What we do is to hash the JSON data string passed to the python, and put that in the file name.
The trouble is that string hashing in Python is intentionally not reliable between python processes (and there is one process per query), so no match is ever found and the same image is generated all over again.
It needs to be fixed separately in both the arch and westminster python backends.
What we do is to hash the JSON data string passed to the python, and put that in the file name. The trouble is that string hashing in Python is intentionally not reliable between python processes (and there is one process per query), so no match is ever found and the same image is generated all over again.
It needs to be fixed separately in both the arch and westminster python backends.