web: support self-hosted Jina Reader

2026-09-15 02:56:10 +08:00 · 2026-08-04 23:59:28 +08:00
parent 7a0969dc4e
commit 1c47a69716
4 changed files with 175 additions and 8 deletions
@@ -177,6 +177,9 @@ object PluginConfig : AutoSavePluginConfig("Config") {
    @ValueDescription("Jina API Key")
    val jinaApiKey by value("")

+    @ValueDescription("Jina Reader API 地址，默认使用在线服务；自托管示例：http://127.0.0.1:4223/")
+    val jinaReaderUrl: String by value("https://r.jina.ai/")
+
    @ValueDescription("SearXNG 搜索引擎地址，如 http://127.0.0.1:8080/search 必须启用允许json格式返回")
    val searXngUrl: String by value("")

@@ -4,10 +4,17 @@ import com.aallam.openai.api.chat.Tool
 import com.aallam.openai.api.core.Parameters
 import io.ktor.client.request.*
 import io.ktor.client.statement.*
+import io.ktor.http.*
+import kotlinx.coroutines.CancellationException
+import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.async
 import kotlinx.coroutines.awaitAll
+import kotlinx.coroutines.withContext
 import kotlinx.serialization.json.*
 import top.jie65535.mirai.config.PluginConfig
+import java.net.InetAddress
+import java.net.URI
+import java.net.UnknownHostException

 class VisitWeb : BaseAgent(
    tool = Tool.function(
@@ -36,8 +43,69 @@ class VisitWeb : BaseAgent(
    )
 ) {
    companion object {
-        // Visit Tool (Using Jina Reader)
-        const val JINA_READER_URL_PREFIX = "https://r.jina.ai/"
+        internal data class ReaderRequest(
+            val endpoint: String,
+            val apiKey: String?,
+        )
+
+        internal fun createReaderRequest(
+            rawUrl: String,
+            readerBaseUrl: String,
+            apiKey: String,
+            resolveAddresses: (String) -> List<InetAddress> = {
+                InetAddress.getAllByName(it).toList()
+            },
+        ): ReaderRequest {
+            val target = parseHttpUrl(rawUrl, "网页地址")
+            require(target.userInfo == null) { "网页地址不能包含用户凭据" }
+
+            val targetHost = requireNotNull(target.host).removeSurrounding("[", "]").lowercase()
+            require(targetHost.contains('.') || targetHost.contains(':')) {
+                "禁止访问单标签或内部主机名"
+            }
+            require(
+                targetHost != "localhost" &&
+                        !targetHost.endsWith(".localhost") &&
+                        !targetHost.endsWith(".local") &&
+                        !targetHost.endsWith(".internal") &&
+                        !targetHost.endsWith(".lan") &&
+                        !targetHost.endsWith(".home.arpa")
+            ) { "禁止访问本机或局域网网页地址" }
+
+            val addresses = try {
+                resolveAddresses(targetHost)
+            } catch (_: UnknownHostException) {
+                // 目标可能只能由 Mihomo 的 DNS 解析，交给代理继续处理。
+                emptyList()
+            }
+            require(addresses.all(VisualImageResolver::isPublicAddress)) {
+                "网页地址解析到非公网地址，已拒绝访问"
+            }
+
+            val readerBase = parseHttpUrl(readerBaseUrl, "Jina Reader API 地址")
+            require(readerBase.query == null && readerBase.fragment == null) {
+                "Jina Reader API 地址不能包含查询参数或片段"
+            }
+
+            return ReaderRequest(
+                endpoint = readerBaseUrl.trim().trimEnd('/') + "/" +
+                        target.toASCIIString().substringBefore('#'),
+                apiKey = apiKey.trim().takeIf(String::isNotEmpty),
+            )
+        }
+
+        private fun parseHttpUrl(rawUrl: String, label: String): URI {
+            val uri = try {
+                URI(rawUrl.trim())
+            } catch (e: Exception) {
+                throw IllegalArgumentException("$label 格式无效", e)
+            }
+            require(uri.scheme?.lowercase() in setOf("http", "https")) {
+                "$label 仅支持 HTTP/HTTPS"
+            }
+            require(!uri.host.isNullOrBlank()) { "$label 缺少有效主机名" }
+            return uri.normalize()
+        }
    }

    override val isEnabled: Boolean
@@ -61,12 +129,26 @@ class VisitWeb : BaseAgent(

    private suspend fun jinaReadPage(url: String): String {
        return try {
-            httpClient.get(JINA_READER_URL_PREFIX + url) {
-                if (PluginConfig.jinaApiKey.isNotEmpty()) {
-                    header("Authorization", "Bearer ${PluginConfig.jinaApiKey}")
-                }
-            }.bodyAsText()
-        } catch (e: Throwable) {
+            val request = withContext(Dispatchers.IO) {
+                createReaderRequest(
+                    rawUrl = url,
+                    readerBaseUrl = PluginConfig.jinaReaderUrl,
+                    apiKey = PluginConfig.jinaApiKey,
+                )
+            }
+            val response = httpClient.get(request.endpoint) {
+                header(HttpHeaders.Accept, ContentType.Text.Plain)
+                request.apiKey?.let { header(HttpHeaders.Authorization, "Bearer $it") }
+            }
+            val body = response.bodyAsText()
+            if (response.status.isSuccess()) {
+                body
+            } else {
+                "Error fetching \"$url\": HTTP ${response.status.value} ${body.take(500)}"
+            }
+        } catch (e: CancellationException) {
+            throw e
+        } catch (e: Exception) {
            "Error fetching \"$url\": ${e.message}"
        }
    }
@@ -0,0 +1,80 @@
+package top.jie65535.mirai.tools
+
+import java.net.InetAddress
+import java.net.UnknownHostException
+import kotlin.test.Test
+import kotlin.test.assertEquals
+import kotlin.test.assertFailsWith
+
+class VisitWebTest {
+    @Test
+    fun `self hosted reader request uses configured base url`() {
+        val request = VisitWeb.createReaderRequest(
+            rawUrl = "https://example.com/docs?q=reader#section",
+            readerBaseUrl = "http://127.0.0.1:4223/",
+            apiKey = "",
+            resolveAddresses = { listOf(InetAddress.getByName("93.184.216.34")) },
+        )
+
+        assertEquals(
+            "http://127.0.0.1:4223/https://example.com/docs?q=reader",
+            request.endpoint,
+        )
+        assertEquals(null, request.apiKey)
+    }
+
+    @Test
+    fun `hosted reader remains backward compatible`() {
+        val request = VisitWeb.createReaderRequest(
+            rawUrl = "https://example.com/",
+            readerBaseUrl = "https://r.jina.ai/",
+            apiKey = "token",
+            resolveAddresses = { listOf(InetAddress.getByName("93.184.216.34")) },
+        )
+
+        assertEquals("https://r.jina.ai/https://example.com/", request.endpoint)
+        assertEquals("token", request.apiKey)
+    }
+
+    @Test
+    fun `private resolved targets are rejected`() {
+        assertFailsWith<IllegalArgumentException> {
+            VisitWeb.createReaderRequest(
+                rawUrl = "https://example.com/",
+                readerBaseUrl = "http://127.0.0.1:4223/",
+                apiKey = "",
+                resolveAddresses = { listOf(InetAddress.getByName("192.168.1.10")) },
+            )
+        }
+    }
+
+    @Test
+    fun `internal hostnames and credentials are rejected`() {
+        assertFailsWith<IllegalArgumentException> {
+            VisitWeb.createReaderRequest(
+                rawUrl = "http://redis/",
+                readerBaseUrl = "http://127.0.0.1:4223/",
+                apiKey = "",
+            )
+        }
+        assertFailsWith<IllegalArgumentException> {
+            VisitWeb.createReaderRequest(
+                rawUrl = "https://user:[email protected]/",
+                readerBaseUrl = "http://127.0.0.1:4223/",
+                apiKey = "",
+            )
+        }
+    }
+
+    @Test
+    fun `unresolved public host may be resolved by reader deployment`() {
+        val request = VisitWeb.createReaderRequest(
+            rawUrl = "https://blocked.example/",
+            readerBaseUrl = "http://127.0.0.1:4223/",
+            apiKey = "",
+            resolveAddresses = { throw UnknownHostException("proxy DNS only") },
+        )
+
+        assertEquals("http://127.0.0.1:4223/https://blocked.example/", request.endpoint)
+    }
+}