Transition and measurement models for road networks #812

sglvladi · 2023-06-02T13:44:29Z

This PR aims to add the following:

RoadNetwork type:
- A subclass of networkx.DiGraph, with some limitations on the allowed node and edge attributes, and some additional methods.
OptimalPathToDestinationTransitionModel transition model:
- A transition model that models a target travelling along the optimal path to a destination, on a given road network.
OptimalPathToDestinationMeasurementModel measurement model:
- A measurement model that projects the target's position on the road network to a 2D position.

TODO

Add tests
Remove/Reformat example

…RoadNetwork add_nodes_from

…ionModel

…mentModel

…to better reflect their purpose

…asurementModel

codecov · 2024-03-18T14:42:47Z

Codecov Report

Attention: Patch coverage is 96.17647% with 13 lines in your changes are missing coverage. Please review.

Project coverage is 93.56%. Comparing base (7fce938) to head (0f13262).
Report is 23 commits behind head on main.

Files	Patch %	Lines
stonesoup/models/transition/graph.py	90.10%	4 Missing and 5 partials ⚠️
stonesoup/models/measurement/graph.py	92.68%	0 Missing and 3 partials ⚠️
stonesoup/functions/graph.py	98.03%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #812      +/-   ##
==========================================
+ Coverage   93.49%   93.56%   +0.06%     
==========================================
  Files         200      204       +4     
  Lines       12715    13055     +340     
  Branches     2615     2685      +70     
==========================================
+ Hits        11888    12215     +327     
- Misses        585      590       +5     
- Partials      242      250       +8

Flag	Coverage Δ
integration	`65.26% <22.05%> (-1.16%)`	⬇️
unittests	`89.26% <96.17%> (+0.22%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

sglvladi · 2024-03-18T14:45:58Z

@jswright-dstl, @sdhiscocks this PR is ready for review. I've left the example in for now to help with the review, but plan to remove it before merging.

sdhiscocks · 2024-03-26T10:41:11Z

stonesoup/models/measurement/graph.py

+        x_k = \left[r_k, \cdots, e_k, d_k, s_k\right]
+
+    and :math:`e_k` denotes the edge the target is currently on, :math:`r_k` is the distance
+    travelled along the edge, :m    ath:`d_k` is the destination node, and :math:`s_k` is


Suggested change

travelled along the edge, :m ath:`d_k` is the destination node, and :math:`s_k` is

travelled along the edge, :math:`d_k` is the destination node, and :math:`s_k` is

sdhiscocks · 2024-03-26T11:23:13Z

stonesoup/models/transition/graph.py

+            v_dest = dict()
+            for edge in unique_edges:
+                # Filter paths that contain the edge
+                filtered_paths = filter(lambda x: np.any(x[1] == edge), s_paths.items())
+                v_dest_tmp = {dest for (_, dest), _ in filtered_paths}
+                if len(v_dest_tmp):
+                    try:
+                        v_dest[edge] |= v_dest_tmp
+                    except KeyError:
+                        v_dest[edge] = v_dest_tmp
+
+            # Perform destination sampling
+            resample_inds = np.flatnonzero(
+                np.random.binomial(1, float(self.destination_resample_probability), num_particles)
+            )
+            for i in resample_inds:
+                try:
+                    v_dest_tmp = list(v_dest[edges[i]])
+                except KeyError:
+                    # If no valid destinations exist for the current edge, keep the current
+                    # destination
+                    continue
+                new_state_vectors[-2, i] = np.random.choice(v_dest_tmp)


I think this reads a little better vs. temp variables. Also change np.random.choice to self.random_state.choice.

Suggested change

v_dest = dict()

for edge in unique_edges:

# Filter paths that contain the edge

filtered_paths = filter(lambda x: np.any(x[1] == edge), s_paths.items())

v_dest_tmp = {dest for (_, dest), _ in filtered_paths}

if len(v_dest_tmp):

try:

v_dest[edge] |= v_dest_tmp

except KeyError:

v_dest[edge] = v_dest_tmp

# Perform destination sampling

resample_inds = np.flatnonzero(

np.random.binomial(1, float(self.destination_resample_probability), num_particles)

)

for i in resample_inds:

try:

v_dest_tmp = list(v_dest[edges[i]])

except KeyError:

# If no valid destinations exist for the current edge, keep the current

# destination

continue

new_state_vectors[-2, i] = np.random.choice(v_dest_tmp)

v_dest = defaultdict(set)

for edge in unique_edges:

# Filter paths that contain the edge

filtered_paths = filter(lambda x: np.any(x[1] == edge), s_paths.items())

v_dest[edge] |= {dest for (_, dest), _ in filtered_paths}

# Perform destination sampling

resample_inds = np.flatnonzero(

np.random.binomial(1, float(self.destination_resample_probability), num_particles)

)

for i in resample_inds:

dests = v_dest[edges[i]]

# If no valid destinations exist for the current edge, keep the current

# destination

if dests:

new_state_vectors[-2, i] = self.random_state.choice(list(dests))

sdhiscocks · 2024-03-27T09:14:47Z

setup.cfg

+    geopandas
+    shapely
+    networkx


Wondering if this should be optional

I have added these under a new "roadnet" optional deps group. However, now the road net tests are getting skipped. Should I add also these to the "dev" dependencies, or should I modify the CircleCI config to install with "roadnet" deps?

