Description

Currently we save every URI ever qname'd forever in here: https://github.com/RDFLib/rdflib/blob/master/rdflib/namespace.py#L289

If you serialize a big graph in a store that is off disk, we may run out of memory.

Either:

• Add a lru_cache (from py3 functools + this 2.7 implementation
• Depend on

Is your feature request related to a problem? Please describe.
CSV files produced in different locales have different default separators, a comma-separated file (",") is but one type of CSV.

In Sweden we usually use a semicolon (";") as separator and many CSV files use a tab as separator ("\t"). Importing CSVs requires reformatting them before import, and causes inoperability issues and f

Debug logging for lifecycle methods like initialization and when data has to be re-inferred (eg. when something is deleted).

Trace logging would be per statement to log what gets inferred.

Following the robustness principle, in HeaderUtil.ts, we might want to omit invalid headers values from the parsed result as opposed to throwing.

Hi, Is it possible to add in readme more information about support?

• What are the feature necessary for LDflex to work?
• Is Node supported?
• Are browsers supported? Which ones?
• Can polyfills help?