Skip to main content

CLI Usage

OpenGIN Tracer provides a Command Line Interface (CLI) to interact with the pipeline.

Installation​

Ensure you have the package installed (preferably in a virtual environment):

pip install -e .

Running a Trace​

To extract data from a PDF file, use the opengin tracer run command. This command supports both local files and URLs.

opengin tracer run <INPUT_SOURCE> [OPTIONS]

Arguments​

  • INPUT_SOURCE: Absolute/relative path to a local PDF file OR a URL (starting with http:// or https://).

Options​

  • --name: A human-readable name for this pipeline run (e.g., "financial_report_2024"). Defaults to run_<timestamp>.
  • --prompt: The extraction prompt text OR a path to a text file containing the prompt. Defaults to "Extract all tables.".
  • --metadata-schema: Path to a YAML file defining the metadata schema to extract for each table.

Examples​

1. Basic Usage (Local File):

opengin tracer run ./data/sample_invoice.pdf --name output_pipeline

2. Using a Remote URL: The CLI will automatically download the file to a temporary location, process it, and clean it up.

opengin tracer run https://raw.githubusercontent.com/LDFLK/opengin-ingestion/main/data/quickstart_sample.pdf --name quickstart-sample

3. Using a Prompt File: For complex instructions, save your prompt in a text file.

opengin tracer run ./data/invoice.pdf --prompt ./prompts/invoice_extraction.txt

4. With Metadata extraction: Extract keys defined in a YAML schema.

opengin tracer run ./data/doc.pdf --metadata-schema ./metadata.yml

Output​

After execution, the results can be found in the pipelines/<pipeline_name>/<run_id>/output/ directory.

The console output will display the run_id for your reference.