@vespermcp/mcp-server 1.2.28 → 1.2.29

package/build/index.js

@@ -339,6 +339,36 @@ function ensureLocalPipelineSource(sourcePath, datasetId, targetDir) {
     }
     return stagedPath;
 }
+function cleanupIntermediateArtifacts(artifactPaths, finalOutputPath) {
+    const finalResolved = path.resolve(finalOutputPath);
+    const finalLineage = `${finalResolved}.lineage.json`;
+    for (const candidate of artifactPaths) {
+        if (!candidate)
+            continue;
+        const resolved = path.resolve(candidate);
+        if (resolved === finalResolved || resolved === finalLineage)
+            continue;
+        try {
+            if (fs.existsSync(resolved) && fs.statSync(resolved).isFile()) {
+                fs.unlinkSync(resolved);
+            }
+        }
+        catch {
+            // Best-effort cleanup.
+        }
+        const sidecar = `${resolved}.lineage.json`;
+        if (sidecar === finalLineage)
+            continue;
+        try {
+            if (fs.existsSync(sidecar) && fs.statSync(sidecar).isFile()) {
+                fs.unlinkSync(sidecar);
+            }
+        }
+        catch {
+            // Best-effort cleanup.
+        }
+    }
+}
 function resolveDatasetLocalPath(datasetIdOrPath, preferredDirs = []) {
     if (fs.existsSync(datasetIdOrPath)) {
         return ensureExportableLocalPath(datasetIdOrPath);
@@ -2898,6 +2928,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
             case "export_dataset": {
                 const datasetId = String(request.params.arguments?.dataset_id);
                 const isDirectLocalInput = isDirectLocalDatasetReference(datasetId);
+                const intermediateArtifacts = new Set();
                 const requestedTargetDir = request.params.arguments?.target_dir
                     ? String(request.params.arguments?.target_dir).trim()
                     : request.params.arguments?.output_dir
@@ -2967,9 +2998,16 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
                     else if (currentExt !== pipelineFmt) {
                         console.error(`[Export] Running quality/cleaning pipeline (use fast=true to skip)...`);
                         try {
+                            const beforeStagingPath = sourcePath;
                             sourcePath = ensureLocalPipelineSource(sourcePath, datasetId, targetDir);
+                            if (path.resolve(beforeStagingPath) !== path.resolve(sourcePath)) {
+                                intermediateArtifacts.add(sourcePath);
+                            }
                             const pipelineResult = await pipelineExecutor.runPipeline(datasetId, sourcePath, pipelineFmt);
                             if (pipelineResult.final_output_path) {
+                                if (path.resolve(pipelineResult.final_output_path) !== path.resolve(sourcePath)) {
+                                    intermediateArtifacts.add(pipelineResult.final_output_path);
+                                }
                                 sourcePath = pipelineResult.final_output_path;
                                 try {
                                     // Update registry to point to pipeline's final output
@@ -3058,6 +3096,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
                         msg += `  Python: \`pd.read_parquet('${result.output_path}').head()\`\n`;
                         msg += `  DuckDB: \`SELECT * FROM '${result.output_path}' LIMIT 50;\`\n`;
                     }
+                    cleanupIntermediateArtifacts(intermediateArtifacts, result.output_path);
                     return { content: [{ type: "text", text: msg }] };
                 }
                 catch (error) {
@@ -3933,6 +3972,7 @@ async function runExportCli(args) {
     const fastMode = args.includes("--fast");
     const preview = args.includes("--preview");
     const isDirectLocalInput = isDirectLocalDatasetReference(datasetId);
+    const intermediateArtifacts = new Set();
     const resolvedTargetDir = path.resolve(targetDir || process.cwd());
     let sourcePath = resolveDatasetLocalPath(datasetId, [resolvedTargetDir, process.cwd()]);
     if (!sourcePath) {
@@ -3954,9 +3994,16 @@ async function runExportCli(args) {
         const pipelineCompatibleInput = currentExt === "csv" || currentExt === "parquet";
         if (pipelineCompatibleInput && currentExt !== pipelineFmt) {
             try {
+                const beforeStagingPath = sourcePath;
                 sourcePath = ensureLocalPipelineSource(sourcePath, datasetId, resolvedTargetDir);
+                if (path.resolve(beforeStagingPath) !== path.resolve(sourcePath)) {
+                    intermediateArtifacts.add(sourcePath);
+                }
                 const pipelineResult = await pipelineExecutor.runPipeline(datasetId, sourcePath, pipelineFmt);
                 if (pipelineResult.final_output_path) {
+                    if (path.resolve(pipelineResult.final_output_path) !== path.resolve(sourcePath)) {
+                        intermediateArtifacts.add(pipelineResult.final_output_path);
+                    }
                     sourcePath = pipelineResult.final_output_path;
                     if (!isDirectLocalInput && shouldTrackExportPath(sourcePath)) {
                         upsertRegistry(datasetId, sourcePath, "completed");
@@ -3987,6 +4034,7 @@ async function runExportCli(args) {
     console.error(`[Export] Resolved output directory: ${outDir}`);
     console.error(`[Export] Output file: ${outputFile}`);
     const result = await dataExporter.export(sourcePath, outputFile, requestedFormat, exportOpts);
+    cleanupIntermediateArtifacts(intermediateArtifacts, result.output_path);
     console.log(`Export complete: ${result.output_path}`);
     console.log(`Format: ${result.format}${result.compression ? ` (${result.compression})` : ""}`);
     if (result.rows !== undefined)

package/build/python/cleaner.py

@@ -182,6 +182,8 @@ def main():
                 output_format = "parquet"
 
         base_name = file_path.rsplit(".", 1)[0]
+        if base_name.endswith("_cleaned"):
+            base_name = base_name[:-8]
         if output_format == "csv":
             output_path = f"{base_name}_cleaned.csv"
             # Stringify ANY column that might not be CSV-safe (List, Struct, Object, etc.)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@vespermcp/mcp-server",
-  "version": "1.2.28",
+  "version": "1.2.29",
   "description": "AI-powered dataset discovery, quality analysis, and preparation MCP server with multimodal support (text, image, audio, video)",
   "type": "module",
   "main": "build/index.js",

package/src/python/cleaner.py

@@ -182,6 +182,8 @@ def main():
                 output_format = "parquet"
 
         base_name = file_path.rsplit(".", 1)[0]
+        if base_name.endswith("_cleaned"):
+            base_name = base_name[:-8]
         if output_format == "csv":
             output_path = f"{base_name}_cleaned.csv"
             # Stringify ANY column that might not be CSV-safe (List, Struct, Object, etc.)