Skip to main content
After processing your documents, FlexOrch lets you group completed jobs into a named dataset and export the structured output in the format your downstream pipeline expects. This page covers building, profiling, exporting, and deleting datasets using the TypeScript SDK.

​
Build a dataset

Pass an array of completed job IDs and a human-readable name to client.datasets.build(). The method returns a Promise<Dataset> that resolves once the dataset is assembled. 
import { FlexOrch } from "@flexorch/sdk";

const client = new FlexOrch({
  apiKey: process.env.FLEXORCH_API_KEY,
});

const dataset = await client.datasets.build({
  jobIds: ["job_abc123", "job_def456", "job_ghi789"],
  name: "q1-invoices",
});

console.log(`Dataset ID:  ${dataset.id}`);
console.log(`Name:        ${dataset.name}`);
console.log(`Status:      ${dataset.status}`);
console.log(`Job count:   ${dataset.jobCount}`);
All job IDs passed to build() must be in the completed state. Jobs with status queued, running, or failed are silently skipped unless you pass strict: true, which causes the Promise to reject with an IncompleteJobError.

​
Build from filtered jobs

A common pattern is to filter jobs by quality grade before building: 
const allJobs = await client.jobs.list({ limit: 200 });

const highQualityIds = allJobs
  .filter((j) => ["A", "B"].includes(j.qualityGrade ?? "") && j.status === "completed")
  .map((j) => j.id);

const dataset = await client.datasets.build({
  jobIds: highQualityIds,
  name: "high-quality-contracts",
});

​
Export a dataset

client.datasets.export() returns a Promise<Buffer> containing the dataset bytes. Write them to disk using Node’s fs module. 
import { writeFileSync } from "fs";

const data = await client.datasets.export("ds_xyz789", { format: "jsonl" });

writeFileSync("q1-invoices.jsonl", data);
console.log("Export saved.");

​
Supported export formats

Formatformat valueBest for
JSON Lines"jsonl"LLM fine-tuning, streaming ingestion
CSV"csv"Spreadsheet tools, quick inspection
Parquet"parquet"Columnar analytics, Spark / DuckDB pipelines
Markdown"markdown"Human review, RAG document stores
Arrow"arrow"High-performance in-memory data exchange
import { writeFileSync } from "fs";

const formats = ["jsonl", "parquet", "csv"] as const;

for (const fmt of formats) {
  const data = await client.datasets.export("ds_xyz789", { format: fmt });
  writeFileSync(`q1-invoices.${fmt}`, data);
  console.log(`Saved q1-invoices.${fmt}`);
}

​
Profile a dataset

client.datasets.profile() returns statistical metadata about your dataset — token counts, field coverage, PII distribution, and grade breakdown — useful for quality checks before fine-tuning. 
const profile = await client.datasets.profile("ds_xyz789");

console.log(`Total records:       ${profile.recordCount}`);
console.log(`Total tokens:        ${profile.totalTokens}`);
console.log(`Grade A records:     ${profile.gradeCounts["A"]}`);
console.log(`Unique PII types:    ${profile.piiTypeCount}`);
console.log(`Avg quality score:   ${profile.avgQualityScore.toFixed(2)}`);
Run profile() after building and before exporting to catch low-quality datasets early — especially when assembling training data for fine-tuning.

​
Delete a dataset

Deleting a dataset removes the assembled export artifact. The underlying jobs and their extracted data are not deleted. 
await client.datasets.delete("ds_xyz789");
console.log("Dataset deleted.");
Deletion is immediate and irreversible. If you need the data again, you must call client.datasets.build() to reassemble it from the original jobs.

​
Next steps

Jobs

Learn how to filter and manage jobs before building datasets.

API Reference

Full parameter and return-type reference for all dataset methods.