Merge pull request #78 from hsj576/main

Merge of feature-lifelong-n branch

README_ospp.md

@@ -0,0 +1,3 @@
+# OSPP 
+I changed the sedna source code to implement my algorithm.   
+Please turn to https://github.com/kubeedge/sedna/pull/378 and https://github.com/nailtu30/sedna/blob/ospp-final/README_ospp.md for more information.

core/common/constant.py

@@ -34,21 +34,46 @@ class ParadigmType(Enum):
     SINGLE_TASK_LEARNING = "singletasklearning"
     INCREMENTAL_LEARNING = "incrementallearning"
     MULTIEDGE_INFERENCE = "multiedgeinference"
+    LIFELONG_LEARNING = "lifelonglearning"
 
 
 class ModuleType(Enum):
     """
     Algorithm module type.
     """
     BASEMODEL = "basemodel"
+
+    # HEM
     HARD_EXAMPLE_MINING = "hard_example_mining"
 
+    # STP
+    TASK_DEFINITION = "task_definition"
+    TASK_RELATIONSHIP_DISCOVERY = "task_relationship_discovery"
+    TASK_ALLOCATION = "task_allocation"
+    TASK_REMODELING = "task_remodeling"
+    INFERENCE_INTEGRATE = "inference_integrate"
+
+    # KM
+    TASK_UPDATE_DECISION = "task_update_decision"
+
+    # UTP
+    UNSEEN_TASK_ALLOCATION = "unseen_task_allocation"
+
+    # UTD
+    UNSEEN_SAMPLE_RECOGNITION = "unseen_sample_recognition"
+    UNSEEN_SAMPLE_RE_RECOGNITION = "unseen_sample_re_recognition"
+
 
 class SystemMetricType(Enum):
     """
     System metric type of ianvs.
     """
+    # pylint: disable=C0103
     SAMPLES_TRANSFER_RATIO = "samples_transfer_ratio"
+    FWT = "FWT"
+    BWT = "BWT"
+    Task_Avg_Acc = "Task_Avg_Acc"
+    Matrix = "Matrix"
 
 class TestObjectType(Enum):
     """

core/storymanager/rank/rank.py

@@ -21,7 +21,7 @@
 import pandas as pd
 
 from core.common import utils
-from core.storymanager.visualization import get_visualization_func
+from core.storymanager.visualization import get_visualization_func, draw_heatmap_picture
 
 
 # pylint: disable=R0902
@@ -107,7 +107,10 @@ def _get_all_module_types(cls, test_cases) -> list:
     def _get_algorithm_hyperparameters(cls, algorithm):
         hps = {}
         for module in algorithm.modules.values():
-            hps.update(**module.hyperparameters)
+            for name, value in module.hyperparameters.items():
+                name = f"{module.type}-{name}"
+                value = str(value)
+                hps.update({name: value})
         return hps
 
     def _get_all_hps_names(self, test_cases) -> list:
@@ -170,6 +173,7 @@ def _get_all(self, test_cases, test_results) -> pd.DataFrame:
         return self._sort_all_df(all_df, self._get_all_metric_names(test_results))
 
     def _save_all(self):
+        # pylint: disable=E1101
         all_df = copy.deepcopy(self.all_df)
         all_df.index = pd.np.arange(1, len(all_df) + 1)
         all_df.to_csv(self.all_rank_file, index_label="rank", encoding="utf-8", sep=" ")
@@ -199,10 +203,21 @@ def _get_selected(self, test_cases, test_results) -> pd.DataFrame:
         return selected_df
 
     def _save_selected(self, test_cases, test_results):
+        # pylint: disable=E1101
         selected_df = self._get_selected(test_cases, test_results)
         selected_df.index = pd.np.arange(1, len(selected_df) + 1)
         selected_df.to_csv(self.selected_rank_file, index_label="rank", encoding="utf-8", sep=" ")
 
+    def _draw_pictures(self, test_cases, test_results):
+        # pylint: disable=E1101
+        for test_case in test_cases:
+            out_put = test_case.output_dir
+            test_result = test_results[test_case.id][0]
+            matrix = test_result.get('Matrix')
+            #print(out_put)
+            for key in matrix.keys():
+                draw_heatmap_picture(out_put, key, matrix[key])
+
     def _prepare(self, test_cases, test_results, output_dir):
         all_metric_names = self._get_all_metric_names(test_results)
         all_hps_names = self._get_all_hps_names(test_cases)
@@ -241,6 +256,11 @@ def save(self, test_cases, test_results, output_dir):
         if self.save_mode == "selected_only":
             self._save_selected(test_cases, test_results)
 
+        if self.save_mode == "selected_and_all_and_picture":
+            self._save_all()
+            self._save_selected(test_cases, test_results)
+            self._draw_pictures(test_cases, test_results)
+
     def plot(self):
         """
         plot rank according to the visual method, include

core/storymanager/visualization/__init__.py

@@ -13,4 +13,4 @@
 # limitations under the License.
 
