Skip to content

Latest commit

 

History

History
122 lines (87 loc) · 3.24 KB

README.md

File metadata and controls

122 lines (87 loc) · 3.24 KB
Raw

Llama Parse CLI

A non-official command-line interface (CLI) for parsing documents using the LlamaIndex Parser.

Installation

To install the llama-parse-cli, you need to have Node.js and npm installed on your system. Then, you can install it globally using:

npm install -g llama-parse-cli

Usage

The CLI provides two main commands: auth and parse.

Authentication

Before using the parse functionality, you need to authenticate with your API key:

llama-parse auth

This command will prompt you to enter your API key, which should start with "llx-". The key will be securely stored in ~/.llama-parse/config.json.

Parsing Documents

To parse a document, use the parse command:

llama-parse parse <file> [options]

Options:

  • -f, --format <format>: Output format (json, markdown, text). Default: markdown
  • -o, --output <output>: The output file
  • -ol, --ocr-language <ocr-language>: The language of the document
  • -pi, --parsing-instructions <parsing-instructions>: The parsing instructions
  • -ps, --page-separator <page-separator>: The page separator
  • -sd, --skip-diagonal-text: Skip diagonal text
  • -ic, --invalidate-cache: Invalidate cache
  • -dc, --do-not-cache: Do not cache
  • -dnc, --do-not-unroll-columns: Do not unroll columns
  • -fm, --fast-mode: Fast mode
  • -gpt4o, --gpt-4o: Use GPT-4o
  • -tp, --target-pages <target-pages>: The target pages (comma-separated list, starting from 0)
  • -v, --verbose: Verbose mode

Examples

  1. Parse a PDF file and output in markdown format:
llama-parse parse example/sample-pdf.pdf
  1. Parse a document and save the output to a file:
llama-parse parse example/sample-pdf.pdf -o output.md
  1. Parse specific pages of a document:
llama-parse parse example/sample-pdf.pdf -tp 0
  1. Parse a document in verbose mode:
llama-parse parse example/sample-pdf.pdf -v

Development

The project is built using TypeScript and uses the following key dependencies:

  • Commander.js for CLI argument parsing
  • Zod for input validation
  • Inquirer for interactive prompts
  • Ora for spinner animations
  • Chalk for colorful console output

To set up the development environment:

  1. Clone the repository
  2. Run pnpm install to install dependencies
  3. Use pnpm dev to run the CLI in development mode
  4. Use pnpm build to build the project
  5. Use pnpm link . to link the CLI globally

The main entry point is in src/index.ts:

#!/usr/bin/env node
import { version } from "../package.json";
import { program } from "commander";
import { createAuthCommand, createParseCommand } from "./commands";

program
  .name("llama-parse")
  .addCommand(createAuthCommand())
  .addCommand(createParseCommand())
  .description("A CLI for parsing documents with llama parse")
  .version(`${version}`);

program.parse(process.argv);

The CLI commands are defined in separate files under the src/commands directory:

export { createAuthCommand } from "./auth";
export { createParseCommand } from "./parse";

Contributing

Contributions to the llama-parse-cli are welcome. Please ensure that your code adheres to the existing style and includes appropriate tests.