pyjevsim 2.1.2

pyjevsim 2.1.2

The 2.1.x line makes pyjevsim interoperable and distributable: a DEVS
model can run, unchanged, as an HLA (IEEE 1516-2010) federate. 2.1.2
completes that story with a true multi-process / multi-host distributed
example and the supporting RTI plumbing, validated live against
Pitch pRTI.

Highlights

• HLA federate support via pluggable RTI backends. Declare HLA
  bindings on existing model ports; an HLAExecutorFactory bridges them
  to a transport selected by name with create_rti(...). The same model
  runs standalone, on an in-process bus, or on a live RTI.
• Distributed ping-pong example. Two federates exchange interactions
  and synchronize an object attribute under RTI-managed logical time —
  runnable offline (no Java) or live, including one federate per OS
  process across hosts.
• Validated on Pitch pRTI (IEEE 1516-2010): federation
  join/publish/subscribe/resign, time-stamp-ordered interactions, object
  registration/reflection, time regulation/constraint with TAR/grant.

New in 2.1.2 (since 2.1.1)

• Distributed (multi-process) HLA ping-pong: one federate per OS process
  (own JVM/LRC) joined to one federation —
  examples/hla_pingpong/run_pitch_federate.py,
  run_pitch_multiprocess.py.
• PitchTransport: federation synchronization-point support
  (start barrier) and a configurable CRC endpoint
  (crc="host:port" / PYJEVSIM_CRC) for multi-host federations.
• Packaging/citation: .zenodo.json, CITATION.cff, Zenodo DOI badge.

The HLA subsystem at a glance (pyjevsim.hla)

Component	Role
RTIConnector	Base class for RTI backends — implement two methods, inherit the rest
RTICapabilities / Codec	Feature negotiation / pluggable FOM (de)serialization
HLAInteraction, HLAAttribute	Port-to-FOM bindings
HLAExecutorFactory, Federate	Wire models to a transport; drive the time-advance loop
Backends	loopback, inprocess (no Java), pitch (Pitch pRTI via JPype)
create_rti / register_rti	Select / add a backend by name

A new RTI (CERTI, Portico, OpenRTI, MÄK, …) is added by subclassing
RTIConnector — see docs/hla/rti_interface.md.

Also in the 2.1.x line

• 2.1.0 — RTI-agnostic HLA interface + Pitch pRTI backend + unified
  V_TIME/R_TIME/HLA two-phase tick (correct confluent / TSO semantics).
• 2.1.1 — canonical republish from main.

Install

pip install pyjevsim==2.1.2
pip install pyjevsim[hla-pitch]   # adds JPype for the Pitch pRTI backend

The loopback / inprocess backends need nothing beyond the core; the
pitch backend additionally needs JPype, Java ≥ 9, and an IEEE 1516-2010
RTI with a running CRC.

Validation

Live, two separate OS processes (Pitch pRTI Free 5.5.2 + Temurin 11),
connecting to the CRC over the network (routable LAN address, not
localhost):

pong received pings: [0, 1, 2, 3]
ping received pongs: [0, 1, 2, 3]
pong reflected hits: [0, 1, 2, 3]   # object synchronization

Test suite: 110 passed, 2 skipped (Pitch live cases skip without a JVM).

Links

• Quick start & HLA guide: https://pyjevsim.readthedocs.io/
• Example: examples/hla_pingpong/
• Full changelog: https://github.com/eventsim/pyjevsim/blob/v2.1.2/CHANGELOG.md
• Compare: v2.1.1...v2.1.2

Citation

Archived on Zenodo — see the DOI badge in the
README and
CITATION.cff.

v2.1.1

pyjevsim 2.1.1

Republish from main. PyPI 2.1.0 shipped an older HLAExecutor inbound
convention (built from the feature branch); 2.1.1 is the canonical build
from main. No in-tree behaviour change versus the 2.1.0 source on main.

v2.1.0

pyjevsim 2.1.0

pyjevsim v1.2.0

pyjevsim, a python-based DEVS environment that allows snapshots of behavioral models and their relationships.
version 1.2.0

pyjevsim v1.1.0

pyjevsim, a python-based DEVS environment that allows snapshots of behavioral models and their relationships.
version 1.1.0

pyjevsim v1.0.0

pyjevsim, a python-based DEVS environment that allows snapshots of behavioral models and their relationships.