- The ratspub shepherd service has 3 environment variables which also act as hints for directories which are needed.
- EDIRECT_PUB_MASTER points to the local copy of the PubMed archive (this can be retrieved with the mcron job pubmed.guile). This is currently at /export2/PubMed. Due to folder permissions it is easiest to run the cronjob to update the PubMed archive as hchen.
- NLTK_DATA contains nltk data which is needed for fetching more PubMed archives, either from the cronjob or during the regular use of rats.pub. This is currently run by Hao and is at /home/hchen/nltk_data
- PERL_LWP_SSL_CA_FILE is because perl-lwp doesn't respect SSL_CA_FILE and needs its own variable to handle SSL. This is set to the standard SSL location, /etc/ssl/certs/ca-certificates.crt.
- ratspub needs to be run from its installed directory (or from $GUIX_ENVIRONMENT if inside a Guix environment with it, 'guix environment --ad-hoc ratspub').