unstructured/examples/ingest/biomed/ingest-with-path.sh

#!/usr/bin/env bash

# Processes the Unstructured-IO/unstructured repository
# through Unstructured's library in 2 processes.

# Structured outputs are stored in biomed-ingest-output-path/

# Biomedical documents can be extracted in one of two ways, in this script is the FTP directory approach.

# The supported ftp directories is:
# https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf

# By providing the path, the documents existing therein are downloaded.
# For example, to download the documents in the path: https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/07/
# The path needed is oa_pdf/07/

SCRIPT_DIR=$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" &>/dev/null && pwd)
cd "$SCRIPT_DIR"/../../.. || exit 1

# The example below will ingest the PDF from the "oa_pdf/07/07/sbaa031.073.PMC7234218.pdf" path.

# You can ingest all the documents in the "oa_pdf/07/07" path by passing "oa_pdf/07/07" instead.
# WARNING: There are many documents in that path.

PYTHONPATH=. ./unstructured/ingest/main.py \
  biomed \
  --path "oa_pdf/07/07/sbaa031.073.PMC7234218.pdf" \
  --output-dir biomed-ingest-output-path \
  --num-processes 2 \
  --verbose \
  --preserve-downloads

# Alternatively, you can call it using:
# unstructured-ingest --biomed-path ...