Skip to content

Feature/access improvements #301

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Mark2000 merged 13 commits into from
Oct 16, 2025
Merged

Feature/access improvements #301

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
e509735
[#309] Cache observation space, option to limit window computations
Mark2000 Jul 31, 2025
bc016dc
[#309] Improve next window handling
Mark2000 Aug 1, 2025
487609e
[#309] Refactor to use single buffer in EOS envs
Mark2000 Aug 1, 2025
cc83cfc
[#309] Only generate expensive log messages when needed
Mark2000 Aug 1, 2025
ba50722
[#309] Pure-python cross product in rv2HN
Mark2000 Aug 1, 2025
71a4e54
[#309] Short circuit observation filter behavior
Mark2000 Aug 1, 2025
9f12c3d
[#309] Check locations for opportunities after specified time
Mark2000 Aug 1, 2025
eb903c3
[#309] Remove slow hash
Mark2000 Aug 2, 2024
4f0b28f
[#309] Remove inaccesible code
Mark2000 Aug 1, 2025
4ef32e4
[#309] Use sets instead of lists in UniqueImageData
Mark2000 Aug 13, 2025
af42322
[#309] Access satellite test updates
Mark2000 Sep 15, 2025
c1832ea
[#309] Update release notes
Mark2000 Sep 15, 2025
598a006
[#309] Make cloud environments compatible
LorenzzoQM Oct 16, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions docs/source/release_notes.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@ Development - |version|
* Add example script for shielded training with action replacement and action masking in `Shielded training with action masking and action replacement <examples/training_with_shield.ipynb>`_.
* Add ``bsk`` as a dependency in ``pyproject.toml``.
* Update the CI/CD workflows to build BSK-RL using the new ``bsk`` dependency.
* Optimize performance of AEOS environments, especially for high request counts.


Version 1.2.0
Expand Down
594 changes: 66 additions & 528 deletions examples/cloud_environment.ipynb
View file
Open in desktop

Large diffs are not rendered by default.

