Teradata MCP Server

import logging import json from teradata_mcp_server.tools.utils import create_response # Define the logger. logger = logging.getLogger("teradata_mcp_server") def get_plot_json_data(conn, table_name, labels, columns, chart_type='line'): """ Helper function to fetch data from a Teradata table and formats it for plotting. Right now, designed only to support line plots from chart.js . """ # Define the colors first. colors = ['rgb(75, 192, 192)', '#99cbba', '#d7d0c4', '#fac778', '#e46c59', '#F9CB99', '#280A3E', '#F2EDD1', '#689B8A'] # Chart properties. Every chart needs different property for colors. chart_properties = {'line': 'borderColor', 'polar': 'backgroundColor', 'pie': 'backgroundColor'} columns = [columns] if isinstance(columns, str) else columns sql = "select {labels}, {columns} from {table_name} order by {labels}".format( labels=labels, columns=','.join(columns), table_name=table_name) # Prepare the statement. with conn.cursor() as cur: recs = cur.execute(sql).fetchall() # Define the structure of the chart data. Below is the structure expected by chart.js # { # labels: labels, # datasets: [{ # label: 'My First Dataset', # data: [65, 59, 80, 81, 56, 55, 40], # fill: false, # borderColor: 'rgb(75, 192, 192)', # tension: 0.1 # }] # } labels = [] datasets = [[] for _ in range(len(columns))] for rec in recs: labels.append(rec[0]) for i_, val in enumerate(rec[1:]): datasets[i_].append(val) # Prepare the datasets for chart.js datasets_ = [] for i, dataset in enumerate(datasets): datasets_.append({ 'label': columns[i], 'data': dataset, 'borderColor': colors[i], 'fill': False }) # For polar plot, every dataset needs different colors. if chart_type in ('polar', 'pie'): for i, dataset in enumerate(datasets_): # Remove borderColor and add backgroundColor dataset.pop('borderColor', None) dataset['backgroundColor'] = colors chart_data = {"labels": [str(l) for l in labels], "datasets": datasets_} logger.debug("Chart data: %s", json.dumps(chart_data, indent=2)) return create_response(data=chart_data, metadata={ "tool_description": "chart js {} plot data".format(chart_type), "table_name": table_name, "labels": labels, "columns": columns }) def get_radar_plot_json_data(conn, table_name, labels, columns): """ Helper function to fetch data from a Teradata table and formats it for plotting. Right now, designed only to support line plots from chart.js . """ logger.debug(f"Tool: get_json_data_for_plotting") # Define the colors first. border_colors = [ 'rgb(255, 99, 132)', 'rgb(54, 162, 235)', '#d7d0c4', '#fac778', '#e46c59', '#F9CB99', '#280A3E', '#F2EDD1', '#689B8A' ] background_colors = [ 'rgba(255, 99, 132, 0.2)', 'rgba(54, 162, 235, 0.2)', 'rgb(222, 232, 206, 0.2)', 'rgb(187, 102, 83, 0.2)', 'rgb(240, 139, 81, 0.2)', 'rgb(255, 248, 232, 0.2)' ] point_background_color = [ 'rgba(255, 99, 132)', 'rgba(54, 162, 235)', 'rgb(222, 232, 206)', 'rgb(187, 102, 83)', 'rgb(240, 139, 81)', 'rgb(255, 248, 232)' ] columns = [columns] if isinstance(columns, str) else columns sql = "select {labels}, {columns} from {table_name} order by {labels}".format( labels=labels, columns=','.join(columns), table_name=table_name) # Execute the SQL query # Prepare the statement. with conn.cursor() as cur: recs = cur.execute(sql).fetchall() labels = [] datasets = [[] for _ in range(len(columns))] for rec in recs: labels.append(rec[0]) for i_, val in enumerate(rec[1:]): datasets[i_].append(val) # Prepare the datasets for chart.js datasets_ = [] for i, dataset in enumerate(datasets): datasets_.append({ 'label': columns[i], 'data': dataset, 'fill': True, "backgroundColor": background_colors[i % len(background_colors)], 'borderColor': border_colors[i % len(border_colors)], "pointBackgroundColor": point_background_color[i % len(point_background_color)], "pointBorderColor": '#fff', "pointHoverBackgroundColor": '#fff', "pointHoverBorderColor": point_background_color[i % len(point_background_color)] }) chart_data = {"labels": [str(l) for l in labels], "datasets": datasets_} logger.debug("Chart data: %s", json.dumps(chart_data, indent=2)) return create_response(data=chart_data, metadata={ "tool_description": "chart js radar plot data", "table_name": table_name, "labels": labels, "columns": columns })