Document Parsing Models - Inference Guide

Overview

The scripts in this folder allow users to extract structured data from unstructured documents using different document parsing services and libraries. Each service follows a standard installation procedure and provides an infer_* script to perform inference on PDF or Image samples.

You can choose from document parsing products such as Upstage DP, AWS Textract, Google Document AI, Microsoft Azure Form Recognizer, LlamaParse, or Unstructured. Most of these services require an API key for access, so ensure you follow specific setup instructions for each product to configure the environment correctly.

Each service generates a JSON output file in a consistent format. You can find detailed information about the output format here.

Upstage

Follow the official Upstage DP Documentation to set up Upstage for Document Parsing.

Note: Ensure that the UPSTAGE_ENDPOINT and UPSTAGE_API_KEY variables are set up to run the code.

Use the script below to make an inference:

$ python infer_upstage.py \
    --data_path <path to the benchmark dataset> \
    --save_path <path to save the .json file>

AWS

To use AWS Textract for document parsing, install AWS CLI and Boto3 for API interaction:

$ curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
$ unzip awscliv2.zip
$ sudo ./aws/install
$ aws configure
$ pip install boto3

Refer to the AWS Textract Documentation for detailed instructions.

Note: To run the AWS inference code, you need to set the following variables: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_REGION, and AWS_S3_BUCKET_NAME.

Use the script below to make an inference:

$ python infer_aws.py \
    --data_path <path to the benchmark dataset> \
    --save_path <path to save the .json file>

Google

Install Google Cloud SDK and Google Document AI for document parsing on Google's platform:

$ apt-get install apt-transport-https ca-certificates gnupg curl
$ curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | gpg --dearmor -o /usr/share/keyrings/cloud.google.gpg
$ echo "deb [signed-by=/usr/share/keyrings/cloud.google.gpg] https://packages.cloud.google.com/apt cloud-sdk main" | tee -a /etc/apt/sources.list.d/google-cloud-sdk.list
$ apt-get update && apt-get install google-cloud-cli
$ gcloud init
$ pip install google-cloud-documentai

More information can be found in the Google Document AI Documentation

Note: To run the Google inference code, you need to set the following variables: GOOGLE_PROJECT_ID, GOOGLE_PROCESSOR_ID, GOOGLE_LOCATION, and GOOGLE_ENDPOINT.

Use the script below to make an inference:

$ python infer_google.py \
    --data_path <path to the benchmark dataset> \
    --save_path <path to save the .json file>

LlamaParse

Refer to the official LlamaParse Documentation to install and use LlamaParse for document analysis.

Note: Ensure that the LLAMAPARSE_API_KEY, LLAMAPARSE_POST_URL, and LLAMAPARSE_GET_URL variables are set before running the code.

Use the script below to make an inference:

$ python infer_llamaparse.py \
    --data_path <path to the benchmark dataset> \
    --save_path <path to save the .json file>

Microsoft

Install the Azure AI Form Recognizer SDK:

$ pip install azure-ai-formrecognizer==3.3.0

See the Microsoft Azure Form Recognizer Documentation for additional details.

Note: Set the MICROSOFT_API_KEY and MICROSOFT_ENDPOINT variables before running the code.

Use the script below to make an inference:

$ python infer_microsoft.py \
    --data_path <path to the benchmark dataset> \
    --save_path <path to save the .json file>

Unstructured

To handle various document formats with Unstructured, follow the steps below:

$ pip install "unstructured-client"

Detailed installation instructions can be found here.

Note: To run the Unstructured inference code, you must set the UNSTRUCTURED_API_KEY and UNSTRUCTURED_URL variables.

Use the script below to make an inference:

$ python infer_unstructured.py \
    --data_path <path to the benchmark dataset> \
    --save_path <path to save the .json file>

Standardize Layout Class Mapping

Within each infer_* script, a CATEGORY_MAP is defined to standardize the mapping of layout elements across different products.
This ensures uniform evaluation by mapping the extracted document layout classes to the standardized layout categories for comparative analysis and evaluation purposes.

Be sure to modify the CATEGORY_MAP in the inference scripts according to the document layout categories you are working with for accurate results.

Below is an example of a CATEGORY_MAP used inside LlamaParse inference script:

CATEGORY_MAP = {
    "text": "paragraph",
    "heading": "heading1",
    "table": "table"
}