add argumentation mining (AM) datasets

[ArneBinder]

This is an overview issue, please create PRs for the following datasets:

• aae2 (requires add generic BRAT dataset #9), see add aae2 dataset #64 and add aae2 dataset - v.2 #92
• abstrct (requires add generic BRAT dataset #9), see add abstrct dataset #68
• argmicro, see Add AM dataset argmicro #56
• cdcp, see Add AM dataset - CDCP #15
• sciarg (requires add generic BRAT dataset #9), see add sciarg dataset #61
• scidtb_argmin, see Add AM dataset SciDTB_Argmin #54

For each dataset, there should be:

• a PIE dataset builder in dataset_builders/pie/DATASET_NAME with
  • a PIE dataset loading script DATASET_NAME.py or a config DATASET_NAME.yaml in the case of BRAT based datasets, and
  • a dataset card in a README.md file
• a HF dataset builder, if not yet publicly available (except on https://huggingface.co/DFKI-SLT), in dataset_builders/hf/DATASET_NAME with
  • a HF dataset loading script DATASET_NAME.py (NOTE: this should be copied from the location mentioned at BASE_DATASET_PATH of the PIE dataset builder!), and
  • a dataset card in a README.md file. Note: this should be directly put next to the HF script, e.g. at https://huggingface.co/DFKI-SLT
• tests for everything

Notes:

• this should build on content from https://github.com/ArneBinder/pie-document-level as much as possible
• the readmes and PIE builder scripts / configs should be in dataset_builders/pie/DATASET_NAME; the tests in tests/dataset_builder/pie/test_DATASET_NAME.py
• this may require to implement new (generic) conversion or processing code as in https://github.com/ArneBinder/pie-document-level. For any of that, there should also be tests in the respective locations under tests.