55 rewrite interface to enable non ai plugins

onair/config/default_config.ini

@@ -4,7 +4,11 @@ TelemetryFile = 700_crash_to_earth_1.csv
 TelemetryMetadataFilePath = onair/data/telemetry_configs/
 MetaFile = data_physics_generation_CONFIG.json
 ParserFileName = onair/data_handling/csv_parser.py
-PluginList = {'generic_plugin':'plugins/generic/generic_plugin.py'}
+
+KnowledgeRepPluginDict = {'generic':'plugins/generic/__init__.py'}
+LearnersPluginDict = {'generic':'plugins/generic/__init__.py'}
+PlannersPluginDict = {'generic':'plugins/generic/__init__.py'}
+ComplexPluginDict = {'generic':'plugins/generic/__init__.py'}
 
 [RUN_FLAGS]
 IO_Flag = true

onair/config/redis_example.ini

@@ -4,7 +4,11 @@ TelemetryFile = 700_crash_to_earth_1.csv
 TelemetryMetadataFilePath = onair/data/telemetry_configs/
 MetaFile = data_physics_generation_CONFIG.json
 ParserFileName = onair/data_handling/redis_adapter.py
-PluginList = {'generic_plugin':'plugins/generic/generic_plugin.py'}
+
+KnowledgeRepPluginDict = {'generic':'plugins/generic/__init__.py'}
+LearnersPluginDict = {'generic':'plugins/generic/__init__.py'}
+PlannersPluginDict = {'generic':'plugins/generic/__init__.py'}
+ComplexPluginDict = {'generic':'plugins/generic/__init__.py'}
 
 [RUN_FLAGS]
 IO_Flag = true

onair/src/ai_components/ai_plugin_abstract/core.py

@@ -36,7 +36,7 @@ def apriori_training(self, batch_data=[]):
         raise NotImplementedError
 
     @abstractmethod
-    def update(self, frame=[]):
+    def update(self, low_level_data=[], high_level_data={}):
         """
         Given streamed data point, system should update internally
         """

onair/src/ai_components/learners_interface.py

@@ -10,30 +10,22 @@
 """
 Data driven learning class for managing all data driven AI components
 """
-import importlib.util
-
+from ..util.plugin_import import import_plugins
 from ..util.data_conversion import *
 
 class LearnersInterface:
-    def __init__(self, headers, _ai_plugins={}):
+    def __init__(self, headers, _learner_plugins={}):
         assert(len(headers)>0), 'Headers are required'
         self.headers = headers
-        self.ai_constructs = []
-        for module_name in list(_ai_plugins.keys()):
-            spec = importlib.util.spec_from_file_location(module_name, _ai_plugins[module_name])
-            module = importlib.util.module_from_spec(spec)
-            spec.loader.exec_module(module)
-            self.ai_constructs.append(module.Plugin(module_name,headers))
-
+        self.ai_constructs = import_plugins(self.headers, _learner_plugins)
+
     def apriori_training(self, batch_data):
         for plugin in self.ai_constructs:
             plugin.apriori_training(batch_data)
 
-    def update(self, curr_data, status):
-        input_data = curr_data
-        output_data = status_to_oneHot(status)
+    def update(self, low_level_data, high_level_data):
         for plugin in self.ai_constructs:
-            plugin.update(input_data)
+            plugin.update(low_level_data)
 
     def check_for_salient_event(self):
         pass

onair/src/ai_components/planners_interface.py

@@ -10,22 +10,16 @@
 """
 Planners interface class for managing all planning-based AI components
 """
-import importlib.util
-
+from ..util.plugin_import import import_plugins
 from ..util.data_conversion import *
 
 class PlannersInterface:
-    def __init__(self, headers, _ai_plugins={}):
+    def __init__(self, headers, _planner_plugins={}):
         assert(len(headers)>0), 'Headers are required'
         self.headers = headers
-        self.ai_constructs = []
-        for module_name in list(_ai_plugins.keys()):
-            spec = importlib.util.spec_from_file_location(module_name, _ai_plugins[module_name])
-            module = importlib.util.module_from_spec(spec)
-            spec.loader.exec_module(module)
-            self.ai_constructs.append(module.Plugin(module_name,headers))
+        self.ai_constructs = import_plugins(self.headers,_planner_plugins)
 
