Official implementation for ChemDT and ChemDTD

ChemDT

The overall extraction workflow for ChemDT"

Run ChemDT to extact IUPACs in PDF

1. pull docker image

docker push registry.cn-beijing.aliyuncs.com/hilab_space/chemdt:v1

docker pull registry.cn-beijing.aliyuncs.com/hilab_space/pdf_process:v1

2. start the docker service

docker run --rm -e is_debug=N -e dev_or_pro=dev -p 5533:5533 chemdt:v1

docker run --rm -p 3114:3080 pdf_process:v1

3. Install envirnment

conda env create -f environment.yaml
conda activate chemdt

4. run main function

cd tools
python client.py

APIs for extaction piplines

1. Convert PDF fild to images.

client = ChemDTClient()
pages = client.pdf2img("path_of_pdf"，"save_dir")

2. IUPAC detection.

client = ChemDTClient()
iupac_det_res = client.iupac_det(origin_img_path)["iupac_result"]

3. OCR

client = ChemDTClient()
box, txt = client.iupac_ocr("path_of_image")

4. Iupac fixer

client = ChemDTClient()
client.iupac_fixer("iupac_text")

5. Full pipline

client = ChemDTClient()
pdf_path = '../test_data/p2.pdf'
save_dir = '../pdf_images'
client.run_pipline(pdf_path, save_dir = save_dir)

ChemDTD

The front end of ChemDTD

The front end of ChemDTD

Run webserver for ChemDTD for visualizing your local database

1. Write extracting data to the local database. You need modify the configuration in conf.yaml

python to_db.py

1. Start webserver and wait a moment to let the front-end to load all your data.

streamlit run app.py --server.addres 0.0.0.0 --server.port 5001

2. View your app in your browser with port 5001