Resource controls

Normal plotting needs no configuration. Curve generation memoises stroke data; coordinate and derived image arrays may still be allocated again on later calls.

An order-k path contains 4**k cells. Coordinates alone cost 16 bytes per cell; images, masks, FFTs and intermediate arrays add to peak memory. The default cell limit is a guard against accidental huge allocations, not a reserved memory budget.

import hilbertplot as hp

previous = hp.config.set_cell_limit(4**10)
try:
    # work within the temporary limit
    pass
finally:
    hp.config.set_cell_limit(previous)
hp.config.clear_cache()

hilbertplot.config.cache_info()

Return statistics for cached recursive curve construction.

Returns:

info – Standard functools cache statistics: hits, misses, maxsize, currsize. maxsize is None for this unbounded memo cache. currsize counts recursive construction entries, not cells or bytes.

Return type:

namedtuple

See also

hilbertplot.config.clear_cache

Release cached construction entries.

Notes

The cache stores recursive stroke arrays shared between curves. It does not report arrays owned by callers or open Matplotlib figures.

Examples

>>> import hilbertplot as hp
>>> hp.config.cache_info().maxsize is None
True

hilbertplot.config.clear_cache()

Clear memoized recursive stroke arrays and reset cache statistics.

Returns:

The generation cache is cleared in place.

Return type:

None

See also

hilbertplot.config.cache_info

Inspect cache entries and hit/miss counts.

Notes

Does not close figures or release arrays referenced elsewhere. A process’s resident memory need not decrease immediately, because Python and NumPy allocators may retain freed memory for reuse.

Examples

>>> import hilbertplot as hp
>>> hp.config.clear_cache()
>>> hp.config.cache_info().currsize
0

hilbertplot.config.cell_limit()

Return the process-wide ceiling for a materialized curve request.

Returns:

cells – Maximum permitted number of cells in one request. The default is 2**28; this is a ceiling, not preallocated or reserved memory.

Return type:

int

See also

hilbertplot.config.set_cell_limit

Change or restore the guard.

Notes

Coordinate pairs alone cost 16 bytes per cell. Temporary arrays, cached strokes, images and rendering can substantially increase peak memory. Direct index_to_xy/xy_to_index calls do not materialize a grid.

hilbertplot.config.set_cell_limit(cells)

Set the ceiling on a single materialised curve and return the previous value.

Parameters:

cells (int or None) – The new limit in cells. None restores the default; a limit of 0 or less is rejected. Pass a very large number to effectively disable the check.

Returns:

The limit that was in force before this call, so it can be restored.

Return type:

int

Examples

>>> import hilbertplot
>>> previous = hilbertplot.config.set_cell_limit(4 ** 10)
>>> hilbertplot.config.cell_limit()
1048576
>>> _ = hilbertplot.config.set_cell_limit(previous)