-    def update(self, curr_raw_tlm, status):
+    def update(self, high_level_data):
         # Raw TLM should be transformed into high-leve state representation here 
         # Can store something as stale unless a planning thread is launched 
         pass

onair/src/reasoning/agent.py

@@ -13,30 +13,37 @@
 """
 from ..ai_components.learners_interface import LearnersInterface
 from ..ai_components.planners_interface import PlannersInterface
+from ..reasoning.complex_reasoning_interface import ComplexReasoningInterface
 from ..reasoning.diagnosis import Diagnosis
 
 class Agent:
-    def __init__(self, vehicle, plugin_list):
+    def __init__(self, vehicle, learners_plugin_dict, planners_plugin_dict, complex_plugin_dict):
+
         self.vehicle_rep = vehicle
         self.mission_status = self.vehicle_rep.get_status()
         self.bayesian_status = self.vehicle_rep.get_bayesian_status()
 
         # AI Interfaces
-        self.learning_systems = LearnersInterface(self.vehicle_rep.get_headers(),plugin_list)
-        self.planning_systems = PlannersInterface(self.vehicle_rep.get_headers(),plugin_list)
+        self.learning_systems = LearnersInterface(self.vehicle_rep.get_headers(),learners_plugin_dict)
+        self.planning_systems = PlannersInterface(self.vehicle_rep.get_headers(),planners_plugin_dict)
+        self.complex_reasoning_systems = ComplexReasoningInterface(self.vehicle_rep.get_headers(),complex_plugin_dict)
 
-    # Markov Assumption holds 
-    def reason(self, frame):
-        # Update with new telemetry 
-        self.vehicle_rep.update(frame)
-        self.mission_status = self.vehicle_rep.get_status() 
-        self.learning_systems.update(frame, self.mission_status)
-        self.planning_systems.update(frame, self.mission_status)
+    def render_reasoning(self):
+        return self.complex_reasoning_systems.render_reasoning()
 
-        # Check for a salient event, needing acionable outcome
-        self.learning_systems.check_for_salient_event()
-        self.planning_systems.check_for_salient_event()
+    def reason(self, frame):
+        self.vehicle_rep.update(frame) 
+        self.learning_systems.update(self.vehicle_rep.curr_data, self.vehicle_rep.get_state_information())
+        self.planning_systems.update(self.vehicle_rep.get_state_information()) 
+
+        aggregate_high_level_info = {'vehicle_rep' : self.vehicle_rep.get_state_information(),
+                                     'learning_systems' : self.learning_systems.render_reasoning(),
+                                     'planning_systems' : self.planning_systems.render_reasoning()}
 
+        self.complex_reasoning_systems.update(aggregate_high_level_info)
+
+        return self.render_reasoning()
+
     def diagnose(self, time_step):
         """ Grab the mnemonics from the """
         learning_system_results = self.learning_systems.render_reasoning()
@@ -45,4 +52,3 @@ def diagnose(self, time_step):
                               self.bayesian_status,
                               self.vehicle_rep.get_current_faulting_mnemonics())
         return diagnosis.perform_diagnosis()
-

onair/src/reasoning/complex_reasoning_interface.py

@@ -0,0 +1,34 @@
+# GSC-19165-1, "The On-Board Artificial Intelligence Research (OnAIR) Platform"
+#
+# Copyright © 2023 United States Government as represented by the Administrator of
+# the National Aeronautics and Space Administration. No copyright is claimed in the
+# United States under Title 17, U.S. Code. All Other Rights Reserved.
+#
+# Licensed under the NASA Open Source Agreement version 1.3
+# See "NOSA GSC-19165-1 OnAIR.pdf"
+
+"""
+Reasoning interface class for managing all complex custom reasoning components
+"""
+
+from ..util.data_conversion import *
+from ..util.plugin_import import import_plugins
+
+class ComplexReasoningInterface:
+    def __init__(self, headers, _reasoning_plugins={}):
+        assert(len(headers)>0), 'Headers are required'
+        self.headers = headers
+        self.reasoning_constructs = import_plugins(self.headers,_reasoning_plugins)
+
+    def update(self, high_level_data):
+        for plugin in self.reasoning_constructs:
+            plugin.update(high_level_data=high_level_data)
+
+    def check_for_salient_event(self):
+        pass
+
+    def render_reasoning(self):
+        intelligent_outcomes = {}
+        for plugin in self.reasoning_constructs:
+            intelligent_outcomes[plugin.component_name] = plugin.render_reasoning()
+        return intelligent_outcomes

onair/src/run_scripts/execution_engine.py

@@ -26,6 +26,7 @@ def __init__(self, config_file='', run_name='', save_flag=False):
 
         # Init Housekeeping 
         self.run_name = run_name
+        self.config_filepath = config_file
 
         # Init Flags 
         self.IO_Flag = False
@@ -49,7 +50,10 @@ def __init__(self, config_file='', run_name='', save_flag=False):
         self.sim = None
 
