@dreamyoungs/trex

Node.js wrapper around the TREX CLI.

Prerequisites

• Node.js 18+
• Recommended: internet access during install (postinstall downloads prebuilt TREX binary)

Install

npm install @dreamyoungs/trex

By default, package install tries to download a matching TREX binary from GitHub Releases. If download is unavailable, set either:

• TREX_BIN=/path/to/trex
• options.binPath in API calls

To skip binary download explicitly:

TREX_SKIP_DOWNLOAD=1 npm install @dreamyoungs/trex

Maintainer: publish release binaries

When bumping package version, upload matching release assets first:

scripts/release/publish_assets.sh --version 0.1.0 --upload

See scripts/release/README.md for details.

Usage

const { extract, extractCsv } = require("@dreamyoungs/trex");

async function run() {
  const tables = await extract("./invoice.pdf", {
    mode: "auto",
    pages: [1, 2],
  });

  const csv = await extractCsv("./invoice.pdf", {
    mode: "lattice",
  });

  console.log(tables.length, csv.length);
}

run().catch(console.error);

Options

• pages: number[] | string (e.g. [1,2,3] or "1-3,7")
• mode: "auto" | "lattice" | "stream" | "dl"
• dlModel, dlMinConfidence, dlFallback
• binPath: override TREX binary path
• timeoutMs: command timeout (default: 120000)
• eventLog, eventDocumentKey, eventTenantId, eventRequestId, eventFeedbackTag, eventTrainingOptIn

Buffer APIs

• extractFromBuffer(buffer, options)
• extractCsvFromBuffer(buffer, options)

These methods write a temporary PDF file, run TREX, then clean up automatically.