Merge pull request #224 from ie3-institute/to/dev

Add graph functionality

CHANGELOG.md

@@ -9,6 +9,7 @@ All notable changes to this project will be documented in this file.
 - Added `Transformers2WResult` dictionary support [#217](https://github.com/ie3-institute/pypsdm/pull/217)
 - `PrimaryData` time series are now protected, introduced `add_time_series` method [#218](https://github.com/ie3-institute/pypsdm/pull/218)
 - Added method to create nodes and lines [#223](https://github.com/ie3-institute/pypsdm/pull/223)
+- Added graph traversal functionality [#224](https://github.com/ie3-institute/pypsdm/pull/224)
 
 ### Changed
 

pypsdm/graph.py

@@ -0,0 +1,84 @@
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+import networkx as nx
+from networkx import Graph
+
+if TYPE_CHECKING:
+    from pypsdm import RawGridContainer
+
+
+def find_branches(G: Graph, start_node):
+    visited = set()
+    visited.add(start_node)
+    branches = []
+
+    def dfs(node, path):
+        visited.add(node)
+        path.append(node)
+
+        for neighbor in G.neighbors(node):
+            if neighbor not in visited:
+                dfs(neighbor, path)
+
+    for neighbor in G.neighbors(start_node):
+        if neighbor not in visited:
+            path = [start_node]
+            dfs(neighbor, path)
+            branches.append(path)
+
+    return branches
+
+
+def find_n_hop_closest_in_slack_direction(
+    uuid: str, n: int, raw_grid: RawGridContainer, candidates: set[str] | None = None
+):
+    slack_ds = raw_grid.find_slack_downstream()
+    G = raw_grid.build_networkx_graph()
+    if candidates is None:
+        candidates = set(raw_grid.nodes.uuid)
+    # find djikstra shortest path between uuid and slck_ds
+    path = nx.shortest_path(G, uuid, slack_ds)
+    closest = find_n_hop_closest_candidates(n, G, uuid, candidates)
+    closest = [n for n in closest if n in path]
+    return closest
+
+
+def find_n_hop_closest_candidates(n: int, G, uuid, candidates):
+    """
+    Find all nodes within candidates that are n candidate hops away from the given node.
+    """
+    closest = set([uuid])
+    for _ in range(n):
+        for node in closest:
+            closest = closest.union(find_closest_candidates(G, node, candidates))
+    return closest
+
+
+def find_closest_candidates(G: nx.Graph, node: str, candidates: set[str]) -> list[str]:
+    """
+    Starting from the node, find the closest nodes in all directions, that are in the
+    candidates set.
+
+    Args:
+        G: The graph to search in.
+        node: The node for which to find the closest candidates.
+        candidates: The set of nodes to search for.
+    """
+    # NOTE: This might be added to pypsdm
+    res = []
+    visited = set()
+
+    def preorder_dfs(cur: str):
+        if cur in visited:
+            return
+        visited.add(cur)
+        if cur in candidates and cur != node:
+            res.append(cur)
+            return
+        for n in G.neighbors(cur):
+            preorder_dfs(n)
+
+    preorder_dfs(node)
+    return res

pypsdm/models/gwr.py

@@ -1,5 +1,6 @@
 from dataclasses import dataclass
 from datetime import datetime
+from pathlib import Path
 from typing import Optional, Tuple, Union
 
 import numpy as np
@@ -274,8 +275,8 @@ def to_csv(
     @classmethod
     def from_csv(
         cls,
-        grid_path: str,
-        result_path: str,
+        grid_path: str | Path,
+        result_path: str | Path,
         grid_delimiter: str | None = None,
         result_delimiter: str | None = None,
         primary_data_delimiter: Optional[str] = None,

pypsdm/models/input/container/grid.py

@@ -2,6 +2,7 @@
 
 from dataclasses import dataclass
 from datetime import datetime
+from pathlib import Path
 from typing import TYPE_CHECKING, Dict, Optional, Union
 
 from pypsdm.models.enums import (
@@ -171,7 +172,7 @@ def plot_grid(
 
     def to_csv(
         self,
-        path: str,
+        path: str | Path,
         include_primary_data: bool = True,
         mkdirs: bool = False,
         delimiter: str = ",",
@@ -185,7 +186,7 @@ def to_csv(
     @classmethod
     def from_csv(
         cls,
-        path: str,
+        path: str | Path,
         delimiter: str | None = None,
         primary_data_delimiter: Optional[str] = None,
     ):

