datarobot-oss/datarobot-agent-skills / skills/datarobot-predictions
datarobot-predictions
VerifiedTools and guidance for making predictions with DataRobot deployments, including real-time predictions, batch scoring, prediction dataset generation, and prediction explanations (SHAP/XEMP). Use when making predictions, running batch scoring, generating prediction datasets, or explaining individual predictions from a deployment.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 14.8 KB |
| scripts/generate_prediction_data_template.pyexecutable | script | 7.6 KB |
| scripts/get_deployment_features.pyexecutable | script | 4.2 KB |
| scripts/make_prediction.pyexecutable | script | 4.8 KB |
| scripts/validate_prediction_data.pyexecutable | script | 4.6 KB |
scripts/validate_prediction_data.py
View on GitHub#!/usr/bin/env python3
# Copyright (c) 2026 DataRobot, Inc. All rights reserved.
# SPDX-License-Identifier: Apache-2.0
"""
Validate prediction data for a deployment.
Usage:
python validate_prediction_data.py <deployment_id> <file_path>
Returns validation report with errors, warnings, and info messages.
"""
import csv
import json
import os
import sys
import datarobot as dr
def validate_prediction_data(deployment_id: str, file_path: str) -> dict:
"""
Validate prediction data for a deployment.
Args:
deployment_id: The deployment ID
file_path: Path to CSV file to validate
Returns:
Validation report with errors, warnings, and info
"""
# Initialize client
dr.Client()
deployment = dr.Deployment.get(deployment_id)
# List of plain dicts with keys name, feature_type, importance, date_format,
# known_in_advance
try:
features = deployment.get_features()
except dr.errors.ClientError as e:
return {
"valid": False,
"errors": [f"Deployment feature metadata unavailable ({e})"],
}
target_name = deployment.model.get(
"target_name"
) # None for unsupervised/anomaly deployments
# Get required features (exclude target)
required_features = {
f["name"]: f["feature_type"] for f in features if f["name"] != target_name
}
# Read CSV data
if not os.path.exists(file_path):
return {"valid": False, "errors": [f"File not found: {file_path}"]}
with open(file_path, "r") as f:
# Skip '#' metadata lines (generate_prediction_data_template.py writes them)
reader = csv.DictReader(line for line in f if not line.lstrip().startswith("#"))
rows = list(reader)
if not rows:
return {"valid": False, "errors": ["CSV file is empty"]}
# Validate
errors = []
warnings = []
info = []
# Check for missing required features (ragged rows park overflow under None)
csv_columns = set(rows[0].keys())
csv_columns.discard(None)
missing_features = set(required_features.keys()) - csv_columns
if missing_features:
errors.append(
f"Missing required features: {', '.join(sorted(missing_features))}"
)
# Check for extra columns
extra_columns = csv_columns - set(required_features.keys())
if extra_columns:
info.append(
f"Extra columns (will be ignored): {', '.join(sorted(extra_columns))}"
)
# Check data types (simplified - actual validation would be more thorough)
for row_num, row in enumerate(rows, start=2): # Start at 2 (header is row 1)
for feature_name, feature_type in required_features.items():
if feature_name in row:
value = row[feature_name]
if value == "":
warnings.append(f"Row {row_num}, {feature_name}: Empty value")
elif feature_type == "Numeric" and value:
try:
float(value)
except ValueError:
errors.append(
f"Row {row_num}, {feature_name}: Expected numeric, got '{value}'"
)
# Use the importance already returned by deployment.get_features() for warnings
# (a model-independent measure of feature/target relationship strength; direction
# is not encoded — negative values mean no detectable relationship, so the signed
# comparison below is intentional: don't take abs()). None means importance is
# unknown for this model class (unsupervised, time-series, custom) — not low.
low_importance_features = [
f["name"]
for f in features
if f["importance"] is not None and f["importance"] < 0.05
]
missing_low_importance = missing_features & set(low_importance_features)
if missing_low_importance:
warnings.append(
f"Missing low-importance features: {', '.join(sorted(missing_low_importance))}"
)
return {
"valid": len(errors) == 0,
"deployment_id": deployment_id,
"row_count": len(rows),
"errors": errors,
"warnings": warnings,
"info": info,
"required_features": sorted(required_features.keys()),
"provided_features": sorted(csv_columns),
}
if __name__ == "__main__":
if len(sys.argv) < 3:
print(
"Usage: python validate_prediction_data.py <deployment_id> <file_path>",
file=sys.stderr,
)
sys.exit(1)
deployment_id = sys.argv[1]
file_path = sys.argv[2]
result = validate_prediction_data(deployment_id, file_path)
print(json.dumps(result, indent=2))
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |