Prerequisites #

Operator usage prerequisites: Enable the Visual Intelligence product - Text Recognition - Intelligent Document Parsing service.

Operator parameters #

Input #

Output #

An array of structs containing parsing results.

• parsed_origin_text: The parsed original Markdown text.
• parsed_plain_text: Plain text with image links removed.
• parsed_detail: Detailed parsing information in JSON format.
• parsed_file_path: Storage path of the parsing result on TOS.
• parsed_image_filenames: List of image filenames in the parsing result.

Parameters #

If a parameter does not have a default value, it is required.

Examples #

The following code demonstrates how to use Daft (for distributed scenarios) to run the operator and parse PDF documents.

# Copyright (c) Beijing Volcano Engine Technology Ltd.

from __future__ import annotations

import os

import daft
from daft import col
from daft.las.functions.doc import PDFParse
from daft.las.functions.udf import las_udf

if __name__ == "__main__":

    # After modifying the parsed content, it will be saved to the specified TOS path. Therefore, you need to set environment variables to ensure permission to write to TOS, including: ACCESS_KEY, SECRET_KEY, TOS_ENDPOINT, TOS_REGION, TOS_TEST_DIR
    TOS_DIR = os.getenv("TOS_TEST_DIR", "tos_bucket")
    output_tos_path = f"tos://{TOS_DIR}/doc/parse/pdf_parse"
    
    if os.getenv("DAFT_RUNNER", "native") == "ray":
        import logging

        import ray

        def configure_logging():
            logging.basicConfig(
                level=logging.INFO,
                format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
                datefmt="%Y-%m-%d %H:%M:%S.%s".format(),
            )
            logging.getLogger("tracing.span").setLevel(logging.WARNING)
            logging.getLogger("daft_io.stats").setLevel(logging.WARNING)
            logging.getLogger("DaftStatisticsManager").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaScheduler").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaDispatcher").setLevel(logging.WARNING)

        ray.init(dashboard_host="0.0.0.0", runtime_env={"worker_process_setup_hook": configure_logging})
        daft.set_runner_ray()
    daft.set_execution_config(actor_udf_ready_timeout=600)
    daft.set_execution_config(min_cpu_per_task=0)

    # Build the URL using environment variables
    tos_dir_url = os.getenv("TOS_DIR_URL", "las-cn-beijing-public-online.tos-cn-beijing.volces.com")
    samples = {
        "input_url": [
            f"https://{tos_dir_url}/public/shared_doc_dataset/sample.pdf"
        ],
        "filename": ["sample.pdf"],
    }
    df = daft.from_pydict(samples)

    constructor_kwargs = {"input_type": "url", "output_tos_path": output_tos_path, "qps": 1}

    # Use Daft for distributed processing
    df = df.with_column(
        "parsed_result",
        las_udf(PDFParse, construct_args=constructor_kwargs, concurrency=1)(col("input_url"), col("filename")),
    )
    df = df.with_column("parsed_text", col("parsed_result")["parsed_origin_text"])
    df.show()

    # ╭────────────────────────────────┬────────────┬────────────────────────────────────────────────────────────────────────────────────────────────────────┬────────────────────────────────╮
    # │ input_url                      ┆ filename   ┆ parsed_result                                                                                          ┆ parsed_text                    │
    # │ ---                            ┆ ---        ┆ ---                                                                                                    ┆ ---                            │
    # │ Utf8                           ┆ Utf8       ┆ Struct[parsed_origin_text: Utf8, parsed_plain_text: Utf8, parsed_detail: Utf8, parsed_file_path: Utf8, ┆ Utf8                           │
    # │                                ┆            ┆ parsed_image_filenames: List[Utf8]]                                                                    ┆                                │
    # ╞════════════════════════════════╪════════════╪════════════════════════════════════════════════════════════════════════════════════════════════════════╪════════════════════════════════╡
    # │  https://las-cn-beijing-publi-… ┆ sample.pdf ┆ {parsed_origin_text: ![fig_94…                                                                         ┆ ![fig_94052](https://pdf-buck… │
    # ╰────────────────────────────────┴────────────┴────────────────────────────────────────────────────────────────────────────────────────────────────────┴────────────────────────────────╯