When iterating over documents, it would be helpful to distinguish between errors specific to a document (such as when an input file is not found) and errors that affect all documents (such as the CoreNLP server being unavailable or we cannot load the NER model). In the former case, processing should advance to the next document. In the latter case, processing should stop with a message to the user on standard out indicating the run was terminated prematurely (details can be in the log file).
When iterating over documents, it would be helpful to distinguish between errors specific to a document (such as when an input file is not found) and errors that affect all documents (such as the CoreNLP server being unavailable or we cannot load the NER model). In the former case, processing should advance to the next document. In the latter case, processing should stop with a message to the user on standard out indicating the run was terminated prematurely (details can be in the log file).
Global problems: