Created base structure for adding discovery and other algorithms

pyproject.toml

@@ -63,7 +63,8 @@ langchain-openai = "0.3.26"
 langchain-huggingface = ">=0.2.0"
 rank-bm25 = ">=0.2.2"
 sentence-transformers = ">=4.1.0"
-onnxruntime = "1.16.3"
+onnxruntime = "1.20.1"
+multidict = "^6.6.3"
 
 [tool.poetry.group.dev.dependencies]
 poethepoet = "^0.33.0"

pywhyllm/data/dataset.py

@@ -0,0 +1,22 @@
+from typing import List, Protocol
+
+class Dataset(Protocol):
+
+    def graph(self):
+        """
+        Output the ground truth graph associated with the dataset. 
+
+        Returns:
+            A networkx graph 
+        """
+        pass
+
+    def data(self):
+        """
+        Output the data associated with the dataset.
+
+        Returns:
+            A numpy array containing the data
+        """
+        pass
+

pywhyllm/data/generate_data.py

@@ -0,0 +1,12 @@
+
+
+def generate_data(graph, *args, **kwargs):
+    """
+    Generate synthetic data based on a causal graph.
+    Arguments may be added as the functionality is developed.
+    Args:
+        graph: A Networkx graph
+    Returns:
+        A numpy array containing the synthetic data
+    """
+    pass

pywhyllm/data/metrics.py

@@ -0,0 +1,25 @@
+
+
+def accuracy(edges, true_edges):
+    """Accuracy metric for predicted edges for a given dataset
+    """
+    pass
+
+
+def precision(edges, true_edges):
+    """Precision metric for predicted edges for a given dataset
+    """
+    pass
+
+
+def recall(edges, true_edges):
+    """Recall metric for predicted edges for a given dataset
+    """
+    pass
+
+
+def F1(edges, true_edges):
+    """F1 metric for predicted edges for a given dataset
+    """
+    pass
+