 # pylint: disable=missing-module-docstring
-from .visualization import get_visualization_func
+from .visualization import get_visualization_func, draw_heatmap_picture

core/storymanager/visualization/visualization.py

@@ -15,7 +15,8 @@
 """Visualization"""
 
 import sys
-
+import os
+import matplotlib.pyplot as plt
 from prettytable import from_csv
 
 
@@ -25,6 +26,23 @@ def print_table(rank_file):
         table = from_csv(file)
         print(table)
 
+def draw_heatmap_picture(output, title, matrix):
+    """
+    draw heatmap for results
+    """
+    plt.figure(figsize=(10, 8), dpi=80)
+    plt.imshow(matrix, cmap='bwr', extent=(0.5, len(matrix)+0.5, 0.5, len(matrix)+0.5),
+                    origin='lower')
+    plt.xticks(fontsize=15)
+    plt.yticks(fontsize=15)
+    plt.xlabel('task round', fontsize=15)
+    plt.ylabel('task', fontsize=15)
+    plt.title(title, fontsize=15)
+    plt.colorbar(format='%.2f')
+    output_dir = os.path.join(output, f"output/{title}-heatmap.png")
+    #print(output_dir)
+    plt.savefig(output_dir)
+    plt.show()
 
 def get_visualization_func(mode):
     """ get visualization func """

core/testcasecontroller/algorithm/algorithm.py

@@ -17,17 +17,19 @@
 import copy
 
 from core.common.constant import ParadigmType
+from core.common.utils import load_module
 from core.testcasecontroller.algorithm.module import Module
 from core.testcasecontroller.algorithm.paradigm import (
     SingleTaskLearning,
     IncrementalLearning,
     MultiedgeInference,
+    LifelongLearning,
 )
 from core.testcasecontroller.generation_assistant import get_full_combinations
 
-
-# pylint: disable=too-few-public-methods
 class Algorithm:
+    # pylint: disable=too-many-instance-attributes
+    # pylint: disable=too-few-public-methods
     """
     Algorithm: typical distributed-synergy AI algorithm paradigm.
     Notes:
@@ -53,14 +55,20 @@ class Algorithm:
     def __init__(self, name, config):
         self.name = name
         self.paradigm_type: str = ""
+        self.third_party_packages: list = []
         self.incremental_learning_data_setting: dict = {
             "train_ratio": 0.8,
             "splitting_method": "default"
         }
+        self.lifelong_learning_data_setting: dict = {
+            "train_ratio": 0.8,
+            "splitting_method": "default"
+        }
         self.initial_model_url: str = ""
         self.modules: list = []
         self.modules_list = None
         self._parse_config(config)
+        self._load_third_party_packages()
 
     def paradigm(self, workspace: str, **kwargs):
         """
@@ -93,6 +101,9 @@ def paradigm(self, workspace: str, **kwargs):
         if self.paradigm_type == ParadigmType.MULTIEDGE_INFERENCE.value:
             return MultiedgeInference(workspace, **config)
 
+        if self.paradigm_type == ParadigmType.LIFELONG_LEARNING.value:
+            return LifelongLearning(workspace, **config)
+
         return None
 
     def _check_fields(self):
@@ -113,6 +124,11 @@ def _check_fields(self):
                 f"algorithm incremental_learning_data_setting"
                 f"({self.incremental_learning_data_setting} must be dictionary type.")
 
+        if not isinstance(self.lifelong_learning_data_setting, dict):
+            raise ValueError(
+                f"algorithm lifelong_learning_data_setting"
+                f"({self.lifelong_learning_data_setting} must be dictionary type.")
+
         if not isinstance(self.initial_model_url, str):
             raise ValueError(
                 f"algorithm initial_model_url({self.initial_model_url}) must be string type.")
@@ -138,7 +154,7 @@ def _parse_modules_config(cls, config):
         for module in modules:
             hps_list = module.hyperparameters_list
             if not hps_list:
-                modules_list.append((module.type, None))
+                modules_list.append((module.type, [module]))
                 continue
 
             module_list = []
@@ -152,3 +168,16 @@ def _parse_modules_config(cls, config):
         module_combinations_list = get_full_combinations(modules_list)
 
         return module_combinations_list
+
+    def _load_third_party_packages(self):
+        if len(self.third_party_packages) == 0:
+            return
+
+        for package in self.third_party_packages:
+            name = package["name"]
+            url = package["url"]
+            try:
+                load_module(url)
+            except Exception as err:
+                raise RuntimeError(f"load third party packages(name={name}, url={url}) failed,"
+                                f" error: {err}.") from err

core/testcasecontroller/algorithm/module/module.py

@@ -23,6 +23,7 @@
 from core.testcasecontroller.generation_assistant import get_full_combinations
 
