SPL parsing for route, sponsor (labeller) and allergenic extracts

[grishagin]

dailymed_get.sh and dailymed_prepare.sh
produce spl_acti_otc.txt, spl_acti_rx.txt, etc.

SPL parsing currently fails to retrieve route of administration.

Comparison with DrugBank parsing of the same files - it would also be nice to get sponsor, for example:

• {
  • kind: "USApprovalRx",
  • source: "FDA NDC",
  • id: "DB10687",
  • sponsor: "Nelco Laboratories, Inc.",
  • product: "Hops",
  • route: "Intradermal; Subcutaneous",
  • created: 1539467255070,
  • jurisdiction: "US",
  • comment: "36987-1775",
  • approvalAppId: "BLA102192",
  • URL: "https://www.drugbank.ca/drugs/DB10687",
  • startDate: "1972-08-29"
• },
• {
  • kind: "USUnapproved",
  • source: "spl_acti_rx.txt",
  • id: "NDC 36987-1777",
  • product: "HOPS",
  • created: 1539467255070,
  • jurisdiction: "US",
  • active: "true",
  • approvalAppId: "BLA102192",
  • URL: "https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=42a03973-e3db-49f2-9264-f76aa831c045",
  • startDate: "1972-08-29",
  • marketingStatus: "BLA",
  • productCategory: "HUMAN PRESCRIPTION DRUG LABEL"
• },

While we attempt to specifically identify allergenic extracts as a class, not all products are clearly marked as such, and we likely need to parse to boxed warning, indication, clinical pharmacology text to more robustly identify such products.
For example:
https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=3b19e1e4-1a28-4515-ac20-cc4a9b2d45e6

Lastly, we would also ideally index products for inactive ingredients --- so see which inactive ingredients, at which strength, are included in currently approved medicines.