155 changes: 78 additions & 77 deletions examples/cloud_environment_with_reimaging.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -160,11 +160,11 @@
"\n",
" def __init__(\n",
" self,\n",
" imaged: Optional[list[\"Target\"]] = None,\n",
" imaged: Optional[set[\"Target\"]] = None,\n",
" duplicates: int = 0,\n",
" known: Optional[list[\"Target\"]] = None,\n",
" cloud_covered: Optional[list[\"Target\"]] = None,\n",
" cloud_free: Optional[list[\"Target\"]] = None,\n",
" known: Optional[set[\"Target\"]] = None,\n",
" cloud_covered: Optional[set[\"Target\"]] = None,\n",
" cloud_free: Optional[set[\"Target\"]] = None,\n",
" ) -> None:\n",
" \"\"\"Construct unit of data to record unique images.\n",
"\n",
Expand All @@ -174,19 +174,19 @@
" ``cloud_covered`` and ``cloud_free`` based on the specified threshold.\n",
"\n",
" Args:\n",
" imaged: List of targets that are known to be imaged.\n",
" imaged: Set of targets that are known to be imaged.\n",
" duplicates: Count of target imaging duplication.\n",
" known: List of targets that are known to exist (imaged and unimaged).\n",
" cloud_covered: List of imaged targets that are known to be cloud covered.\n",
" cloud_free: List of imaged targets that are known to be cloud free.\n",
" known: Set of targets that are known to exist (imaged and unimaged).\n",
" cloud_covered: Set of imaged targets that are known to be cloud covered.\n",
" cloud_free: Set of imaged targets that are known to be cloud free.\n",
" \"\"\"\n",
" super().__init__(imaged=imaged, duplicates=duplicates, known=known)\n",
" if cloud_covered is None:\n",
" cloud_covered = []\n",
" cloud_covered = set()\n",
" if cloud_free is None:\n",
" cloud_free = []\n",
" self.cloud_covered = list(set(cloud_covered))\n",
" self.cloud_free = list(set(cloud_free))\n",
" cloud_free = set()\n",
" self.cloud_covered = set(cloud_covered)\n",
" self.cloud_free = set(cloud_free)\n",
"\n",
" def __add__(self, other: \"CloudImageBinaryData\") -> \"CloudImageBinaryData\":\n",
" \"\"\"Combine two units of data.\n",
Expand All @@ -198,17 +198,17 @@
" Combined unit of data.\n",
" \"\"\"\n",
"\n",
" imaged = list(set(self.imaged + other.imaged))\n",
" imaged = self.imaged | other.imaged\n",
" duplicates = (\n",
" self.duplicates\n",
" + other.duplicates\n",
" + len(self.imaged)\n",
" + len(other.imaged)\n",
" - len(imaged)\n",
" )\n",
" known = list(set(self.known + other.known))\n",
" cloud_covered = list(set(self.cloud_covered + other.cloud_covered))\n",
" cloud_free = list(set(self.cloud_free + other.cloud_free))\n",
" known = self.known | other.known\n",
" cloud_covered = self.cloud_covered | other.cloud_covered\n",
" cloud_free = self.cloud_free | other.cloud_free\n",
"\n",
" return self.__class__(\n",
" imaged=imaged,\n",
Expand Down Expand Up @@ -236,27 +236,25 @@
" Returns:\n",
" list: Targets imaged at new_state that were unimaged at old_state\n",
" \"\"\"\n",
" update_idx = np.where(new_state - old_state > 0)[0]\n",
" imaged = []\n",
" for idx in update_idx:\n",
" message = self.satellite.dynamics.storageUnit.storageUnitDataOutMsg\n",
" target_id = message.read().storedDataName[int(idx)]\n",
" imaged.append(\n",
" [target for target in self.data.known if target.id == target_id][0]\n",
" )\n",
"\n",
" cloud_covered = []\n",
" cloud_free = []\n",
" for target in imaged:\n",
" cloud_coverage = target.cloud_cover_true\n",
" if cloud_coverage > target.reward_threshold:\n",
" cloud_covered.append(target)\n",
" data_increase = new_state - old_state\n",
" if data_increase <= 0:\n",
" return UniqueImageData()\n",
" else:\n",
" assert self.satellite.latest_target is not None\n",
" self.update_target_colors([self.satellite.latest_target])\n",
" cloud_coverage = self.satellite.latest_target.cloud_cover_true\n",
" cloud_threshold = self.satellite.latest_target.reward_threshold\n",
" if cloud_coverage > cloud_threshold:\n",
" cloud_covered = [self.satellite.latest_target]\n",
" cloud_free = []\n",
" else:\n",
" cloud_free.append(target)\n",
"\n",
" return CloudImageBinaryData(\n",
" imaged=imaged, cloud_covered=cloud_covered, cloud_free=cloud_free\n",
" )\n",
" cloud_covered = []\n",
" cloud_free = [self.satellite.latest_target]\n",
" return CloudImageBinaryData(\n",
" imaged={self.satellite.latest_target},\n",
" cloud_covered=cloud_covered,\n",
" cloud_free=cloud_free,\n",
" )\n",
"\n",
"\n",
"class CloudImageBinaryRewarder(UniqueImageReward):\n",
Expand All @@ -276,14 +274,20 @@
" reward: Cumulative reward across satellites for one step\n",
" \"\"\"\n",
" reward = {}\n",
" imaged_counts = {}\n",
" for new_data in new_data_dict.values():\n",
" for target in new_data.imaged:\n",
" if target not in imaged_counts:\n",
" imaged_counts[target] = 0\n",
" imaged_counts[target] += 1\n",
"\n",
" for sat_id, new_data in new_data_dict.items():\n",
" reward[sat_id] = 0.0\n",
" for target in new_data.cloud_free:\n",
" reward[sat_id] += self.reward_fn(target.priority)\n",
"\n",
" for new_data in new_data_dict.values():\n",
" self.data += new_data\n",
" if target not in self.data.imaged:\n",
" reward[sat_id] += (\n",
" self.reward_fn(target.priority) / imaged_counts[target]\n",
" )\n",
" return reward\n",
"\n",
"\n",
Expand Down Expand Up @@ -326,9 +330,9 @@
" def __init__(\n",
