Add types to python bindings.

.gitignore

@@ -4,3 +4,4 @@
 .idea
 cleora-light-rust.iml
 *.out
+*.so

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "pycleora"
-version = "2.0.0"
+version = "2.1.0"
 edition = "2018"
 license-file = "LICENSE"
 readme = "README.md"

pycleora/pycleora.pyi

@@ -0,0 +1,38 @@
+from typing import Any, Iterable, Optional, Self
+
+import numpy as np
+from numpy.typing import NDArray
+
+
+class SparseMatrix:
+    def __new__(cls, *args: Any) -> Self:
+        pass
+
+    @classmethod
+    def from_iterator(
+        cls, hyperedges: Iterable[str], columns: str, hyperedge_trim_n: int = 16, num_workers: Optional[int] = None
+    ) -> Self:
+        pass
+
+    @classmethod
+    def from_files(
+        cls, filepaths: list[str], columns: str, hyperedge_trim_n: int = 16, num_workers: Optional[int] = None
+    ) -> Self:
+        pass
+
+    def left_markov_propagate(self, x: NDArray[np.float32], num_workers: Optional[int] = None) -> NDArray[np.float32]:
+        pass
+
+    def symmetric_markov_propagate(
+        self, x: NDArray[np.float32], num_workers: Optional[int] = None
+    ) -> NDArray[np.float32]:
+        pass
+
+    def get_entity_column_mask(self, column_name: str) -> NDArray[np.bool]:
+        pass
+
+    def entity_degrees(self) -> NDArray[np.float32]:
+        pass
+
+    def initialize_deterministically(self, feature_dim: int, seed: int = 0) -> NDArray[np.float32]:
+        pass

pyproject.toml

@@ -10,7 +10,7 @@ classifiers = [
     "Programming Language :: Python :: Implementation :: CPython",
     "Programming Language :: Python :: Implementation :: PyPy",
 ]
-version = "2.0.0"
+version = "2.1.0"
 description = "Sparse hypergraph structure and markov-propagation for node embeddings embeddings exposed via Python bindings."
 readme = { file = "README.md", content-type = "text/markdown" }
 authors = [