R=Finder is a subprogram of the DECIMER project developed by Kohulan Rajan and Otto Brinkhaus. Based on a separately trained tesseract-ocr model and using regular expressions (REGEX), R=Finder identifies R-groups in journals (pdf files) and maps them to their structure found by DECIMER-image-segmentation and complements/corrects the given SMILES string generated by a "predictor" program.
Installation
To use R=Finder ... it is very complex because of tensorflow dependencies... git pull ... TODO
