onesvm-browser-server/server/internal/dock/dock_test.go

// dock_test.go：适配器单测（无外网，httptest 环回 + 样本形状对齐 bench）。
//
// T1 契约：mock 响应形状 = bench 实测样本形状
// （searxng-cn/samples/t1-1.excerpt.json、trafilatura-http/samples/t2-1 与
// t2-cross-govcn.excerpt.json）。
package dock

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

	"onesvm.com/onesvm/browser-server/internal/contract"
)

// sampleSearxJSON 构造与 bench t1-1.excerpt.json 同形状的响应（字段名一致）。
func sampleSearxJSON() string {
	return `{
	  "query": "跨境电商 出口退税 政策 2026",
	  "number_of_results": null,
	  "unresponsive_engines": [["sogou", "Suspended: CAPTCHA"]],
	  "results": [
	    {"title": "海关总署公告", "url": "https://hainan.chinatax.gov.cn/xxgk_6_1/06163393.html",
	     "content": "对自2026年1月1日至2027年12月31日期间…免征进口关税", "engine": "baidu"},
	    {"title": "雨果跨境", "url": "https://m.cifnews.com/",
	     "content": "雨果跨境以雨果网作为流量依托", "engine": "bing", "publishedDate": "2026-08-23T00:00:00+08:00"}
	  ]
	}`
}

// TestSearxExecute 样本形状解析 + 查询参数对齐（format=json/safesearch=0）。
func TestSearxExecute(t *testing.T) {
	var gotPath string
	var gotQuery string
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		gotPath = r.URL.Path
		gotQuery = r.URL.RawQuery
		_, _ = w.Write([]byte(sampleSearxJSON()))
	}))
	defer srv.Close()
	a := NewSearx("searxng-cn", srv.URL, "zh-CN")
	if err := a.Init(context.Background()); err != nil {
		t.Fatalf("Init: %v", err)
	}
	job := contract.JobEnvelope{Intent: "search", Search: &contract.SearchInput{Query: "跨境电商 出口退税"}}
	raw, eb := a.Execute(context.Background(), job)
	if eb != nil {
		t.Fatalf("Execute 不应失败: %+v", eb)
	}
	if gotPath != "/search" {
		t.Errorf("路径应 /search，得 %s", gotPath)
	}
	for _, want := range []string{"format=json", "safesearch=0", "language=zh-CN"} {
		if !strings.Contains(gotQuery, want) {
			t.Errorf("query 缺 %s: %s", want, gotQuery)
		}
	}
	items, unres := mapSearxItems(raw)
	if len(items) != 2 {
		t.Fatalf("应解析 2 条，得 %d", len(items))
	}
	if items[0].Engine != "baidu" || items[0].URL == "" {
		t.Errorf("首条形状不符: %+v", items[0])
	}
	if len(unres) != 1 || !strings.Contains(unres[0], "sogou") {
		t.Errorf("unresponsive 应含 sogou: %v", unres)
	}
	// 健康位。
	if h := a.Health(); !h.OK {
		t.Errorf("健康位应为 true: %+v", h)
	}
}

// mapSearxItems 测试辅助：从 Extra 取回条目（与 scheduler 侧解析同形状）。
type mapSearxItem struct {
	Title         string  `json:"title"`
	URL           string  `json:"url"`
	Content       string  `json:"content"`
	Engine        string  `json:"engine"`
	PublishedDate *string `json:"publishedDate"`
}

func mapSearxItems(raw *contract.RawResult) ([]mapSearxItem, []string) {
	var items []mapSearxItem
	if v, ok := raw.Extra["searx_results"]; ok {
		if b, err := json.Marshal(v); err == nil {
			_ = json.Unmarshal(b, &items)
		}
	}
	var unres []string
	if v, ok := raw.Extra["unresponsive_engines"]; ok {
		if b, err := json.Marshal(v); err == nil {
			var pairs [][]string
			_ = json.Unmarshal(b, &pairs)
			for _, p := range pairs {
				if len(p) == 2 {
					unres = append(unres, p[0]+"("+p[1]+")")
				}
			}
		}
	}
	return items, unres
}