         # Init plugins
-        self.plugin_list = ['']
+        self.knowledge_rep_plugin_dict = ['']
+        self.learners_plugin_dict = ['']
+        self.planners_plugin_dict = ['']
+        self.complex_plugin_dict = ['']
 
         self.save_flag = save_flag
         self.save_name = run_name
@@ -78,25 +82,21 @@ def parse_configs(self, config_filepath):
             ## Parse Required Data: Names
             self.parser_file_name = config['DEFAULT']['ParserFileName']
 
-            ## Parse Required Data: Plugin name to path dict
-            config_plugin_list = config['DEFAULT']['PluginList']
-            ast_plugin_list = self.ast_parse_eval(config_plugin_list)
-            if isinstance(ast_plugin_list.body, ast.Dict) and len(ast_plugin_list.body.keys) > 0:
-                temp_plugin_list = ast.literal_eval(config_plugin_list)
-            else:
-                raise ValueError(f"{config_plugin_list} is an invalid PluginList. It must be a dict of at least 1 key/value pair.")
-            for plugin_name in temp_plugin_list.values():
-                if not(os.path.exists(plugin_name)):
-                    raise FileNotFoundError(f"In config file '{config_filepath}', path '{plugin_name}' does not exist or is formatted incorrectly.")
-            self.plugin_list = temp_plugin_list
+            ## Parse Required Data: Plugins
+            self.knowledge_rep_plugin_dict = self.parse_plugins_dict(config['DEFAULT']['KnowledgeRepPluginDict'])
+            self.learners_plugin_dict = self.parse_plugins_dict(config['DEFAULT']['LearnersPluginDict'])
+            self.planners_plugin_dict = self.parse_plugins_dict(config['DEFAULT']['PlannersPluginDict'])
+            self.complex_plugin_dict = self.parse_plugins_dict(config['DEFAULT']['ComplexPluginDict'])
+
+            ## Parse Optional Data: Flags
+            ## 'RUN_FLAGS' must exist, but individual flags return False if missing
+            self.IO_Flag = config['RUN_FLAGS'].getboolean('IO_Flag')
+            self.Dev_Flag = config['RUN_FLAGS'].getboolean('Dev_Flag')
+            self.Viz_Flag = config['RUN_FLAGS'].getboolean('Viz_Flag')
+
         except KeyError as e:
             new_message = f"Config file: '{config_filepath}', missing key: {e.args[0]}"
             raise KeyError(new_message) from e
-
-        ## Parse Optional Data: Flags
-        self.IO_Flag = config['RUN_FLAGS'].getboolean('IO_Flag')
-        self.Dev_Flag = config['RUN_FLAGS'].getboolean('Dev_Flag')
-        self.Viz_Flag = config['RUN_FLAGS'].getboolean('Viz_Flag')
 
         ## Parse Optional Data: Benchmarks
         try:
@@ -106,14 +106,31 @@ def parse_configs(self, config_filepath):
         except:
             pass
 
+    def parse_plugins_dict(self, config_plugin_dict): 
+        ## Parse Required Data: Plugin name to path dict
+        ast_plugin_dict = self.ast_parse_eval(config_plugin_dict)
+        if isinstance(ast_plugin_dict.body, ast.Dict):
+            temp_plugin_dict = ast.literal_eval(config_plugin_dict)
+        else:
+            raise ValueError(f"Plugin dict {config_plugin_dict} from {self.config_filepath} is invalid. It must be a dict.")
+
+        for plugin_file in temp_plugin_dict.values():
+            if not(os.path.exists(plugin_file)):
+                raise FileNotFoundError(f"In config file '{self.config_filepath}' Plugin path '{plugin_file}' does not exist.")
+        return temp_plugin_dict
+
     def parse_data(self, parser_file_name, data_file_name, metadata_file_name, subsystems_breakdown=False):
         data_source_spec = importlib.util.spec_from_file_location('data_source', parser_file_name)
         data_source_module = importlib.util.module_from_spec(data_source_spec)
         data_source_spec.loader.exec_module(data_source_module)
         self.simDataParser = data_source_module.DataSource(data_file_name, metadata_file_name, subsystems_breakdown)
 
     def setup_sim(self):
-        self.sim = Simulator(self.simDataParser, self.plugin_list)
+        self.sim = Simulator(self.simDataParser, 
+                             self.knowledge_rep_plugin_dict, 
+                             self.learners_plugin_dict,
+                             self.planners_plugin_dict,
+                             self.complex_plugin_dict)
         try:
             fls = ast.literal_eval(self.benchmarkFiles)
             fp = os.path.dirname(os.path.realpath(__file__)) + '/../..' + self.benchmarkFilePath

