This is a small program which will grab all the papers posted in the last day, convert them to pure text, and, in each of them, search for the regular expressions listed in the configuration file.
First install the required dependencies:
$ pip install arxiv dateutil slate3k
Then clone this repository
$ git clone https://github.com/bfichera/arxivscraper
Finally, make a file called conf.py in the base directory with your configuration details. See example_conf.py.
Just run
$ python arxivscraper.py --config-file /path/to/my/config/file
See example_conf.py. The search terms which are chemical formulas should go into the chem_terms field, with spaces between elements; they will be converted to a regular expression which I've found matches well most of the time (different authors like to format TaS2 like TaS2, TaS$_2$, TaS$_{2}$, etc. and the chem_terms tries to catch all of them). Otherwise, use the terms field.