LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows
Python ai-agentsdata-engineeringdocument-aigenerative-aiidpjson-extractionllmmcp-serverocrpdf-extractionprompt-engineeringstructured-output
import { afterEach, describe, expect, it, vi } from "vitest";
import { saasOnlyErrorProps } from "./PostHogDeployment";
const setHost = (hostname) =>
vi.spyOn(window, "location", "get").mockReturnValue({ hostname });
describe("saasOnlyErrorProps", () => {
afterEach(() => vi.restoreAllMocks());
it("drops error text on self-hosted deployments", () => {
setHost("unstract.customer.internal");
expect(saasOnlyErrorProps("connect to https://10.0.0.5 failed")).toEqual(
{},
);
});
it("keeps error text, capped at 200 chars, on SaaS", () => {
setHost("us-central.unstract.com");
const props = saasOnlyErrorProps("x".repeat(500));
expect(props.error).toHaveLength(200);
});
it("adds nothing when there is no string error", () => {
setHost("eu-west.unstract.com");
expect(saasOnlyErrorProps(undefined)).toEqual({});
expect(saasOnlyErrorProps([{ detail: "limit" }])).toEqual({});
});
});