// TestSearxConcurrencyClamp 并发钳 6 + host 最小间隔：20 个并发请求串行化
// （150ms 间隔 → 总耗时 ≥150ms×(需等待次数)，至少验证无并发穿透）。
func TestSearxConcurrencyClamp(t *testing.T) {
	var mu sync.Mutex
	concurrent := 0
	maxConcurrent := 0
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		mu.Lock()
		concurrent++
		if concurrent > maxConcurrent {
			maxConcurrent = concurrent
		}
		mu.Unlock()
		time.Sleep(30 * time.Millisecond)
		mu.Lock()
		concurrent--
		mu.Unlock()
		_, _ = w.Write([]byte(sampleSearxJSON()))
	}))
	defer srv.Close()
	a := NewSearx("searxng-cn", srv.URL, "zh-CN")
	// 信号量钳 6 是硬上限：并发 20 打入，maxConcurrent ≤ 6。
	var wg sync.WaitGroup
	for i := 0; i < 20; i++ {
		wg.Add(1)
		go func() {
			defer wg.Done()
			_, _ = a.Execute(context.Background(), contract.JobEnvelope{
				Intent: "search", Search: &contract.SearchInput{Query: "q"}})
		}()
	}
	wg.Wait()
	if maxConcurrent > searxMaxConcurrent {
		t.Errorf("并发穿透：max=%d > %d", maxConcurrent, searxMaxConcurrent)
	}
}

// TestTrafilaturaContract 契约：请求体 {url,max_chars}；响应样本形状（成功 + empty_extract）。
func TestTrafilaturaContract(t *testing.T) {
	var gotBody map[string]any
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
		switch r.URL.Path {
		case "/health":
			_, _ = w.Write([]byte(`{"ok":true}`))
		case "/v1/read":
			_ = json.NewDecoder(r.Body).Decode(&gotBody)
			// t2-1.excerpt.json 同形状（成功）。
			_, _ = w.Write([]byte(`{"ok":true,"title":"跨境电子商务出口退运商品税收优惠政策公告",
		  "markdown":"| 索引号 | 11460000008174507Q/2026-14228 |\n\n为支持跨境电子商务新业态发展，现将…公告如下：",
		  "char_count":1122,"truncated":false,"url":"https://hainan.chinatax.gov.cn/xxgk_6_1/06163393.html"}`))
		}
	}))
	defer srv.Close()
	a := NewTrafilatura(srv.URL)
	_ = a.Init(context.Background())
	job := contract.JobEnvelope{Intent: "read",
		Read: &contract.ReadInput{URL: "https://hainan.chinatax.gov.cn/xxgk_6_1/06163393.html", MaxChars: 20000}}
	raw, eb := a.Execute(context.Background(), job)
	if eb != nil {
		t.Fatalf("Execute 不应失败: %+v", eb)
	}
	if gotBody["url"] != job.Read.URL || gotBody["max_chars"] != float64(20000) {
		t.Errorf("请求体契约不符: %v", gotBody)
	}
	if raw.Title == "" || raw.Markdown == "" {
		t.Errorf("RawResult 字段缺失: %+v", raw)
	}
	if raw.Extra["char_count"] != 1122 {
		t.Errorf("char_count 应透传 1122，得 %v", raw.Extra["char_count"])
	}
}

// TestTrafilaturaEmptyExtract t2-cross-govcn 形状：ok=false + fail_class=empty_extract
// → 上层触发降级链（此处断言错误为 upstream + 消息含 empty_extract）。
func TestTrafilaturaEmptyExtract(t *testing.T) {
	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
		_, _ = w.Write([]byte(`{"ok":false,"title":"","char_count":null,"truncated":null,
		  "url":"https://www.gov.cn/lianbo/202602/content_7057577.htm",
		  "error":"empty_extract","fail_class":"empty_extract","markdown_head":""}`))
	}))
	defer srv.Close()
	a := NewTrafilatura(srv.URL)
	job := contract.JobEnvelope{Intent: "read", Read: &contract.ReadInput{URL: "https://www.gov.cn/lianbo/x.htm"}}
	raw, eb := a.Execute(context.Background(), job)
	if raw != nil || eb == nil {
		t.Fatalf("空正文应返回错误: raw=%v eb=%v", raw, eb)
	}
	if eb.Code != contract.CodeUpstream {
		t.Errorf("empty_extract 应为 upstream（触发降级链），得 %s", eb.Code)
	}
}

