Blocking Scheme¶
- class halmd.observables.dynamics.blocking_scheme(args)¶
Construct blocking scheme.
- Parameters:
args – keyword arguments
args.max_lag (number) – maximum lag time in MD units
args.every (number) – sampling interval of lowest coarse-graining level in integration steps
args.size (number) – size of each block, determines coarse-graining factor
args.shift (number) – coarse-graining shift between odd and even levels (default: \(\lfloor\sqrt\text{size}\rfloor\))
args.separation (number) – minimal separation of samples for time averages in sampling steps (default: size)
args.flush (number) – interval in seconds for flushing the accumulated results to the file (default: 900)
- disconnect()¶
Disconnect blocking scheme from sampler.
- class correlation(args)¶
Compute time correlation function.
- Parameters:
args (table) – keyword arguments
args.tcf – time correlation function
args.file – instance of
halmd.io.writers.h5md
args.location (string table) – location within file (optional)
The argument
tcf
specifies the time correlation function. It is expected to provide the attributesacquire
(1 or 2 callables that yield the samples to be correlated) anddesc
(module description) as well as a methodwriter
(file writer). Suitable modules are found inhalmd.observables.dynamics
, see there for details.The argument
location
specifies a path in a structured file format like H5MD given as a table of strings. If omitted it is defined by the time correlation function, typically {"dynamics"
, particle group, name of correlation function}.- disconnect()¶
Disconnect correlation function from blocking scheme.