pypsdm/models/input/container/mixins.py

@@ -7,6 +7,7 @@
 from dataclasses import replace
 from datetime import datetime
 from functools import partial
+from pathlib import Path
 from typing import TYPE_CHECKING, Self, Tuple
 
 from loguru import logger
@@ -112,7 +113,7 @@ def entity_keys(cls):
     @classmethod
     def entities_from_csv(
         cls,
-        simulation_data_path: str,
+        simulation_data_path: str | Path,
         simulation_end: datetime | None = None,
         grid_container: GridContainer | None = None,
         delimiter: str | None = None,

pypsdm/models/input/container/participants.py

@@ -1,4 +1,5 @@
 from dataclasses import dataclass
+from pathlib import Path
 from typing import Union
 
 import pandas as pd
@@ -151,7 +152,7 @@ def subset(self, uuids):
 
     @staticmethod
     def from_csv(
-        path: str, delimiter: str | None = None
+        path: str | Path, delimiter: str | None = None
     ) -> "SystemParticipantsContainer":
         loads = Loads.from_csv(path, delimiter, must_exist=False)
         fixed_feed_ins = FixedFeedIns.from_csv(path, delimiter, must_exist=False)

pypsdm/models/input/container/raw_grid.py

@@ -1,8 +1,10 @@
 from dataclasses import dataclass
+from pathlib import Path
 from typing import Union
 
 from networkx import Graph
 
+from pypsdm.graph import find_branches
 from pypsdm.models.enums import RawGridElementsEnum
 from pypsdm.models.input.connector.lines import Lines
 from pypsdm.models.input.connector.switches import Switches
@@ -61,7 +63,7 @@ def get_branches(self, as_graphs=False) -> Union[list[list[str]], list[Graph]]:
             raise ValueError("Did not find a slack node!")
         slack_connected_node = slack_connected_node.pop()
         graph = self.build_networkx_graph()
-        branches = self._find_branches(graph, slack_connected_node)
+        branches = find_branches(graph, slack_connected_node)
         branches = [[slack_connected_node] + branch for branch in branches]
         if as_graphs:
             return [graph.subgraph(branch).copy() for branch in branches]
@@ -110,30 +112,40 @@ def admittance_matrix(self, uuid_to_idx: dict):
         transformers_admittance = self.transformers_2_w.admittance_matrix(uuid_to_idx)
         return lines_admittance + transformers_admittance
 
-    @staticmethod
-    def _find_branches(G: Graph, start_node):
-        visited = set()
-        visited.add(start_node)
-        branches = []
-
-        def dfs(node, path):
-            visited.add(node)
-            path.append(node)
-
-            for neighbor in G.neighbors(node):
-                if neighbor not in visited:
-                    dfs(neighbor, path)
-
-        for neighbor in G.neighbors(start_node):
-            if neighbor not in visited:
-                path = []
-                dfs(neighbor, path)
-                branches.append(path)
+    def find_slack_downstream(self) -> str:
+        """
+        Find the downstream node of the slack node, which is the node on the transformer's
+        lower voltage side.
+        """
+        from pypsdm import Transformers2W
 
-        return branches
+        slack_node = self.nodes.get_slack_nodes()
+        if len(slack_node.data) != 1:
+            raise ValueError("Currently only implemented for singular slack nodes.")
+        transformers = self.transformers_2_w
+        slack_transformers = Transformers2W(
+            transformers.data[
+                (transformers.node_a.isin(slack_node.uuid.to_list()))
+                | (transformers.node_b.isin(slack_node.uuid.to_list()))
+            ]
+        )
+        slack_connected_node = (
+            set(slack_transformers.node_a)
+            .union(slack_transformers.node_b)
+            .difference(slack_node.uuid)
+        )
+        if len(slack_connected_node) > 1:
+            raise ValueError(
+                "There are multiple nodes connected to the slack node via a transformer."
+            )
+        elif len(slack_connected_node) == 0:
+            raise ValueError("Did not find a slack node!")
+        return slack_connected_node.pop()
 
