Quick Install
import datapulse # Initialize client with API key client = datapulse.Client(api_key="your_api_key_here") # Fetch predictive analytics for revenue forecasting response = client.analytics.predict( model="revenue_forecast_v2", dataset="q3_sales_data", horizon_days=30 ) print(response.json())
import { DataPulse } from "@datapulse/sdk"; const client = new DataPulse({ apiKey: process.env.DP_API_KEY }); // Execute customer segmentation pipeline const segments = await client.models.cluster({ source: "customer_behavior_logs", algorithm: "k-means++", k: 5 }); console.log(segments.results);
import com.datapulse.sdk.DataPulseClient; import com.datapulse.sdk.models.PredictionRequest; DataPulseClient client = DataPulseClient.builder() .apiKey("your_api_key_here") .build(); PredictionRequest req = PredictionRequest.builder() .model("churn_prediction_v3") .features("user_activity_stream") .build(); var result = client.analytics.predict(req);
package main import ( "fmt" "github.com/datapulse/sdk-go" ) func main() { client := datapulse.NewClient( datapulse.WithAPIKey("your_api_key_here"), datapulse.WithRegion("us-east-1"), ) resp, err := client.Predict(context.Background(), &datapulse.PredictOpts{ Model: "demand_forecasting", Payload: []byte(`{"features": [0.8, 0.9, 0.7]}`), }) if err != nil { panic(err) } fmt.Println(string(resp.Data)) }
Type-Safe & Reliable
Fully typed interfaces, comprehensive error handling, and automatic retry logic with exponential backoff built into every SDK.
Async & Streaming
Native support for async/await, websockets, and server-sent events for real-time analytics and model inference pipelines.
Observability Ready
OpenTelemetry integration, structured logging, and request tracing out of the box. Monitor every API call in your stack.
Enterprise Security
OAuth 2.0, API key rotation, RBAC enforcement, and SOC 2 Type II compliant authentication flows baked in.
Recent Releases
- Python: Added async context manager support for long-running inference jobs
- Node.js: Fixed memory leak in WebSocket event listeners
- All: Updated OpenTelemetry trace propagation headers
- Performance optimization: 40% faster JSON deserialization in Java & Go clients
- Added support for custom HTTP headers in retry configurations
- Documentation updates for streaming prediction endpoints
- Major: Introduced unified `DatapulseClient` interface across all languages
- New `models.cluster()` and `models.classify()` convenience methods
- Dropped support for Python 3.7 and Node.js 14
Frequently Asked Questions
We officially support Python (3.8+), JavaScript/Node.js (18+), Java (11+), Go (1.19+), and C# (.NET 6+). Community-maintained wrappers exist for Ruby, PHP, and Rust, but are not covered by enterprise SLAs.
Yes. All official SDKs are released under the MIT License and are available on GitHub under the `DataPulse-Org` organization. Pull requests and issue reports are welcome.
All SDKs include an exponential backoff retry mechanism that activates on 429 and 5xx responses. You can configure max retries, base delay, and jitter via the client options object or environment variables.
Yes. Enterprise customers can deploy our SDK-compatible inference endpoints within your VPC or air-gapped environments. Contact sales for architecture guidance and licensing.