| Name | goalEDP JSON |
| Version |
0.0.7
JSON |
| download |
| home_page | None |
| Summary | Library for processing goal-oriented agents. Processing is done on an event-driven mechanism. Events are saved to be used by explanation generating methods. |
| upload_time | 2024-03-20 18:58:08 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.12 |
| license | None |
| keywords |
scientific
paradigms
agents
explainability
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# goalEDP
Library for processing goal-oriented agents. Processing is done on an
event-driven mechanism. Events are saved to be used by explanation generating
methods.
## Contents
- [Sample application](#sample-application)
- [References](#references)
- [About](#about)
## Sample application
This package is published in the official Python repository. To install, use:
```
python3 -m pip install goalEDP
```
Application example:
```python
from goalEDP.extensions.goal import BeliefsReviewer, GoalStatusPromoter, Goal, Action, Agent, GoalBroker, Conflict
from goalEDP.explainers.simple_explainer import SimpleExplainer
from goalEDP.storages.in_memory import InMemoryHistory
from goalEDP.core import Event
from goalEDP.extensions.web_gui import WebGUI
# Describe how the agent works:
# The Agent has the following flow:
# 1) revise beliefs
# 2) promotes goals statuses
# 3) detects conflicts
# 4) pursues goals
class ReviewAccident(BeliefsReviewer):
def __init__(self):
# Call to superclass constructor
super().__init__(
desc="Review accident data", attrs=["accident"]) # Attributes can be external events, inputted or not by a simulator. They can also be events that represent communication between agents.
async def reviewBeliefs(self):
# The event queue is cleared each time the entity is processed. Maybe you want to save the beliefs in an object variable.
beliefs = dict()
if ("accident" in self.eventQueueByTopic):
# "-1" to get the last one in queue
# Use only JSON-serializable variables as belief values.
# In fact, in this framework all event values must be serializable to JSON.
# If you want to use complex structured data, use compositions with Python's dict structure.
beliefs["accident.coord"] = self.eventQueueByTopic["accident"][-1].value["coord"]
beliefs["accident.hasAtLeastTwoVictims"] = len(
self.eventQueueByTopic["accident"][-1].value["victims"]) >= 2
beliefs["accident.severityIsHigh"] = False
for victim in self.eventQueueByTopic["accident"][-1].value["victims"]:
if victim["bpm"] < 60 or victim["bpm"] > 100:
beliefs["accident.severityIsHigh"] = True
if victim["mmHg"][0] > 14 or victim["mmHg"][1] < 6:
beliefs["accident.severityIsHigh"] = True
return beliefs
# All goal statuses must be True for it to be pursued
class RescuePromoter(GoalStatusPromoter):
def __init__(self):
super().__init__(
desc="Promotes rescue goal statuses", beliefs=["accident.hasAtLeastTwoVictims", "accident.severityIsHigh"], promotionNames=["intention"])
async def promoteOrDemote(self):
# The event queue is cleared each time the entity is processed. Maybe you want to save the promotions in an object variable.
promotions = dict()
if ("accident.severityIsHigh" in self.eventQueueByTopic) and ("accident.hasAtLeastTwoVictims" in self.eventQueueByTopic):
if (self.eventQueueByTopic["accident.severityIsHigh"][-1].value) or (self.eventQueueByTopic["accident.hasAtLeastTwoVictims"][-1].value):
promotions["intention"] = True
else:
promotions["intention"] = False
return promotions
class RescueAction(Action):
def __init__(self):
super().__init__(
desc="Action to rescue victims", beliefs=["accident.coord"])
async def procedure(self) -> None:
print("Rescue at: " +
str(self.eventQueueByTopic["accident.coord"][-1].value))
# Goals with highest priority are pursued first
class Rescue(Goal):
def __init__(self):
super().__init__(desc="Objective of rescuing victims of serious accidents",
promoter=RescuePromoter(), plan=[RescueAction()], priority=0)