     @classmethod
-    def from_csv(cls, path: str, delimiter: str | None = None) -> "RawGridContainer":
+    def from_csv(
+        cls, path: str | Path, delimiter: str | None = None
+    ) -> "RawGridContainer":
         nodes = Nodes.from_csv(path, delimiter)
         lines = Lines.from_csv(path, delimiter, must_exist=False)
         transformers_2_w = Transformers2W.from_csv(path, delimiter, must_exist=False)

pypsdm/models/input/create/grid_elements.py

@@ -1,8 +1,10 @@
-from pypsdm.models.input import Nodes, Lines
-from pypsdm.models.input.create.utils import create_data
 from uuid import uuid4
+
 import pandas as pd
 
+from pypsdm.models.input import Lines, Nodes
+from pypsdm.models.input.create.utils import create_data
+
 
 def create_nodes(data_dict):
     return Nodes(create_data(data_dict, create_nodes_data))

pypsdm/models/input/entity.py

@@ -344,7 +344,7 @@ def copy(
     @classmethod
     def from_csv(
         cls: Type[Self],
-        path: str,
+        path: str | Path,
         delimiter: str | None = None,
         must_exist: bool = True,
     ) -> Self:
@@ -364,7 +364,7 @@ def from_csv(
     @classmethod
     def _from_csv(
         cls: Type[Self],
-        path: str,
+        path: str | Path,
         entity: EntitiesEnum,
         delimiter: str | None = None,
         must_exist: bool = True,

pypsdm/models/result/container/grid.py

@@ -3,6 +3,7 @@
 import os
 from dataclasses import dataclass
 from datetime import datetime
+from pathlib import Path
 from typing import TYPE_CHECKING, Optional, Union
 
 from pypsdm.io.utils import check_filter
@@ -133,7 +134,7 @@ def concat(self, other: "GridResultContainer", deep: bool = True, keep="last"):
     @classmethod
     def from_csv(
         cls,
-        simulation_data_path: str,
+        simulation_data_path: str | Path,
         delimiter: str | None = None,
         simulation_end: Optional[datetime] = None,
         grid_container: Optional[GridContainer] = None,

pypsdm/models/result/container/participants.py

@@ -1,5 +1,6 @@
 from dataclasses import dataclass
 from datetime import datetime
+from pathlib import Path
 from typing import Iterable, Optional, Self, Tuple, Union
 
 import pandas as pd
@@ -204,7 +205,7 @@ def to_csv(self, path: str, delimiter: str = ",", mkdirs=False):
     @classmethod
     def from_csv(
         cls,
-        simulation_data_path: str,
+        simulation_data_path: str | Path,
         simulation_end: Optional[datetime] = None,
         grid_container: Optional[GridContainer] = None,
         delimiter: Optional[str] = None,

pypsdm/models/result/container/raw_grid.py

@@ -1,5 +1,6 @@
 from dataclasses import dataclass
 from datetime import datetime
+from pathlib import Path
 from typing import Optional, Self, Union
 
 from pypsdm.models.enums import RawGridElementsEnum
@@ -112,7 +113,7 @@ def entity_keys(cls):
     @classmethod
     def from_csv(
         cls,
-        simulation_data_path: str,
+        simulation_data_path: str | Path,
         simulation_end: Optional[datetime] = None,
         grid_container: Optional[GridContainer] = None,
         delimiter: Optional[str] = None,

pypsdm/processing/series.py

@@ -9,19 +9,19 @@
 def duration_weighted_series(series: Series):
     series.sort_index(inplace=True)
     values = series[:-1].reset_index(drop=True)
-    time = (
+    duration = (
         (series.index[1::] - series.index[:-1])
         .to_series()
         .apply(lambda x: x.total_seconds() / 3600)
         .reset_index(drop=True)
     )
-    return pd.concat([values.rename("values"), time.rename("time")], axis=1)
+    return pd.concat([values.rename("values"), duration.rename("duration")], axis=1)
 
 
 def weighted_series_sum(weighted_series: DataFrame) -> float:
     if len(weighted_series) == 0:
         return 0.0
-    return (weighted_series["values"] * weighted_series["time"]).sum()
+    return (weighted_series["values"] * weighted_series["duration"]).sum()
 
 
 def duration_weighted_sum(series: Series) -> float: