fix(websearch): 修复 UUID 生成逻辑，确保唯一性；更新 API Base URL 错误提示信息；新增消息引用缓存机制

Author: piexian

astrbot/builtin_stars/web_searcher/main.py

@@ -396,7 +396,7 @@ async def search_from_tavily(
             return "Error: Tavily web searcher does not return any results."
 
         ret_ls = []
-        ref_uuid = str(uuid.uuid4())[:4]
+        ref_uuid = str(uuid.uuid4())
         for idx, result in enumerate(results, 1):
             index = f"{ref_uuid}.{idx}"
             ret_ls.append(
@@ -425,7 +425,7 @@ async def tavily_extract_web_page(
         """Extract the content of a web page using Tavily.
 
         Args:
-            url(string): Required. An URl to extract content from.
+            url(string): Required. A URL to extract content from.
             extract_depth(string): Optional. The depth of the extraction, must be one of 'basic', 'advanced'. Default is "basic".
             timeout(number): Optional. Request timeout in seconds. Minimum is 30. Default is 30.
 
@@ -595,7 +595,7 @@ async def search_from_bocha(
             return "Error: BoCha web searcher does not return any results."
 
         ret_ls = []
-        ref_uuid = str(uuid.uuid4())[:4]
+        ref_uuid = str(uuid.uuid4())
         for idx, result in enumerate(results, 1):
             index = f"{ref_uuid}.{idx}"
             ret_ls.append(
@@ -717,7 +717,7 @@ async def search_from_exa(
             return "Error: Exa web searcher does not return any results."
 
         ret_ls = []
-        ref_uuid = str(uuid.uuid4())[:4]
+        ref_uuid = str(uuid.uuid4())
         for idx, result in enumerate(results, 1):
             index = f"{ref_uuid}.{idx}"
             ret_ls.append(
@@ -869,7 +869,7 @@ async def find_similar_links(
             return "Error: Exa find similar does not return any results."
 
         ret_ls = []
-        ref_uuid = str(uuid.uuid4())[:4]
+        ref_uuid = str(uuid.uuid4())
         for idx, result in enumerate(results, 1):
             index = f"{ref_uuid}.{idx}"
             ret_ls.append(

astrbot/core/utils/web_search_utils.py

@@ -25,7 +25,9 @@ def normalize_web_search_base_url(
     parsed = urlparse(normalized)
     if parsed.scheme not in {"http", "https"} or not parsed.netloc:
         raise ValueError(
-            f"Error: {provider_name} API Base URL must start with http:// or https://.",
+            f"Error: {provider_name} API Base URL must be a base host URL starting "
+            f"with http:// or https:// (for example, {default}), not a full endpoint "
+            f"path. Received: {normalized!r}.",
         )
     return normalized
 

dashboard/src/components/chat/MessageList.vue

@@ -274,7 +274,8 @@ export default {
                 url: ''
             },
             // Web search results mapping: { 'uuid.idx': { url, title, snippet } }
-            webSearchResults: {}
+            webSearchResults: {},
+            messageRefsCache: new WeakMap()
         };
     },
     async mounted() {
@@ -359,13 +360,66 @@ export default {
             return refs;
         },
 
+        buildMessageRefsCacheKey(messageParts) {
+            if (!Array.isArray(messageParts)) {
+                return '';
+            }
+
+            const cacheParts = [];
+
+            messageParts.forEach(part => {
+                if (part.type !== 'tool_call' || !Array.isArray(part.tool_calls)) {
+                    return;
+                }
+
+                part.tool_calls.forEach(toolCall => {
+                    if (!WEB_SEARCH_REFERENCE_TOOLS.includes(toolCall?.name) || !toolCall.result) {
+                        return;
+                    }
+
+                    const rawResult = typeof toolCall.result === 'string'
+                        ? toolCall.result
+                        : JSON.stringify(toolCall.result);
+
+                    cacheParts.push(`${toolCall.id || toolCall.name}:${rawResult}`);
+                });
+            });
+
+            return cacheParts.join('||');
+        },
+
+        getCachedMessageRefs(content) {
+            if (!content || typeof content !== 'object') {
+                return null;
+            }
+
+            const cacheKey = this.buildMessageRefsCacheKey(content.message);
+            if (!cacheKey) {
+                return null;
+            }
+
+            const cachedEntry = this.messageRefsCache.get(content);
+            if (cachedEntry?.key === cacheKey) {
+                return cachedEntry.refs;
+            }
+
+            const refs = this.collectMessageWebSearchRefs(content.message);
+            const normalizedRefs = refs.length ? { used: refs } : null;
+
+            this.messageRefsCache.set(content, {
+                key: cacheKey,
+                refs: normalizedRefs
+            });
+
+            return normalizedRefs;
+        },
+
         getMessageRefs(content) {
             if (content?.refs?.used?.length) {
                 return content.refs;
             }
 
-            const fallbackRefs = this.collectMessageWebSearchRefs(content?.message);
-            return fallbackRefs.length ? { used: fallbackRefs } : null;
+            return this.getCachedMessageRefs(content);
         },
 
         // 从消息中提取网页搜索结果映射
@@ -377,7 +431,12 @@ export default {
                     return;
                 }
 
-                this.collectMessageWebSearchRefs(msg.content.message).forEach(ref => {
+                const refs = this.getMessageRefs(msg.content);
+                if (!refs?.used?.length) {
+                    return;
+                }
+
+                refs.used.forEach(ref => {
                     results[ref.index] = {
                         url: ref.url,
                         title: ref.title,

tests/unit/test_web_search_utils.py

@@ -1,9 +1,12 @@
 import json
 
+import pytest
+
 from astrbot.core.utils.web_search_utils import (
     build_web_search_refs,
     collect_web_search_ref_items,
     collect_web_search_results,
+    normalize_web_search_base_url,
 )
 
 
@@ -88,3 +91,18 @@ def test_build_web_search_refs_ignores_tool_call_id_and_falls_back():
     )
 
     assert [ref["index"] for ref in refs["used"]] == ["a152.1", "a152.2"]
+
+
+def test_normalize_web_search_base_url_reports_invalid_value():
+    with pytest.raises(ValueError) as exc_info:
+        normalize_web_search_base_url(
+            "exa.ai/search",
+            default="https://api.exa.ai",
+            provider_name="Exa",
+        )
+
+    assert str(exc_info.value) == (
+        "Error: Exa API Base URL must be a base host URL starting with "
+        "http:// or https:// (for example, https://api.exa.ai), not a full "
+        "endpoint path. Received: 'exa.ai/search'."
+    )