Merge branch 'dev' of git.nju.edu.cn:learnware/learnware-market into dev

examples/example_market_db/example_db.py

@@ -9,11 +9,12 @@ import joblib
 import numpy as np
 import os
 
-# database_ops.load_market_from_db()
-
 
 def prepare_learnware(learnware_num=10):
     for i in range(learnware_num):
+        dir_path = f"./learnware_pool/svm{i}"
+        os.makedirs(dir_path, exist_ok=True)
+
         print("Preparing Learnware: %d" % (i))
         data_X = np.random.randn(5000, 20)
         data_y = np.random.randn(5000)
@@ -21,28 +22,34 @@ def prepare_learnware(learnware_num=10):
 
         clf = svm.SVC(kernel="linear")
         clf.fit(data_X, data_y)
-        joblib.dump(clf, "./svm/svm_%d.pkl" % (i))
+        joblib.dump(clf, os.path.join(dir_path, "svm.pkl"))
 
         spec = specification.utils.generate_rkme_spec(X=data_X, gamma=0.1, cuda_idx=0)
-        spec.save("./svm/spec_%d.json" % (i))
+        spec.save(os.path.join(dir_path, "spec.json"))
+
+        init_file = os.path.join(dir_path, "__init__.py")
+        os.system(f"cp example_init.py {init_file}")
 
 
 def test_market():
     easy_market = EasyMarket()
     print("Total Item:", len(easy_market))
-    root_path = "./svm"
-    os.makedirs(root_path, exist_ok=True)
     test_learnware_num = 10
     prepare_learnware(test_learnware_num)
+    root_path = "./learnware_pool"
+    os.makedirs(root_path, exist_ok=True)
+
     for i in range(test_learnware_num):
-        model_path = os.path.join(root_path, "svm_%d.pkl" % (i))
-        stat_spec_path = os.path.join(root_path, "spec_%d.json" % (i))
+        dir_path = f"./learnware_pool/svm{i}"
+        model_path = os.path.join(dir_path, "__init__.py")
+        stat_spec_path = os.path.join(dir_path, "spec.json")
         easy_market.add_learnware(
             "learnware_%d" % (i), model_path, stat_spec_path, {"desc": "test_learnware_number_%d" % (i)}
         )
     print("Total Item:", len(easy_market))
     curr_inds = easy_market._get_ids()
     print("Available ids:", curr_inds)
+
     easy_market.delete_learnware(curr_inds[4])
     easy_market.delete_learnware(curr_inds[8])
     curr_inds = easy_market._get_ids()

examples/example_market_db/example_init.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm0/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm1/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm2/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm3/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm4/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm5/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm6/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm7/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm8/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

examples/example_market_db/learnware_pool/svm9/__init__.py

@@ -0,0 +1,19 @@
+import os
+import joblib
+import numpy as np
+from learnware.model import BaseModel
+
+
+class Model(BaseModel):
+    def __init__(self):
+        dir_path = os.path.dirname(os.path.abspath(__file__))
+        self.model = joblib.load(os.path.join(dir_path, "svm.pkl"))
+
+    def fit(self, X: np.ndarray, y: np.ndarray):
+        pass
+
+    def predict(self, X: np.ndarray) -> np.ndarray:
+        return self.model.predict(X)
+
+    def fintune(self, X: np.ndarray, y: np.ndarray):
+        pass

learnware/learnware/base.py

@@ -8,7 +8,7 @@ from ..utils import get_module_by_module_path
 
 
 class Learnware:
-    def __init__(self, id: str, name: str, model: Union[BaseModel, dict], specification: Specification):
+    def __init__(self, id: str, name: str, model: Union[BaseModel, str], specification: Specification):
         self.id = id
         self.name = name
         self.model = self._import_model(model)
@@ -42,8 +42,6 @@ class Learnware:
         elif isinstance(model, dict):
             model_module = get_module_by_module_path(model["module_path"])
             return getattr(model_module, model["class_name"])()
-        elif isinstance(model, str):
-            return model  # For test purpose
         else:
             raise TypeError("model must be BaseModel or dict")
 

learnware/market/easy.py

@@ -72,8 +72,7 @@ class EasyMarket(BaseMarket):
         specification = Specification(semantic_spec=semantic_spec, stat_spec=stat_spec)
         # Commented for test purpose. Uncomment when Learnware class is implemented.
         # model_dict = {"module_path": model_path, "class_name": "BaseModel"}
-        model_dict = model_path
-        new_learnware = Learnware(id=id, name=learnware_name, model=model_dict, specification=specification)
+        new_learnware = Learnware(id=id, name=learnware_name, model=model_path, specification=specification)
         self.learnware_list[id] = new_learnware
         self.count += 1
         add_learnware_to_db(