add node and edge sampler

Author: thunderock

examples/Untitled.ipynb

@@ -0,0 +1,67 @@
+from typing import Optional
+import numpy as np
+import pytorch_lightning as pl
+import random
+from ..utils import utils
+
+
+class Sampler(object):
+    def _set_seed(self):
+        np.random.seed(self.seed)
+        pl.seed_everything(self.seed)
+        random.seed(self.seed)
+
+    def _set_number_target(self, number_target):
+        if number_target is None:
+            self.number_target = number_target
+        return self.number_target
+
+    def __init__(self, window_length: int = 10, number_target: Optional[int] = None, seed: Optional[int] = None, degree_agnostic: Boolen) -> None:
+        """
+        Parameters
+        ----------
+        window_length : int
+            Number of nodes to sample in the context window
+        number_target : int
+               Number of target nodes or edges to sample, can be none because fit can take centers
+        seed : int
+            Seed for random number generator
+        """
+        self.window_length = window_length
+        self.number_target = number_target
+        self.seed = utils.get_formatted_environ_variable("SEED", int, 42) if seed is None else seed
+        self._set_seed()
+
+
+    def num_nodes(self, A):
+        return A.shape[0]
+
+    def num_edges(self, A):
+        return A.nnz
+
+    def _generate_centers(self, A):
+        return np.random.choice(self.num_nodes(A), self.number_target, replace=False)
+    def sample(self, centers: Optional[np.ndarray], padding_mask: int = 0) -> np.ndarray:
+        raise NotImplementedError
+
+
+"""
+
+Did:
+1. Reviewed, tested and merged STD team's PR to fix principal failure issue.
+2. Raised PR for funding type and reserve amount in ltd-cdf-pipeline.
+3. Completed Pru backtesting and asked Drew to review.
+4. Oncall deployment duties.
+Will do:
+1. Address any comments on PR for funding type and reserve amount in ltd-cdf-pipeline.
+2. Start testing Prudential on ltd-raw-data-preprocessor.
+
+Did:
+1. Merged PR for funding type and reserve amount in ltd-cdf-pipeline.
+2. Finished first version of Prudential on ltd-raw-data-preprocessor.
+Will do:
+1. Issue today for other clients due to funding type and reserve amount. Will check in local and then raise PR.
+2. Keep working on Prudential on ltd-raw-data-preprocessor. I am stuck with how to use export date asset. Will reach out to platform oncall
+in case not able to figure it out.
+3. Postpartum meeting for oneamerica oncall issue. Ask questions I had in mind.
+"""