Dataset Viewer MCP Server

Instructions

Implementation Reference

• src/dataset_viewer/server.py:507-517 (handler)

  MCP tool handler execution logic for 'get_first_rows': extracts arguments, instantiates DatasetViewerAPI, calls get_first_rows method, and returns JSON-formatted result as TextContent.

  elif name == "get_first_rows":
      dataset = arguments["dataset"]
      config = arguments["config"]
      split = arguments["split"]
      first_rows = await DatasetViewerAPI(auth_token=auth_token).get_first_rows(dataset, config=config, split=split)
      return [
          types.TextContent(
              type="text",
              text=json.dumps(first_rows, indent=2)
          )
      ]

• src/dataset_viewer/server.py:274-304 (registration)

  Tool registration in @server.list_tools(), defining name, description, and input schema for 'get_first_rows'.

  types.Tool(
      name="get_first_rows",
      description="Get first rows from a Hugging Face dataset split",
      inputSchema={
          "type": "object",
          "properties": {
              "dataset": {
                  "type": "string",
                  "description": "Hugging Face dataset identifier in the format owner/dataset",
                  "pattern": "^[^/]+/[^/]+$",
                  "examples": ["ylecun/mnist", "stanfordnlp/imdb"]
              },
              "config": {
                  "type": "string",
                  "description": "Dataset configuration/subset name. Use get_info to list available configs",
                  "examples": ["default", "en", "es"]
              },
              "split": {
                  "type": "string",
                  "description": "Dataset split name. Splits partition the data for training/evaluation",
                  "examples": ["train", "validation", "test"]
              },
              "auth_token": {
                  "type": "string",
                  "description": "Hugging Face auth token for private/gated datasets",
                  "optional": True
              }
          },
          "required": ["dataset", "config", "split"],
      }
  ),

• src/dataset_viewer/server.py:93-102 (helper)

  Core helper function in DatasetViewerAPI class that makes HTTP GET request to '/first-rows' endpoint of Hugging Face datasets-server to retrieve first rows of specified dataset/config/split.

  async def get_first_rows(self, dataset: str, config: str, split: str) -> dict:
      """Get first few rows of a dataset split"""
      params = {
          "dataset": dataset,
          "config": config,
          "split": split
      }
      response = await self.client.get("/first-rows", params=params)
      response.raise_for_status()
      return response.json()