages WHERE url = '/pages/post'`).Scan(&owned)
if owned != 0 {
t.Fatal("a path owned by a content type was inserted")
}
if got := env.resolve(t, "/o-nas"); got["action"] != "ok" {
t.Fatalf("resolve /o-nas = %v, want ok", got)
}
// The editor changes the reported page; the same report again changes
// only reported_at.
id := jsonID(about.ID)
env.expect(t, http.StatusOK, http.MethodPut, pagesAPI+"/"+id, admin, map[string]any{"title": map[string]any{"en": "About us"}, "kind": "offer", "note": "kept"})
var revisions int64
env.db.Raw(`SELECT COUNT(*) FROM golem15_seo_revisions`).Scan(&revisions)
before := env.pageRow(t, "/o-nas")
out = env.reportRoutes(t, "report-token", []string{"/", "/kontakt", "/o-nas"})
if out["inserted"] != float64(0) || out["updated"] != float64(3) {
t.Fatalf("second report = %v, want 0 inserted, 3 updated", out)
}
after := env.pageRow(t, "/o-nas")
if after.Title != "About us" || after.Kind != "offer" || after.Note != "kept" || after.Origin != "reported" || after.Stamp == before.Stamp {
t.Fatalf("after the second report /o-nas = %+v (before %+v), want only reported_at changed", after, before)
}
var revisionsAfter int64
env.db.Raw(`SELECT COUNT(*) FROM golem15_seo_revisions`).Scan(&revisionsAfter)
if revisionsAfter != revisions {
t.Fatalf("revisions %d -> %d: a report that inserts nothing writes no history", revisions, revisionsAfter)
}
// A report without /o-nas marks it missing and deletes nothing.
env.reportRoutes(t, "report-token", []string{"/", "/kontakt"})
missing := env.pageRow(t, "/o-nas")
if missing.Title != "About us" {
t.Fatalf("missing page = %+v, want it kept with its values", missing)
}
var list struct {
Meta struct {
RowStates map[string][]string `json:"row_states"`
} `json:"meta"`
}
if err := json.Unmarshal(env.expect(t, http.StatusOK, http.MethodGet, pagesAPI, admin, nil).Body.Bytes(), &list); err != nil {
t.Fatal(err)
}
if states := list.Meta.RowStates[id]; len(states) != 1 || states[0] != "negative" {
t.Fatalf("row states = %v, want /o-nas (%s) negative", list.Meta.RowStates, id)
}
if states := list.Meta.RowStates[jsonID(contact.ID)]; len(states) != 0 {
t.Fatalf("the manual page has row states %v", states)
}
status := env.expect(t, http.StatusOK, http.MethodGet, pagesAPI+"/partials/page_status?id="+id, admin, nil).Body.String()
if !strings.Contains(status, "summer-callout--warning") || !strings.Contains(status, "did not include this address") {
t.Fatalf("P9 of the missing page = %s", status)
}
if home := env.expect(t, http.StatusOK, http.MethodGet, pagesAPI+"/partials/page_status?id="+jsonID(env.pageRow(t, "/").ID), admin, nil).Body.String(); !strings.Contains(home, "Reported by the site on") || strings.Contains(home, "--warning") {
t.Fatalf("P9 of a reported page = %s", home)
}
if manual := env.expect(t, http.StatusOK, http.MethodGet, pagesAPI+"/partials/page_status?id="+jsonID(contact.ID), admin, nil).Body.String(); strings.Contains(manual, "summer-callout") {
t.Fatalf("P9 of a manual page = %s, want nothing", manual)
}
missingList := env.expect(t, http.StatusOK, http.MethodGet, "/_seo/api/v1/pages?missing=only", admin, nil).Body.String()
if !jsonContains(missingList, `"total":1`) || !jsonContains(missingList, `"missing":true`) {
t.Fatalf("missing=only list = %s", missingList)
}
// Reported again, it is back on the site.
env.reportRoutes(t, "report-token", []string{"/", "/kontakt", "/o-nas"})
if again := env.expect(t, http.StatusOK, http.MethodGet, "/_seo/api/v1/pages?missing=only", admin, nil).Body.String(); !jsonContains(again, `"total":0`) {
t.Fatalf("missing=only after the page is reported again = %s", again)
}
}
// outboundOf reads the resolve answer's outbound rules.
func outboundOf(t *testing.T, got map[string]any) map[string]any {
t.Helper()
out, ok := got["outbound"].(map[string]any)
if !ok {
t.Fatalf("resolve answer without outbound: %v", got)
}
return out
}
// TestSettingsLinks covers the Links tab (UI-SPEC C20, D-83, P33-16): the
// new-tab switch on by default, the domain lists with their line-numbered
// refusal, the MCP-writable keys and outbound in the resolve answer.
func TestSettingsLinks(t *testing.T) {
env := bootEnv(t)
token := env.admin(t, true)
findField(t, env.formSchema(t, token, "golem15/seo/settings"), "nofollow_domains")
env.createPage(t, token, map[string]any{"title": "About", "slug": "about", "published": true})
out := outboundOf(t, env.resolve(t, "/pages/about"))
if out["new_tab"] != true || len(out["nofollow_domains"].([]any)) != 0 || len(out["sponsored_domains"].([]any)) != 0 {
t.Fatalf("default outbound = %v, want new_tab true and no domains", out)
}
if got := env.resolve(t, "/pages/missing"); got["outbound"] != nil {
t.Fatalf("a not_found answer carries outbound: %v", got)
}
rec := env.call(t, http.MethodPut, adminAPI("/golem15/seo/settings/1"), token, map[string]any{"nofollow_domains": "example.com\nhttps://x.pl/a"})
if rec.Code != http.StatusUnprocessableEntity || !strings.Contains(rec.Body.String(), `Line 2: “https://x.pl/a” is not a domain.`) {
t.Fatalf("admin save of a bad domain = %d %s, want 422 naming line 2", rec.Code, rec.Body.String())
}
env.expect(t, http.StatusOK, http.MethodPut, adminAPI("/golem15/seo/settings/1"), token, map[string]any{"nofollow_domains": "Example.com\n\n shop.example.org "})
var stored string
if err := env.db.Raw(`SELECT nofollow_domains FROM golem15_seo_settings WHERE id = 1`).Scan(&stored).Error; err != nil || stored != "example.com\nshop.example.org" {
t.Fatalf("stored nofollow_domains = %q err=%v", stored, err)
}
api := "/_seo/api/v1/settings"
rec = env.call(t, http.MethodPut, api, token, map[string]any{"sponsored_domains": "partner.pl\nnot a domain"})
if rec.Code != http.StatusUnprocessableEntity || !strings.Contains(rec.Body.String(), "sponsored_domains") || !strings.Contains(rec.Body.String(), "Line 2") {
t.Fatalf("API save of a bad domain = %d %s", rec.Code, rec.Body.String())
}
env.expect(t, http.StatusOK, http.MethodPut, api, token, map[string]any{"sponsored_domains": "partner.pl", "links_new_tab": false})
show := env.expect(t, http.StatusOK, http.MethodGet, api, token, nil).Body.String()
if !jsonContains(show, `"links_new_tab":false`) || !jsonContains(show, `"sponsored_domains":"partner.pl"`) {
t.Fatalf("settings = %s", show)
}
out = outboundOf(t, env.resolve(t, "/pages/about"))
raw, _ := json.Marshal(out)
if string(raw) != `{"new_tab":false,"nofollow_domains":["example.com","shop.example.org"],"sponsored_domains":["partner.pl"]}` {
t.Fatalf("outbound = %s", raw)
}
}