refactor: improve seed source handling with additional type support, enhanced parsing logic, and text chunking optimization

studio/frontend/src/features/recipe-studio/dialogs/seed/seed-dialog.tsx

@@ -33,6 +33,7 @@ import {
   TabsList,
   TabsTrigger,
 } from "@/components/ui/tabs";
+import { RecursiveCharacterTextSplitter } from "@langchain/textsplitters";
 import mammoth from "mammoth";
 import { type ReactElement, useEffect, useMemo, useState } from "react";
 import { extractText, getDocumentProxy } from "unpdf";
@@ -60,6 +61,10 @@ const UNSTRUCTURED_ACCEPT = ".txt,.pdf,.docx";
 const MAX_UPLOAD_BYTES = 50 * 1024 * 1024;
 const CHUNK_SIZE = 1200;
 const CHUNK_OVERLAP = 200;
+const UNSTRUCTURED_SPLITTER = new RecursiveCharacterTextSplitter({
+  chunkSize: CHUNK_SIZE,
+  chunkOverlap: CHUNK_OVERLAP,
+});
 
 type SeedDialogProps = {
   config: SeedConfig;
@@ -84,20 +89,12 @@ function stringifyCell(value: unknown): string {
   }
 }
 
-function chunkText(input: string): string[] {
+async function chunkText(input: string): Promise<string[]> {
   const text = input.replace(/\r\n/g, "\n").replace(/\n{3,}/g, "\n\n").trim();
   if (!text) return [];
 
-  const chunks: string[] = [];
-  let cursor = 0;
-  const overlap = Math.max(0, Math.min(CHUNK_OVERLAP, CHUNK_SIZE - 1));
-  while (cursor < text.length) {
-    const end = Math.min(cursor + CHUNK_SIZE, text.length);
-    chunks.push(text.slice(cursor, end).trim());
-    if (end >= text.length) break;
-    cursor = Math.max(0, end - overlap);
-  }
-  return chunks.filter(Boolean);
+  const chunks = await UNSTRUCTURED_SPLITTER.splitText(text);
+  return chunks.map((chunk) => chunk.trim()).filter(Boolean);
 }
 
 async function fileToBase64Payload(file: File): Promise<string> {
@@ -218,7 +215,7 @@ export function SeedDialog({ config, onUpdate }: SeedDialogProps): ReactElement
       }
 
       const text = await extractUnstructuredText(unstructuredFile);
-      const chunks = chunkText(text);
+      const chunks = await chunkText(text);
       if (chunks.length === 0) {
         throw new Error("No text found in file.");
       }

studio/frontend/src/features/recipe-studio/stores/recipe-studio.ts

@@ -300,12 +300,12 @@ export const useRecipeStudioStore = create<RecipeStudioState>((set, get) => ({
       if (!existing) {
         return buildAddedNodeState(state, "seed", type);
       }
-      const nextSourceType: SeedSourceType =
-        type === "seed_local"
-          ? "local"
-          : type === "seed_unstructured"
-            ? "unstructured"
-            : "hf";
+      let nextSourceType: SeedSourceType = "hf";
+      if (type === "seed_local") {
+        nextSourceType = "local";
+      } else if (type === "seed_unstructured") {
+        nextSourceType = "unstructured";
+      }
 
       const nextConfig = {
         ...existing,

studio/frontend/src/features/recipe-studio/utils/import/importer.ts

@@ -4,6 +4,7 @@ import type {
   LlmToolConfig,
   NodeConfig,
   RecipeProcessorConfig,
+  SeedSourceType,
 } from "../../types";
 import { buildEdges } from "./edges";
 import { isRecord, parseJson, readString } from "./helpers";
@@ -29,6 +30,7 @@ type RecipeInput = {
 type UiInput = {
   nodes?: unknown;
   edges?: unknown;
+  seed_source_type?: unknown;
 };
 
 function parseProcessors(input: unknown): RecipeProcessorConfig[] {
@@ -219,11 +221,18 @@ export function importRecipePayload(input: string): ImportResult {
   const nameToId = new Map<string, string>();
 
   let nextId = 1;
+  const uiSeedSourceTypeRaw = readString(ui?.seed_source_type);
+  const uiSeedSourceType: SeedSourceType | undefined =
+    uiSeedSourceTypeRaw === "hf" ||
+    uiSeedSourceTypeRaw === "local" ||
+    uiSeedSourceTypeRaw === "unstructured"
+      ? uiSeedSourceTypeRaw
+      : undefined;
 
   if (recipe.seed_config) {
     const id = `n${nextId}`;
     nextId += 1;
-    const seedConfig = parseSeedConfig(recipe.seed_config, id);
+    const seedConfig = parseSeedConfig(recipe.seed_config, id, uiSeedSourceType);
     if (seedConfig) {
       configs.push(seedConfig);
     }

studio/frontend/src/features/recipe-studio/utils/import/parsers/seed-config-parser.ts

@@ -65,7 +65,7 @@ function parseSeedSettings(seedConfigRaw: unknown): Partial<SeedConfig> {
   let hf_endpoint = "https://huggingface.co";
   let hf_repo_id = "";
   let local_file_name = "";
-  let unstructured_file_name = "";
+  const unstructured_file_name = "";
   const sourceRaw = seedConfigRaw.source;
   if (isRecord(sourceRaw)) {
     const seedType = readString(sourceRaw.seed_type);
@@ -127,12 +127,23 @@ function parseSeedSettings(seedConfigRaw: unknown): Partial<SeedConfig> {
 export function parseSeedConfig(
   seedConfigRaw: unknown,
   id: string,
+  preferredSourceType?: SeedSourceType,
 ): SeedConfig | null {
   if (!seedConfigRaw) {
     return null;
   }
+  const parsed = parseSeedSettings(seedConfigRaw);
+  let sourceType: SeedSourceType = "hf";
+  if (parsed.seed_source_type === "hf") {
+    sourceType = "hf";
+  } else if (preferredSourceType) {
+    sourceType = preferredSourceType;
+  } else if (parsed.seed_source_type) {
+    sourceType = parsed.seed_source_type;
+  }
   return {
     ...makeDefaultSeedConfig(id),
-    ...parseSeedSettings(seedConfigRaw), // payload-only fields override ui defaults
+    ...parsed, // payload-only fields override ui defaults
+    seed_source_type: sourceType,
   };
 }

studio/frontend/src/features/recipe-studio/utils/payload/build-payload.ts

@@ -241,6 +241,7 @@ export function buildRecipePayload(
       ui: {
         nodes: uiNodes,
         edges: uiEdges,
+        ...(firstSeed && { seed_source_type: firstSeed.seed_source_type }),
       },
     },
   };

studio/frontend/src/features/recipe-studio/utils/payload/types.ts

@@ -30,6 +30,8 @@ export type RecipePayload = {
   ui: {
     nodes: { id: string; x: number; y: number }[];
     edges: { from: string; to: string; type?: string }[];
+    // ui-only, used to preserve seed block mode across imports/refresh
+    seed_source_type?: "hf" | "local" | "unstructured";
   };
 };