Google Books Ngram JSON: use corpus=26 for possessives; query terms singly
The undocumented endpoint books.google.com/ngrams/json?content=...&corpus=X returns [] for any phrase containing an apostrophe ("Prim's algorithm", "Prim 's algorithm") when corpus=en; with the numeric id corpus=26 (English 2019) the same queries work and come back normalized as "Prim ' s algorithm", with hyphens as "Floyd -- Warshall". Because names are normalized, comma-joining several terms makes it hard to tell which ones were absent, so send one term per request: [] then means the n-gram is below the 40-occurrence inclusion threshold, which is itself a finding. Values are relative frequencies; multiply by the yearly totals in storage.googleapis.com/books/ngrams/books/20200217/eng/totalcounts-1 to get approximate raw counts, and report those. Many niche-term series turn out to be only a handful of books per year.
Correction from later use: [] does not always mean below threshold. Spaced-hyphen spellings like "Roy - Warshall" or "Newcomb - Benford" can return [] while "Roy-Warshall" and "Newcomb-Benford" return data. Try several spellings (hyphen spaced and unspaced, " 's" and "'s") before concluding a term is absent.