I can't think of a thread safe way to:
- append generated haiku to
generated.csv
- dynamically update the upper bound for
/generated/{n} on file updates
- return the
nth generated haiku from generated.csv
without stupid amounts of file I/O.
The same applies to the training set, but to a lesser degree.
I can't think of a thread safe way to:
generated.csv/generated/{n}on file updatesnth generated haiku fromgenerated.csvwithout stupid amounts of file I/O.
The same applies to the training set, but to a lesser degree.