# If conflict is detected, the goal with the highest priority is chosen, with the others discarded.
#
# class ConflictOne(Conflict):
# def __init__(self):
# super().__init__(
# desc="Recharge battery instead of saving victim", conflictingGoals=[goalInstance1, goalInstance2])
agent1 = Agent(desc="Robot that saves victims", beliefsReviewers=[
ReviewAccident()], goals=[Rescue()], conflicts=[])
# Instantiates history to save events.
history = InMemoryHistory()
# Instantiates the event broker, which also controls the execution sequence
broker = GoalBroker(agents=[agent1], history=history)
broker.startProcess()
# Enter external events, from a simulator for example.
broker.inputExternalEvents([
Event("accident", {"victims": [
{"mmHg": [120, 80], "bpm":50}], "coord": [30, 40]})
])
# instantiates the explanation manager
explainer = SimpleExplainer(broker, broker.history)
# Instantiates the explanation generator web graphical interface.
interface = WebGUI(explainer)
# Starts the web graphical interface server.
# The IP and port will be printed on the terminal.
# You must open the respective ip and port in the browser.
interface.server.run()
```
## References
<a id="1">[1]</a> C. Castelfranchi and F. Paglieri. The role of beliefs in goal
dynamics: prolegomena to a constructive theory of intentions. Synthese, 155,
237–263, 2007. doi: https://doi.org/10.1007/s11229-006-9156-3
<a id="2">[2]</a> H. Jasinski and C. Tacla Denerating contrastive explanations
for BDI-based goal selection. url:
http://repositorio.utfpr.edu.br/jspui/handle/1/29522
<a id="3">[3]</a> J. M. Simão, C. A. Tacla, P. C. Stadzisz and R. F.
Banaszewski, "Notification Oriented Paradigm (NOP) and Imperative Paradigm: A
Comparative Study," Journal of Software Engineering and Applications, Vol. 5 No.
6, 2012, pp. 402-416. doi: https://www.doi.org/10.4236/jsea.2012.56047
## About
Author: Henrique Emanoel Viana, a Brazilian computer scientist, enthusiast of
web technologies, cel: +55 (41) 99999-4664. URL:
https://sites.google.com/view/henriqueviana
Improvements and suggestions are welcome!
Raw data
{
"_id": null,
"home_page": null,
"name": "goalEDP",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.12",
"maintainer_email": null,
"keywords": "scientific, paradigms, agents, explainability",
"author": null,
"author_email": "Henrique Emanoel Viana <hv5088@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/7d/ec/ce603d65bc073a2cbce830064ea0a8a86ead789d5a3c43852e5c873ef638/goalEDP-0.0.7.tar.gz",
"platform": null,
"description": "# goalEDP\r\n\r\nLibrary for processing goal-oriented agents. Processing is done on an\r\nevent-driven mechanism. Events are saved to be used by explanation generating\r\nmethods.\r\n\r\n## Contents\r\n\r\n- [Sample application](#sample-application)\r\n- [References](#references)\r\n- [About](#about)\r\n\r\n## Sample application\r\n\r\nThis package is published in the official Python repository. To install, use:\r\n\r\n```\r\npython3 -m pip install goalEDP\r\n```\r\n\r\nApplication example:\r\n\r\n```python\r\nfrom goalEDP.extensions.goal import BeliefsReviewer, GoalStatusPromoter, Goal, Action, Agent, GoalBroker, Conflict\r\nfrom goalEDP.explainers.simple_explainer import SimpleExplainer\r\nfrom goalEDP.storages.in_memory import InMemoryHistory\r\nfrom goalEDP.core import Event\r\nfrom goalEDP.extensions.web_gui import WebGUI\r\n\r\n\r\n# Describe how the agent works:\r\n# The Agent has the following flow:\r\n# 1) revise beliefs\r\n# 2) promotes goals statuses\r\n# 3) detects conflicts\r\n# 4) pursues goals\r\n\r\nclass ReviewAccident(BeliefsReviewer):\r\n def __init__(self):\r\n # Call to superclass constructor\r\n super().