 
+# pylint: disable=too-few-public-methods
 class Module:
     """
     Algorithm Module:
@@ -52,8 +53,8 @@ def __init__(self, config):
         self.type: str = ""
         self.name: str = ""
         self.url: str = ""
-        self.hyperparameters = None
-        self.hyperparameters_list = None
+        self.hyperparameters = {}
+        self.hyperparameters_list = []
         self._parse_config(config)
 
     def _check_fields(self):
@@ -71,75 +72,59 @@ def _check_fields(self):
         if not isinstance(self.url, str):
             raise ValueError(f"module url({self.url}) must be string type.")
 
-    def basemodel_func(self):
+    def get_module_instance(self, module_type):
         """
-        get basemodel module function of the module.
+        get function of algorithm module by using module type
+
+        Parameters
+        ---------
+        module_type: string
+            module type, e.g.: basemodel, hard_example_mining, etc.
 
         Returns
-        --------
+        ------
         function
 
         """
+        class_factory_type = ClassType.GENERAL
+        if module_type in [ModuleType.HARD_EXAMPLE_MINING.value]:
+            class_factory_type = ClassType.HEM
 
-        if not self.url:
-            raise ValueError(f"url({self.url}) of basemodel module must be provided.")
+        elif module_type in [ModuleType.TASK_DEFINITION.value,
+                             ModuleType.TASK_RELATIONSHIP_DISCOVERY.value,
+                             ModuleType.TASK_REMODELING.value,
+                             ModuleType.TASK_ALLOCATION.value,
+                             ModuleType.INFERENCE_INTEGRATE.value]:
+            class_factory_type = ClassType.STP
 
-        try:
-            utils.load_module(self.url)
-            # pylint: disable=E1134
-            basemodel = ClassFactory.get_cls(type_name=ClassType.GENERAL,
-                                             t_cls_name=self.name)(**self.hyperparameters)
-        except Exception as err:
-            raise RuntimeError(f"basemodel module loads class(name={self.name}) failed, "
-                            f"error: {err}.") from err
+        elif module_type in [ModuleType.TASK_UPDATE_DECISION.value]:
+            class_factory_type = ClassType.KM
 
-        return basemodel
+        elif module_type in [ModuleType.UNSEEN_TASK_ALLOCATION.value]:
+            class_factory_type = ClassType.UTP
 
-    def hard_example_mining_func(self):
-        """
-        get hard example mining function of the module.
-
-        Returns:
-        --------
-        function
-
-        """
+        elif module_type in [ModuleType.UNSEEN_SAMPLE_RECOGNITION.value,
+                             ModuleType.UNSEEN_SAMPLE_RE_RECOGNITION.value]:
+            class_factory_type = ClassType.UTD
 
         if self.url:
             try:
                 utils.load_module(self.url)
                 # pylint: disable=E1134
                 func = ClassFactory.get_cls(
-                    type_name=ClassType.HEM, t_cls_name=self.name)(**self.hyperparameters)
+                    type_name=class_factory_type, t_cls_name=self.name)(**self.hyperparameters)
 
                 return func
             except Exception as err:
-                raise RuntimeError(f"hard_example_mining module loads class"
-                                f"(name={self.name}) failed, error: {err}.") from err
+                raise RuntimeError(f"module(type={module_type} loads class(name={self.name}) "
+                                f"failed, error: {err}.") from err
 
-        # call built-in hard example mining function
-        hard_example_mining = {"method": self.name}
+        # call lib built-in module function
+        module_func = {"method": self.name}
         if self.hyperparameters:
-            hard_example_mining["param"] = self.hyperparameters
-
-        return hard_example_mining
+            module_func["param"] = self.hyperparameters
 
-    def get_module_func(self, module_type):
-        """
-        get function of algorithm module by using module type
-
-        Parameters
-        ---------
-        module_type: string
-            module type, e.g.: basemodel, hard_example_mining, etc.
-
-        Returns
-        ------
-        function
-
-        """
-        func_name = f"{module_type}_func"
-        return getattr(self, func_name)
+        return module_func
 
     def _parse_config(self, config):
         # pylint: disable=C0103

core/testcasecontroller/algorithm/paradigm/__init__.py

@@ -16,3 +16,4 @@
 from .incremental_learning import IncrementalLearning
 from .singletask_learning import SingleTaskLearning
 from .multiedge_inference import MultiedgeInference
+from .lifelong_learning import LifelongLearning