data_median.py•1.8 kB
import json
from typing import Any
from mcp import types
from pydantic import ConfigDict
from ..make_logger import make_logger
from .model import Data
logger = make_logger(__name__)
class DataMedianInputSchema(Data):
model_config = ConfigDict(
validate_assignment=True,
frozen=True,
extra="forbid",
arbitrary_types_allowed=True,
)
@staticmethod
def input_schema() -> dict:
return {
"type": "object",
"properties": {
"input_data_file_path": {
"type": "string",
"description": "Path to the input data file",
},
},
}
@staticmethod
def from_schema(input_data_file_path: str) -> "DataMedianInputSchema":
data = Data.from_file(input_data_file_path)
return DataMedianInputSchema(df=data.df)
@staticmethod
def from_args(arguments: dict[str, Any]) -> "DataMedianInputSchema":
input_data_file_path = arguments["input_data_file_path"]
return DataMedianInputSchema.from_schema(input_data_file_path=input_data_file_path)
async def handle_data_median(
arguments: dict[str, Any],
) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:
data_median_input = DataMedianInputSchema.from_args(arguments)
median_df = data_median_input.df.median()
# Convert the DataFrame to a dictionary format
median_dict = {
"description": "Median values for each column",
"median_values": {
col: str(val) if val is not None else None for col, val in zip(median_df.columns, median_df.row(0))
},
}
return [
types.TextContent(
type="text",
text=json.dumps(median_dict),
)
]