__init__(\r\n desc=\"Review accident data\", attrs=[\"accident\"]) # Attributes can be external events, inputted or not by a simulator. They can also be events that represent communication between agents.\r\n\r\n async def reviewBeliefs(self):\r\n # The event queue is cleared each time the entity is processed. Maybe you want to save the beliefs in an object variable.\r\n beliefs = dict()\r\n if (\"accident\" in self.eventQueueByTopic):\r\n # \"-1\" to get the last one in queue\r\n # Use only JSON-serializable variables as belief values.\r\n # In fact, in this framework all event values \u200b\u200bmust be serializable to JSON.\r\n # If you want to use complex structured data, use compositions with Python's dict structure.\r\n beliefs[\"accident.coord\"] = self.eventQueueByTopic[\"accident\"][-1].value[\"coord\"]\r\n beliefs[\"accident.hasAtLeastTwoVictims\"] = len(\r\n self.eventQueueByTopic[\"accident\"][-1].value[\"victims\"]) >= 2\r\n beliefs[\"accident.severityIsHigh\"] = False\r\n for victim in self.eventQueueByTopic[\"accident\"][-1].value[\"victims\"]:\r\n if victim[\"bpm\"] < 60 or victim[\"bpm\"] > 100:\r\n beliefs[\"accident.severityIsHigh\"] = True\r\n if victim[\"mmHg\"][0] > 14 or victim[\"mmHg\"][1] < 6:\r\n beliefs[\"accident.severityIsHigh\"] = True\r\n return beliefs\r\n\r\n# All goal statuses must be True for it to be pursued\r\n\r\n\r\nclass RescuePromoter(GoalStatusPromoter):\r\n def __init__(self):\r\n super().__init__(\r\n desc=\"Promotes rescue goal statuses\", beliefs=[\"accident.hasAtLeastTwoVictims\", \"accident.severityIsHigh\"], promotionNames=[\"intention\"])\r\n\r\n async def promoteOrDemote(self):\r\n # The event queue is cleared each time the entity is processed. Maybe you want to save the promotions in an object variable.\r\n promotions = dict()\r\n if (\"accident.severityIsHigh\" in self.eventQueueByTopic) and (\"accident.hasAtLeastTwoVictims\" in self.eventQueueByTopic):\r\n if (self.eventQueueByTopic[\"accident.severityIsHigh\"][-1].value) or (self.eventQueueByTopic[\"accident.hasAtLeastTwoVictims\"][-1].value):\r\n promotions[\"intention\"] = True\r\n else:\r\n promotions[\"intention\"] = False\r\n return promotions\r\n\r\n\r\nclass RescueAction(Action):\r\n def __init__(self):\r\n super().__init__(\r\n desc=\"Action to rescue victims\", beliefs=[\"accident.coord\"])\r\n\r\n async def procedure(self) -> None:\r\n print(\"Rescue at: \" +\r\n str(self.eventQueueByTopic[\"accident.coord\"][-1].value))\r\n\r\n# Goals with highest priority are pursued first\r\n\r\n\r\nclass Rescue(Goal):\r\n def __init__(self):\r\n super().__init__(desc=\"Objective of rescuing victims of serious accidents\",\r\n promoter=RescuePromoter(), plan=[RescueAction()], priority=0)\r\n\r\n# If conflict is detected, the goal with the highest priority is chosen, with the others discarded.\r\n#\r\n# class ConflictOne(Conflict):\r\n# def __init__(self):\r\n# super().__init__(\r\n# desc=\"Recharge battery instead of saving victim\", conflictingGoals=[goalInstance1, goalInstance2])\r\n\r\n\r\nagent1 = Agent(desc=\"Robot that saves victims\", beliefsReviewers=[\r\n ReviewAccident()], goals=[Rescue()], conflicts=[])\r\n\r\n# Instantiates history to save events.\r\nhistory = InMemoryHistory()\r\n\r\n# Instantiates the event broker, which also controls the execution sequence\r\nbroker = GoalBroker(agents=[agent1], history=history)\r\n\r\nbroker.startProcess()\r\n\r\n# Enter external events, from a simulator for example.