" self,\n",
" imaged: Optional[list[\"Target\"]] = None,\n",
" imaged_complete: Optional[list[\"Target\"]] = None,\n",
" imaged_complete: Optional[set[\"Target\"]] = None,\n",
" list_belief_update_var: Optional[list[float]] = None,\n",
" known: Optional[list[\"Target\"]] = None,\n",
" known: Optional[set[\"Target\"]] = None,\n",
" ) -> None:\n",
" \"\"\"Construct unit of data to record unique images.\n",
"\n",
Expand All @@ -337,22 +341,22 @@
"\n",
" Args:\n",
" imaged: List of targets that are known to be imaged.\n",
" imaged_complete: List of targets that are known to be completely imaged (P(S=1) >= reward_threshold).\n",
" imaged_complete: Set of targets that are known to be completely imaged (P(S=1) >= reward_threshold).\n",
" list_belief_update_var: List of belief update variations for each target after each picture.\n",
" known: List of targets that are known to exist (imaged and not imaged)\n",
" \"\"\"\n",
" if imaged is None:\n",
" imaged = []\n",
" if imaged_complete is None:\n",
" imaged_complete = []\n",
" imaged_complete = set()\n",
" if list_belief_update_var is None:\n",
" list_belief_update_var = []\n",
" if known is None:\n",
" known = []\n",
" self.known = list(set(known))\n",
" known = set()\n",
" self.known = set(known)\n",
"\n",
" self.imaged = list(imaged)\n",
" self.imaged_complete = list(set(imaged_complete))\n",
" self.imaged = imaged\n",
" self.imaged_complete = imaged_complete\n",
" self.list_belief_update_var = list(list_belief_update_var)\n",
"\n",
" def __add__(\n",
Expand All @@ -367,13 +371,13 @@
" Combined unit of data.\n",
" \"\"\"\n",
"\n",
" imaged = list(self.imaged + other.imaged)\n",
" imaged_complete = list(set(self.imaged_complete + other.imaged_complete))\n",
" list_belief_update_var = list(\n",
" imaged = self.imaged + other.imaged\n",
" imaged_complete = self.imaged_complete | other.imaged_complete\n",
" list_belief_update_var = (\n",
" self.list_belief_update_var + other.list_belief_update_var\n",
" )\n",
"\n",
" known = list(set(self.known + other.known))\n",
" known = self.known | other.known\n",
" return self.__class__(\n",
" imaged=imaged,\n",
" imaged_complete=imaged_complete,\n",
Expand Down Expand Up @@ -401,9 +405,8 @@
" Returns:\n",
" array: storedData from satellite storage unit\n",
" \"\"\"\n",
" return np.array(\n",
" self.satellite.dynamics.storageUnit.storageUnitDataOutMsg.read().storedData\n",
" )\n",
" msg = self.satellite.dynamics.storageUnit.storageUnitDataOutMsg.read()\n",
" return msg.storedData[0]\n",
"\n",
" def compare_log_states(\n",
" self, old_state: np.ndarray, new_state: np.ndarray\n",
Expand All @@ -421,22 +424,18 @@
" Returns:\n",
" list: Targets imaged at new_state that were unimaged at old_state\n",
" \"\"\"\n",
" update_idx = np.where(new_state - old_state > 0)[0]\n",
" imaged = []\n",
" for idx in update_idx:\n",
" message = self.satellite.dynamics.storageUnit.storageUnitDataOutMsg\n",
" target_id = message.read().storedDataName[int(idx)]\n",
" imaged.append(\n",
" [target for target in self.data.known if target.id == target_id][0]\n",
" )\n",
"\n",
" list_imaged_complete = []\n",
" list_belief_update_var = []\n",
" data_increase = new_state - old_state\n",
" if data_increase <= 0:\n",
" return CloudImageProbabilityData()\n",
" else:\n",
" assert self.satellite.latest_target is not None\n",
" # return UniqueImageData(imaged={self.satellite.latest_target})\n",
"\n",
" current_sim_time = self.satellite.simulator.sim_time\n",
" belief_update_func = self.satellite.belief_update_func\n",
" target = self.satellite.latest_target\n",
" current_sim_time = self.satellite.simulator.sim_time\n",
" belief_update_func = self.satellite.belief_update_func\n",
"\n",
" for target in imaged:\n",
" target_prev_obs = (\n",
" target.prev_obs\n",
" ) # Time at which the target was previously observed\n",
Expand All @@ -457,18 +456,20 @@
" target.prev_obs = current_sim_time # Update the previous observation time\n",
"\n",
" if updated_belief[1] > target.reward_threshold:\n",
" list_imaged_complete.append(target)\n",
" list_belief_update_var.append(target.belief_update_var)\n",
" list_imaged_complete = [target]\n",
" else:\n",
" list_imaged_complete = []\n",
" list_belief_update_var = target.belief_update_var\n",
"\n",
" return CloudImageProbabilityData(\n",
" imaged=imaged,\n",
" imaged_complete=list_imaged_complete,\n",
" list_belief_update_var=list_belief_update_var,\n",
" )\n",
" return CloudImageProbabilityData(\n",
" imaged=[target],\n",
" imaged_complete=set(list_imaged_complete),\n",
" list_belief_update_var=[list_belief_update_var],\n",
" )\n",
"\n",
"\n",
"class CloudImageProbabilityRewarder(GlobalReward):\n",
" datastore_type = CloudImageProbabilityDataStore\n",
" data_store_type = CloudImageProbabilityDataStore\n",
"\n",
" def __init__(\n",
" self,\n",
Expand Down Expand Up @@ -505,6 +506,7 @@
" Returns:\n",
" reward: Cumulative reward across satellites for one step\n",
" \"\"\"\n",
"\n",
" reward = {}\n",
"\n",
" for sat_id, new_data in new_data_dict.items():\n",
Expand All @@ -519,7 +521,6 @@
" reward[sat_id] += self.reward_fn(\n",
" target.priority, None, self.alpha, reach_threshold=True\n",
" )\n",
"\n",
" return reward\n",
"\n",
"\n",
Expand Down
Loading