@vespermcp/mcp-server 1.1.2 → 1.1.3

package/build/ingestion/hf-downloader.js

@@ -1,4 +1,5 @@
 import { listFiles } from "@huggingface/hub";
+import path from "path";
 import { RobustDownloader } from "../utils/downloader.js";
 export class HFDownloader {
     hfToken;
@@ -14,13 +15,26 @@ export class HFDownloader {
     async findBestFile(repoId) {
         try {
             const files = [];
+            const blacklist = [
+                ".gitattributes",
+                ".gitignore",
+                ".git",
+                "README.md",
+                "LICENSE",
+                "package.json",
+                "requirements.txt",
+                "setup.py"
+            ];
             for await (const file of listFiles({
                 repo: { type: "dataset", name: repoId },
                 recursive: true,
                 ...(this.hfToken ? { accessToken: this.hfToken } : {})
             })) {
                 if (file.type === "file") {
-                    files.push(file.path);
+                    const fileName = path.basename(file.path);
+                    if (!blacklist.includes(fileName) && !fileName.startsWith(".")) {
+                        files.push(file.path);
+                    }
                 }
             }
             // Priority logic for data scientists
@@ -39,7 +53,13 @@ export class HFDownloader {
                 if (match)
                     return match;
             }
-            return files.length > 0 ? files[0] : null;
+            // Strict fallback: Only return the first file if it has a data-like extension
+            const dataExtensions = [".csv", ".parquet", ".jsonl", ".json", ".txt", ".tsv", ".avro", ".orc"];
+            const fallback = files.find(f => {
+                const ext = path.extname(f).toLowerCase();
+                return dataExtensions.includes(ext);
+            });
+            return fallback || null;
         }
         catch (error) {
             console.error(`[HF] Failed to list files for ${repoId}:`, error.message);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@vespermcp/mcp-server",
-  "version": "1.1.2",
+  "version": "1.1.3",
   "description": "AI-powered dataset discovery, quality analysis, and preparation MCP server with multimodal support (text, image, audio, video)",
   "type": "module",
   "main": "build/index.js",