From ac183a90e8e392d1930a81ee3bf062ec5afcb3bb Mon Sep 17 00:00:00 2001
From: ItzCrazyKns <95534749+ItzCrazyKns@users.noreply.github.com>
Date: Thu, 18 Dec 2025 13:56:26 +0530
Subject: [PATCH] feat(academic-search): add academic search

---
 .../researcher/actions/academicSearch.ts      | 129 ++++++++++++++++++
 .../agents/search/researcher/actions/index.ts |   2 +
 2 files changed, 131 insertions(+)

diff --git a/src/lib/agents/search/researcher/actions/academicSearch.ts b/src/lib/agents/search/researcher/actions/academicSearch.ts
index e69de29..72e1f4b 100644
--- a/src/lib/agents/search/researcher/actions/academicSearch.ts
+++ b/src/lib/agents/search/researcher/actions/academicSearch.ts
@@ -0,0 +1,129 @@
+import z from 'zod';
+import { ResearchAction } from '../../types';
+import { Chunk, SearchResultsResearchBlock } from '@/lib/types';
+import { searchSearxng } from '@/lib/searxng';
+
+const schema = z.object({
+  queries: z.array(z.string()).describe('List of academic search queries'),
+});
+
+const academicSearchDescription = `
+Use this tool to perform academic searches for scholarly articles, papers, and research studies relevant to the user's query. Provide a list of concise search queries that will help gather comprehensive academic information on the topic at hand.
+You can provide up to 3 queries at a time. Make sure the queries are specific and relevant to the user's needs.
+
+For example, if the user is interested in recent advancements in renewable energy, your queries could be:
+1. "Recent advancements in renewable energy 2024"
+2. "Cutting-edge research on solar power technologies"
+3. "Innovations in wind energy systems"
+
+If this tool is present and no other tools are more relevant, you MUST use this tool to get the needed academic information.
+`;
+
+const academicSearchAction: ResearchAction<typeof schema> = {
+  name: 'academic_search',
+  schema: schema,
+  getDescription: () => academicSearchDescription,
+  getToolDescription: () =>
+    "Use this tool to perform academic searches for scholarly articles, papers, and research studies relevant to the user's query. Provide a list of concise search queries that will help gather comprehensive academic information on the topic at hand.",
+  enabled: (config) =>
+    config.sources.includes('academic') &&
+    config.classification.classification.skipSearch === false &&
+    config.classification.classification.academicSearch === true,
+  execute: async (input, additionalConfig) => {
+    input.queries = input.queries.slice(0, 3);
+
+    const researchBlock = additionalConfig.session.getBlock(
+      additionalConfig.researchBlockId,
+    );
+
+    if (researchBlock && researchBlock.type === 'research') {
+      researchBlock.data.subSteps.push({
+        type: 'searching',
+        id: crypto.randomUUID(),
+        searching: input.queries,
+      });
+
+      additionalConfig.session.updateBlock(additionalConfig.researchBlockId, [
+        {
+          op: 'replace',
+          path: '/data/subSteps',
+          value: researchBlock.data.subSteps,
+        },
+      ]);
+    }
+
+    const searchResultsBlockId = crypto.randomUUID();
+    let searchResultsEmitted = false;
+
+    let results: Chunk[] = [];
+
+    const search = async (q: string) => {
+      const res = await searchSearxng(q, {
+        engines: ['arxiv', 'google scholar', 'pubmed'],
+      });
+
+      const resultChunks: Chunk[] = res.results.map((r) => ({
+        content: r.content || r.title,
+        metadata: {
+          title: r.title,
+          url: r.url,
+        },
+      }));
+
+      results.push(...resultChunks);
+
+      if (
+        !searchResultsEmitted &&
+        researchBlock &&
+        researchBlock.type === 'research'
+      ) {
+        searchResultsEmitted = true;
+
+        researchBlock.data.subSteps.push({
+          id: searchResultsBlockId,
+          type: 'search_results',
+          reading: resultChunks,
+        });
+
+        additionalConfig.session.updateBlock(additionalConfig.researchBlockId, [
+          {
+            op: 'replace',
+            path: '/data/subSteps',
+            value: researchBlock.data.subSteps,
+          },
+        ]);
+      } else if (
+        searchResultsEmitted &&
+        researchBlock &&
+        researchBlock.type === 'research'
+      ) {
+        const subStepIndex = researchBlock.data.subSteps.findIndex(
+          (step) => step.id === searchResultsBlockId,
+        );
+
+        const subStep = researchBlock.data.subSteps[
+          subStepIndex
+        ] as SearchResultsResearchBlock;
+
+        subStep.reading.push(...resultChunks);
+
+        additionalConfig.session.updateBlock(additionalConfig.researchBlockId, [
+          {
+            op: 'replace',
+            path: '/data/subSteps',
+            value: researchBlock.data.subSteps,
+          },
+        ]);
+      }
+    };
+
+    await Promise.all(input.queries.map(search));
+
+    return {
+      type: 'search_results',
+      results,
+    };
+  },
+};
+
+export default academicSearchAction;
diff --git a/src/lib/agents/search/researcher/actions/index.ts b/src/lib/agents/search/researcher/actions/index.ts
index 6d7e58f..90559b9 100644
--- a/src/lib/agents/search/researcher/actions/index.ts
+++ b/src/lib/agents/search/researcher/actions/index.ts
@@ -1,3 +1,4 @@
+import academicSearchAction from './academicSearch';
 import doneAction from './done';
 import planAction from './plan';
 import ActionRegistry from './registry';
@@ -10,5 +11,6 @@ ActionRegistry.register(doneAction);
 ActionRegistry.register(planAction);
 ActionRegistry.register(scrapeURLAction);
 ActionRegistry.register(uploadsSearchAction);
+ActionRegistry.register(academicSearchAction);
 
 export { ActionRegistry };