sdhiscocks · 2024-03-27T09:16:33Z

stonesoup/models/transition/graph.py

+
+            # Perform destination sampling
+            resample_inds = np.flatnonzero(
+                np.random.binomial(1, float(self.destination_resample_probability), num_particles)


Suggested change

np.random.binomial(1, float(self.destination_resample_probability), num_particles)

self.random_state.binomial(1, float(self.destination_resample_probability), num_particles)

sdhiscocks · 2024-03-27T09:18:21Z

stonesoup/models/transition/graph.py

+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.random_state = np.random.RandomState(self.seed) if self.seed is not None else None


Suggested change

self.random_state = np.random.RandomState(self.seed) if self.seed is not None else None

if self.seed is not None:

self.random_state = np.random.RandomState(self.seed)

else:

self.random_state = np.random.mtrand._rand

… and remove use of temp variables

…cstring

sglvladi requested review from sdhiscocks and jswright-dstl June 2, 2023 13:44

sglvladi requested a review from a team as a code owner June 2, 2023 13:44

jswright-dstl requested a review from orosoman-dstl June 5, 2023 09:19

sglvladi added 6 commits February 19, 2024 15:18

Add RoadNetwork type

85c493c

Add ShortestPathToDestinationTransitionModel

6b1bef6

Add ShortestPathToDestinationMeasurementModel

e743c84

Add dependencies

fe7bdc0

Update documentation

20b88e4

Add example

cb85b51

sglvladi force-pushed the road_net branch from 6629ccc to cb85b51 Compare February 19, 2024 15:23

sglvladi added 3 commits February 28, 2024 16:16

Add minn_2.mat file

19857f9

Minor fix in ShortestPathToDestinationMeasurementModel

6a2d44d

Minor update in example

dd02dd3

sglvladi force-pushed the road_net branch from a07dc62 to dd02dd3 Compare February 28, 2024 16:20

sglvladi added 12 commits March 11, 2024 13:56

Ensure node ids are positive integers and correct error is raised in …

93f4789

…RoadNetwork add_nodes_from

Tidy up RoadNetwork.to_gdf method and improve docs

65cfdfc

Remove plotting functions from RoadNetwork

4e58345

Fix flake8 errors in types/graph.py

d9c0ccb

Minor corrections and refactoring in ShortestPathToDestinationTransit…

ced4a66

…ionModel

Minor corrections and refactoring in ShortestPathToDestinationMeasure…

36f487b

…mentModel

Docs and minor fixes in graph functions

767a661

Added tests

ffd8ff1

Minor example update

8d43bd5

Replace minn_2.mat with road_net.pkl in example

3df31c5

Renamed ShortestPathToDestination models to OptimalPathToDestination …

520743a

…to better reflect their purpose

Improve tests coverage for RoadNetwork and OptimalPathToDestinationMe…

67a3edd

…asurementModel

sglvladi force-pushed the road_net branch from 0f3fb9f to 67a3edd Compare March 12, 2024 12:32

Amend references in OptimalPathToDestination models to fix sphinx errors

0f13262

sdhiscocks approved these changes Mar 27, 2024

View reviewed changes

sglvladi added 3 commits May 30, 2024 11:21

Ensure OptimalPathToDestinationTransitionModel uses self.random_state…

3943735

… and remove use of temp variables

Minor typo fixed in OptimalPathToDestinationMeasurementModel class do…

7d04ff3

…cstring

Make roadnet dependencies optional

2f29455

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Transition and measurement models for road networks #812

Transition and measurement models for road networks #812

sglvladi commented Jun 2, 2023 •

edited

codecov bot commented Mar 18, 2024

sglvladi commented Mar 18, 2024

sdhiscocks Mar 26, 2024

sdhiscocks Mar 26, 2024

sdhiscocks Mar 27, 2024

sglvladi May 30, 2024 •

edited

sdhiscocks Mar 27, 2024

sdhiscocks Mar 27, 2024

	travelled along the edge, :m ath:`d_k` is the destination node, and :math:`s_k` is
	travelled along the edge, :math:`d_k` is the destination node, and :math:`s_k` is

	np.random.binomial(1, float(self.destination_resample_probability), num_particles)
	self.random_state.binomial(1, float(self.destination_resample_probability), num_particles)

Transition and measurement models for road networks #812

Are you sure you want to change the base?

Transition and measurement models for road networks #812

Conversation

sglvladi commented Jun 2, 2023 • edited

TODO

codecov bot commented Mar 18, 2024

Codecov Report

sglvladi commented Mar 18, 2024

sdhiscocks Mar 26, 2024

Choose a reason for hiding this comment

sdhiscocks Mar 26, 2024

Choose a reason for hiding this comment

sdhiscocks Mar 27, 2024

Choose a reason for hiding this comment

sglvladi May 30, 2024 • edited

Choose a reason for hiding this comment

sdhiscocks Mar 27, 2024

Choose a reason for hiding this comment

sdhiscocks Mar 27, 2024

Choose a reason for hiding this comment

sglvladi commented Jun 2, 2023 •

edited

sglvladi May 30, 2024 •

edited