geo/server/internal/tenant/app/monitoring_citation_match.go

package app

import (
	"context"
	"database/sql"
	"net"
	"net/url"
	"sort"
	"strings"

	"github.com/jackc/pgx/v5"

	"github.com/geo-platform/tenant-api/internal/shared/response"
)

type monitoringPublishedAlias struct {
	ArticleID       int64
	PublishRecordID *int64
	ArticleTitle    string
	PublishPlatform string
	Ambiguous       bool
}

type monitoringPublishedAliasIndex map[string]monitoringPublishedAlias

type monitoringAliasQuerier interface {
	Query(ctx context.Context, sql string, args ...interface{}) (pgx.Rows, error)
}

func loadMonitoringPublishedAliasIndex(ctx context.Context, q monitoringAliasQuerier, tenantID int64) (monitoringPublishedAliasIndex, error) {
	rows, err := q.Query(ctx, `
		SELECT
			article_id,
			publish_record_id,
			COALESCE(article_title_snapshot, '未命名文章') AS article_title,
			COALESCE(publish_platform_name_snapshot, '已发布内容') AS publish_platform,
			original_url,
			normalized_url
		FROM monitoring_article_url_aliases
		WHERE tenant_id = $1
		  AND confidence = 'high'
		ORDER BY updated_at DESC, id DESC
	`, tenantID)
	if err != nil {
		return nil, responseInternalAliasLookupError()
	}
	defer rows.Close()

	index := make(monitoringPublishedAliasIndex)
	for rows.Next() {
		var item monitoringPublishedAlias
		var originalURL string
		var normalizedURL string
		var publishRecordID sql.NullInt64
		if scanErr := rows.Scan(&item.ArticleID, &publishRecordID, &item.ArticleTitle, &item.PublishPlatform, &originalURL, &normalizedURL); scanErr != nil {
			return nil, responseInternalAliasScanError()
		}
		item.PublishRecordID = nullableInt64Value(publishRecordID)
		for _, key := range monitoringCitationCandidateKeys(originalURL, normalizedURL) {
			existing, exists := index[key]
			if exists && existing.ArticleID != item.ArticleID {
				existing.Ambiguous = true
				index[key] = existing
				continue
			}
			if !exists {
				index[key] = item
			}
		}
	}
	if err := rows.Err(); err != nil {
		return nil, responseInternalAliasScanError()
	}

	return index, nil
}

func (index monitoringPublishedAliasIndex) Match(rawURL, normalizedURL string) (monitoringPublishedAlias, bool) {
	for _, key := range monitoringCitationCandidateKeys(rawURL, normalizedURL) {
		item, ok := index[key]
		if ok && !item.Ambiguous {
			return item, true
		}
	}
	return monitoringPublishedAlias{}, false
}

func monitoringCitationCandidateKeys(rawURL, normalizedURL string) []string {
	keys := []string{
		monitoringCitationURLMatchKey(rawURL),
		monitoringCitationURLMatchKey(normalizedURL),
	}
	result := make([]string, 0, len(keys))
	seen := make(map[string]struct{}, len(keys))
	for _, key := range keys {
		if key == "" {
			continue
		}
		if _, ok := seen[key]; ok {
			continue
		}
		seen[key] = struct{}{}
		result = append(result, key)
	}
	return result
}

func monitoringCitationURLMatchKey(raw string) string {
	trimmed := strings.TrimSpace(raw)
	if trimmed == "" {
		return ""
	}

	parsed, err := url.Parse(trimmed)
	if err != nil || parsed.Hostname() == "" {
		return ""
	}

	host := monitoringCanonicalCitationHost(parsed.Hostname())
	if host == "" {
		return ""
	}

	if strings.HasSuffix(host, "163.com") {
		if id := monitoringNeteaseArticleID(trimmed); id != "" {
			return "163.com/dy/article/" + strings.ToLower(id)
		}
	}

	if host == "baijiahao.baidu.com" {
		if id := extractBaijiahaoArticleIDFromURL(&trimmed); id != "" {
			return host + "/s?id=" + strings.ToLower(id)
		}
	}

	path := strings.TrimRight(parsed.EscapedPath(), "/")
	if path == "" {
		path = "/"
	}

	if host == "mp.weixin.qq.com" {
		if strings.HasPrefix(path, "/s/") && len(path) > len("/s/") {
			return host + path
		}
		if query := monitoringCanonicalQuery(parsed, []string{"__biz", "mid", "idx", "sn"}); query != "" {
			return host + path + "?" + query
		}
		return ""
	}

	if path == "/" {
		if query := monitoringCanonicalQuery(parsed, monitoringGenericContentQueryKeys()); query != "" {
			return host + path + "?" + query
		}
		return ""
	}

	if query := monitoringCanonicalQuery(parsed, monitoringGenericContentQueryKeys()); query != "" {
		return host + path + "?" + query
	}
	return host + path
}

func monitoringCanonicalCitationHost(value string) string {
	host := strings.ToLower(strings.TrimSpace(value))
	if host == "" {
		return ""
	}
	if withoutPort, _, err := net.SplitHostPort(host); err == nil {
		host = withoutPort
	}
	host = strings.Trim(host, ".")
	parts := strings.Split(host, ".")
	if len(parts) <= 2 {
		return host
	}
	switch parts[0] {
	case "www", "m", "wap", "mobile", "3g", "amp":
		return strings.Join(parts[1:], ".")
	default:
		return host
	}
}

func monitoringCanonicalQuery(parsed *url.URL, allowedKeys []string) string {
	if parsed == nil || len(allowedKeys) == 0 {
		return ""
	}
	values := parsed.Query()
	pairs := make([]string, 0, len(allowedKeys))
	for _, key := range allowedKeys {
		value := strings.TrimSpace(values.Get(key))
		if value == "" {
			continue
		}
		pairs = append(pairs, url.QueryEscape(key)+"="+url.QueryEscape(value))
	}
	sort.Strings(pairs)
	return strings.Join(pairs, "&")
}

func monitoringGenericContentQueryKeys() []string {
	return []string{"id", "article_id", "articleId", "docid", "docId", "postId"}
}

func monitoringNeteaseArticleID(rawURL string) string {
	if id := extractWangyihaoArticleIDFromURL(&rawURL); id != "" {
		return id
	}
	parsed, err := url.Parse(strings.TrimSpace(rawURL))
	if err != nil {
		return ""
	}
	segments := strings.Split(strings.Trim(parsed.Path, "/"), "/")
	for index := len(segments) - 1; index >= 0; index-- {
		id := strings.TrimSuffix(strings.TrimSpace(segments[index]), ".html")
		if monitoringLooksLikeNeteaseArticleID(id) {
			return id
		}
	}
	return ""
}

func monitoringLooksLikeNeteaseArticleID(value string) bool {
	if len(value) < 8 || len(value) > 32 {
		return false
	}
	hasDigit := false
	hasLetter := false
	for _, r := range value {
		switch {
		case r >= '0' && r <= '9':
			hasDigit = true
		case r >= 'A' && r <= 'Z':
			hasLetter = true
		case r >= 'a' && r <= 'z':
			hasLetter = true
		default:
			return false
		}
	}
	return hasDigit && hasLetter
}

func responseInternalAliasLookupError() error {
	return response.ErrInternal(50041, "alias_lookup_failed", "failed to resolve article aliases")
}

func responseInternalAliasScanError() error {
	return response.ErrInternal(50041, "alias_scan_failed", "failed to parse article alias resolution")
}