\r\nbroker.inputExternalEvents([\r\n Event(\"accident\", {\"victims\": [\r\n {\"mmHg\": [120, 80], \"bpm\":50}], \"coord\": [30, 40]})\r\n])\r\n\r\n# instantiates the explanation manager\r\nexplainer = SimpleExplainer(broker, broker.history)\r\n\r\n# Instantiates the explanation generator web graphical interface.\r\ninterface = WebGUI(explainer)\r\n\r\n# Starts the web graphical interface server.\r\n# The IP and port will be printed on the terminal.\r\n# You must open the respective ip and port in the browser.\r\ninterface.server.run()\r\n```\r\n\r\n## References\r\n\r\n<a id=\"1\">[1]</a> C. Castelfranchi and F. Paglieri. The role of beliefs in goal\r\ndynamics: prolegomena to a constructive theory of intentions. Synthese, 155,\r\n237\u2013263, 2007. doi: https://doi.org/10.1007/s11229-006-9156-3\r\n\r\n<a id=\"2\">[2]</a> H. Jasinski and C. Tacla Denerating contrastive explanations\r\nfor BDI-based goal selection. url:\r\nhttp://repositorio.utfpr.edu.br/jspui/handle/1/29522\r\n\r\n<a id=\"3\">[3]</a> J. M. Sim\u00e3o, C. A. Tacla, P. C. Stadzisz and R. F.\r\nBanaszewski, \"Notification Oriented Paradigm (NOP) and Imperative Paradigm: A\r\nComparative Study,\" Journal of Software Engineering and Applications, Vol. 5 No.\r\n6, 2012, pp. 402-416. doi: https://www.doi.org/10.4236/jsea.2012.56047\r\n\r\n## About\r\n\r\nAuthor: Henrique Emanoel Viana, a Brazilian computer scientist, enthusiast of\r\nweb technologies, cel: +55 (41) 99999-4664. URL:\r\nhttps://sites.google.com/view/henriqueviana\r\n\r\nImprovements and suggestions are welcome!\r\n",
"bugtrack_url": null,
"license": null,
"summary": "Library for processing goal-oriented agents. Processing is done on an event-driven mechanism. Events are saved to be used by explanation generating methods.",
"version": "0.0.7",
"project_urls": {
"Homepage": "https://github.com/hviana/goalEDP"
},
"split_keywords": [
"scientific",
" paradigms",
" agents",
" explainability"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "15190bc62b4f60c5dd4731dd6533ef0a9e7ddfe6bf8491176f3f690acc46894f",
"md5": "0c25f2f5daddba3a1294b1f1a129d21d",
"sha256": "7225a5d799dd60012206a3bbc92b5b358afe1382fd79517c067c912237ee2461"
},
"downloads": -1,
"filename": "goalEDP-0.0.7-py3-none-any.whl",
"has_sig": false,
"md5_digest": "0c25f2f5daddba3a1294b1f1a129d21d",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.12",
"size": 20522,
"upload_time": "2024-03-20T18:58:06",
"upload_time_iso_8601": "2024-03-20T18:58:06.510532Z",
"url": "https://files.pythonhosted.org/packages/15/19/0bc62b4f60c5dd4731dd6533ef0a9e7ddfe6bf8491176f3f690acc46894f/goalEDP-0.0.7-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "7decce603d65bc073a2cbce830064ea0a8a86ead789d5a3c43852e5c873ef638",
"md5": "ca56e1421fe9a35d0d912e2ba582930a",
"sha256": "91b86f0c07b99a1b4096652907fbc9f156852c485f0e6f23e7b28efae9bd2467"
},
"downloads": -1,
"filename": "goalEDP-0.0.7.tar.gz",
"has_sig": false,
"md5_digest": "ca56e1421fe9a35d0d912e2ba582930a",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.12",
"size": 21267,
"upload_time": "2024-03-20T18:58:08",
"upload_time_iso_8601": "2024-03-20T18:58:08.564555Z",
"url": "https://files.pythonhosted.org/packages/7d/ec/ce603d65bc073a2cbce830064ea0a8a86ead789d5a3c43852e5c873ef638/goalEDP-0.0.7.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-03-20 18:58:08",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "hviana",
"github_project": "goalEDP",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "goaledp"
}