// TestDetectVendor detectVendor 正则族（cdp_fetch.mjs 移植）逐类断言。
func TestDetectVendor(t *testing.T) {
	cases := []struct {
		title, text, htmlHead, hdr string
		looksBlocked               bool
		status, textLen            int
		want                       string
	}{
		{"Just a moment...", "", "<html><head>cf-challenge</head>", "", false, 403, 500, "cloudflare"},
		{"", "Please verify you are a human (captcha)", "", "", false, 200, 300, "captcha"},
		{"Access Denied", "", "", "server:AkamaiGHost", false, 403, 50, "waf"},
		{"", "", "", "", false, 200, 30, "empty"},
		{"Normal Page", strings.Repeat("正文内容", 100), "", "", false, 200, 400, "none"},
	}
	for i, c := range cases {
		got := detectVendor(c.title, c.text, c.htmlHead, c.hdr, c.looksBlocked, c.status, c.textLen)
		if got != c.want {
			t.Errorf("case %d: detectVendor=%s want=%s", i, got, c.want)
		}
	}
}

// TestExtractResultGolden CDP evaluate 返回形状 golden（字段名锁死）。
func TestExtractResultGolden(t *testing.T) {
	ex := extractResult{Title: "T", Text: "body", HTMLLen: 100, TextLen: 4,
		HTMLHead: "<html>", FinalURL: "https://x/", ReadyState: "complete", LooksBlocked: false}
	b, err := json.Marshal(ex)
	if err != nil {
		t.Fatal(err)
	}
	var m map[string]any
	if err := json.Unmarshal(b, &m); err != nil {
		t.Fatal(err)
	}
	for _, k := range []string{"title", "text", "htmlLen", "textLen", "htmlHead", "finalUrl", "readyState", "looksBlocked"} {
		if _, ok := m[k]; !ok {
			t.Errorf("extractResult 缺字段 %s（cdp_fetch.mjs EXTRACT_JS 形状漂移）", k)
		}
	}
}

// TestRegistryHealth 摘除语义：探活失败 → Health().ok=false → HealthyAdapters 过滤。
func TestRegistryHealth(t *testing.T) {
	bad := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
		w.WriteHeader(http.StatusServiceUnavailable)
	}))
	defer bad.Close()
	a := NewTrafilatura(bad.URL)
	_ = a.Init(context.Background())
	if h := a.Health(); h.OK {
		t.Fatalf("503 探活后应不健康: %+v", h)
	}
	reg := NewRegistry()
	reg.Register(a)
	if got := reg.HealthyAdapters("read"); len(got) != 0 {
		t.Errorf("不健康适配器应被摘除，得 %v", got)
	}
}

// TestTextToMarkdown 纯文本段落保持（\n\n 保段）。
func TestTextToMarkdown(t *testing.T) {
	in := "第一段第一行\n第一段第二行\n\n第二段"
	got := textToMarkdown(in)
	if !strings.Contains(got, "第一段第一行\n第一段第二行") {
		t.Errorf("段内换行应保留: %q", got)
	}
	if !strings.Contains(got, "\n\n第二段") {
		t.Errorf("段间应 \\n\\n: %q", got)
	}
}

// TestRewriteWsURL rewriteWs 语义（cdp_fetch.mjs：ws host 对齐 http 端点）。
func TestRewriteWsURL(t *testing.T) {
	got := rewriteWsURL("ws://127.0.0.1:9222/devtools/browser/abc", "lightpanda:9222")
	if got != "ws://lightpanda:9222/devtools/browser/abc" {
		t.Errorf("rewriteWsURL: %s", got)
	}
}

// TestSearxCaps 区域/意图能力标签。
func TestSearxCaps(t *testing.T) {
	cn := NewSearx("searxng-cn", "http://x", "zh-CN").Capabilities()
	if err := cn.Match("search", contract.RegionDomestic, contract.RenderNone); err != nil {
		t.Errorf("cn 应匹配 domestic search: %v", err)
	}
	if err := cn.Match("search", contract.RegionOverseas, contract.RenderNone); err == nil {
		t.Error("cn 不应匹配 overseas")
	}
	gl := NewSearx("searxng-global", "http://x", "en").Capabilities()
	if err := gl.Match("search", contract.RegionOverseas, contract.RenderNone); err != nil {
		t.Errorf("global 应匹配 overseas: %v", err)
	}
}

var _ = time.Second // 保导入