onair/src/run_scripts/sim.py

@@ -24,12 +24,11 @@
 DIAGNOSIS_INTERVAL = 100
 
 class Simulator:
-    def __init__(self, dataParser, plugin_list):
+    def __init__(self, dataParser, knowledge_rep_plugin_dict, learners_plugin_dict, planners_plugin_dict, complex_plugin_dict):
         self.simData = dataParser
-
         headers, tests = dataParser.get_vehicle_metadata()
-        vehicle = VehicleRepresentation(headers, tests)
-        self.agent = Agent(vehicle, plugin_list)
+        vehicle = VehicleRepresentation(headers, tests, knowledge_rep_plugin_dict)
+        self.agent = Agent(vehicle, learners_plugin_dict, planners_plugin_dict, complex_plugin_dict)
 
     def run_sim(self, IO_Flag=False, dev_flag=False, viz_flag = True):
         if IO_Flag == True: print_sim_header()

onair/src/systems/vehicle_rep.py

@@ -16,22 +16,36 @@
 from .telemetry_test_suite import TelemetryTestSuite
 
 from ..util.print_io import *
+from ..util.plugin_import import import_plugins
+# from ..util.data_conversion import *
 
 class VehicleRepresentation:
-    def __init__(self, headers, tests):
+    def __init__(self, headers, tests, _knowledge_rep_plugins={}):
         assert(len(headers) == len(tests))
-        self.status = Status('MISSION')
         self.headers = headers
+        self.knowledge_synthesis_constructs = import_plugins(self.headers,_knowledge_rep_plugins)
+
+        self.status = Status('MISSION')
         self.test_suite = TelemetryTestSuite(headers, tests)
-        self.curr_data = ['-']* len(self.headers)
+
+        self.curr_data = ['-']* len(self.headers) #stale data
 
     ##### UPDATERS #################################
     def update(self, frame):
+        # Update constructs
+        self.update_curr_data(frame)
+        self.test_suite.execute_suite(self.curr_data)
+        self.status.set_status(*self.test_suite.get_suite_status())
+        self.update_constructs(self.curr_data)
+
+    def update_constructs(self, frame):
+        for construct in self.knowledge_synthesis_constructs:
+            construct.update(frame)
+
+    def update_curr_data(self, frame):
         for i in range(len(frame)):
             if frame[i] != '-':
                 self.curr_data[i] = frame[i]
-        self.test_suite.execute_suite(frame)
-        self.status.set_status(*self.test_suite.get_suite_status())
 
     ##### GETTERS AND SETTERS #####
     def get_headers(self):
@@ -55,4 +69,10 @@ def get_bayesian_status(self):
     def get_batch_status_reports(self, batch_data):
         return
 
+    def get_state_information(self, scope=['status']):
+        state_info = {}
+        for construct in self.knowledge_synthesis_constructs:
+            state_info[construct.component_name] = construct.render_reasoning()
+        return state_info
+
 

onair/src/util/plugin_import.py

@@ -0,0 +1,27 @@
+# GSC-19165-1, "The On-Board Artificial Intelligence Research (OnAIR) Platform"
+#
+# Copyright © 2023 United States Government as represented by the Administrator of
+# the National Aeronautics and Space Administration. No copyright is claimed in the
+# United States under Title 17, U.S. Code. All Other Rights Reserved.
+#
+# Licensed under the NASA Open Source Agreement version 1.3
+# See "NOSA GSC-19165-1 OnAIR.pdf"
+
+"""
+plugin_import.py
+Function to import user-specified plugins (from config files) into interfaces
+"""
+
+import importlib.util
+import sys
+
+def import_plugins(headers, module_dict):
+    plugin_list = []
+    for module_name in list(module_dict.keys()):
+                spec = importlib.util.spec_from_file_location(module_name, module_dict[module_name])
+                module = importlib.util.module_from_spec(spec)
+                spec.loader.exec_module(module)
+                sys.modules[module_name] = module         
+                plugin = __import__(f'{module_name}.{module_name}_plugin', fromlist=[f'{module_name}_plugin'])            
+                plugin_list.append(plugin.Plugin(module_name,headers))
+    return(plugin_list)

plugins/generic/__init__.py

@@ -1 +0,0 @@
-from .generic_plugin import Plugin

plugins/generic/generic_plugin.py

@@ -17,7 +17,7 @@ def apriori_training(self,batch_data=[]):
         """
         pass
 
-    def update(self,frame=[]):
+    def update(self,low_level_data=[], high_level_data={}):
         """
         Given streamed data point, system should update internally
         """