For pre-calculated networks (can be done in advance with uniprot_scraper scripts for every UniProt ID / HGNC symbol) it then works very fast, but for any other query, using multiple gene names, it does not store the results. If it were just like in the scraper script, then another user's request for the same set of genes might perform much faster. And, it's ok to store (or cache permanently), because the PC2 pathway data is static, read-only, i.e., doesn't change between Pathway Commons releases (for several months).
For pre-calculated networks (can be done in advance with uniprot_scraper scripts for every UniProt ID / HGNC symbol) it then works very fast, but for any other query, using multiple gene names, it does not store the results. If it were just like in the scraper script, then another user's request for the same set of genes might perform much faster. And, it's ok to store (or cache permanently), because the PC2 pathway data is static, read-only, i.e., doesn't change between Pathway Commons releases (for several months).