package webfetch

import (
	"context"
	"fmt"
	"io"
	"net/http"
	"regexp"
	"strings"
	"time"

	"git.yeij.top/AskaEth/Cyrene_v2/internal/plugin"
)

type Plugin struct{ client *http.Client }

func New() *Plugin { return &Plugin{client: &http.Client{Timeout: 20 * time.Second}} }

func (p *Plugin) Name() string        { return "web-fetch" }
func (p *Plugin) Version() string     { return "1.0.0" }
func (p *Plugin) Description() string { return "网页抓取" }
func (p *Plugin) Init(_ context.Context, _ plugin.PluginConfig) error { return nil }
func (p *Plugin) Close() error        { return nil }
func (p *Plugin) Enabled() bool       { return true }

func (p *Plugin) Tools() []plugin.ToolDefinition {
	return []plugin.ToolDefinition{{
		Name:        "web_fetch",
		Description: "Fetch a web page and return its text content (HTML stripped, truncated).",
		Parameters: map[string]any{
			"type": "object",
			"properties": map[string]any{
				"url": map[string]string{"type": "string", "description": "Page URL (http/https)"},
			},
			"required": []string{"url"},
		},
	}}
}

var (
	reScript   = regexp.MustCompile(`(?is)<(script|style|noscript)[^>]*>.*?</(script|style|noscript)>`)
	reTags     = regexp.MustCompile(`(?s)<[^>]+>`)
	reBlank    = regexp.MustCompile(`\n\s*\n+`)
	maxBody    = 1 << 20   // 1MB
	maxContent = 4000      // runes returned to the LLM
)

func (p *Plugin) Execute(ctx context.Context, _ string, params map[string]any) (string, error) {
	rawURL, _ := params["url"].(string)
	if !strings.HasPrefix(rawURL, "http://") && !strings.HasPrefix(rawURL, "https://") {
		return "", fmt.Errorf("url must start with http:// or https://")
	}

	req, err := http.NewRequestWithContext(ctx, "GET", rawURL, nil)
	if err != nil {
		return "", err
	}
	req.Header.Set("User-Agent", "Mozilla/5.0 (compatible; Cyrene/2.0)")

	resp, err := p.client.Do(req)
	if err != nil {
		return "", fmt.Errorf("fetch: %w", err)
	}
	defer resp.Body.Close()
	if resp.StatusCode != 200 {
		return "", fmt.Errorf("fetch API %d", resp.StatusCode)
	}

	body, err := io.ReadAll(io.LimitReader(resp.Body, int64(maxBody)))
	if err != nil {
		return "", fmt.Errorf("read body: %w", err)
	}

	text := htmlToText(string(body))
	if runes := []rune(text); len(runes) > maxContent {
		text = string(runes[:maxContent]) + "\n…（内容过长已截断）"
	}
	if text == "" {
		return "页面无可用文本内容", nil
	}
	return text, nil
}

// htmlToText strips scripts/styles/tags and collapses whitespace.
func htmlToText(html string) string {
	html = reScript.ReplaceAllString(html, " ")
	html = reTags.ReplaceAllString(html, "\n")
	html = reBlank.ReplaceAllString(html, "\n")
	var lines []string
	for _, line := range strings.Split(html, "\n") {
		if s := strings.TrimSpace(line); s != "" {
			lines = append(lines, s)
		}
	}
	return strings.Join(lines, "\n")
}