Is it possible to dump the memory snapshot for offline analysis? about guppy3 HOT 3 CLOSED

Guppy can dump the statistics from an existing guppy session

IdentitySet.dump is the function.

>>> import guppy
>>> guppy.hpy().heap().dump('test.hpy')

makes:

.loader: _load_stat
.format: SetFormat
.timemade: 1628756373.710652
.count: 38146
.size: 4497412
.kindname: 
.kindheader: Kind (class / dict of class)
.numrows: 127
.r: 11118 985591 str
.r: 7725 535840 tuple
.r: 576 480952 type
.r: 2668 471304 types.CodeType
.r: 5242 375945 bytes
.r: 2448 332928 function
[...]
.r: 1 16 builtins.NoneType
.r: 1 16 builtins.NotImplementedType
.r: 1 16 builtins.ellipsis
.end: .loader: _load_stat

... which you can get by guppy.hpy().heap().all (use repr() or str() to convert it to string)

from guppy3.

Unfortunately, not really :( Guppy can dump the statistics from an existing guppy session, but any 'more advanced' analysis, such as trying to filter by reference types, needs to be analyzed from within.

from guppy3.

I'm gonna close this as dup of #22 and #24. Feel free to reopen if you have more questions.

from guppy3.