HA Websocket for service calls, catch zero rates #3168

springfall2008 · 2026-01-02T12:09:13Z

No description provided.

Copilot

Pull request overview

This PR refactors the HA websocket service call mechanism to use a queue-based architecture and fixes a bug where zero/missing rates caused crashes in cost calculations.

Key Changes:

Refactored websocket service calls from direct async calls to a queue-based system where async_call_service_websocket_command() queues commands that are processed by the existing socketLoop()
Added thread synchronization infrastructure (command queue, pending requests tracker, bridge thread) to coordinate between sync and async contexts
Fixed KeyError crashes in output.py by using .get(minute, 0) instead of direct dictionary indexing when rates are missing
Updated tests to use receive() method instead of __aiter__ pattern for websocket mocking
Removed duplicate wait_api_started() method from GECloudData

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
apps/predbat/ha.py	Major refactoring: Added queue infrastructure for websocket commands, bridge thread for sync/async coordination, and integrated command processing into socketLoop
apps/predbat/output.py	Bug fix: Changed rate dictionary access from `rate_import[minute]` to `rate_import.get(minute, 0)` to handle missing rate data
apps/predbat/tests/test_hainterface_websocket.py	Updated test mocking from `__aiter__` pattern to `receive()` method to match new websocket handling
apps/predbat/tests/test_hainterface_service.py	Refactored tests to use queue-based architecture with background threads simulating socketLoop processing
apps/predbat/gecloud.py	Cleanup: Removed duplicate `wait_api_started()` method (inherited from ComponentBase)
apps/predbat/fox.py	Changed entity refresh to run every 5 minutes (removed "every minute" comment)
apps/predbat/component_base.py	Increased default wait_api_started timeout from 5 to 10 minutes

Copilot · 2026-01-02T12:13:12Z

apps/predbat/ha.py

+            self.ws_sync_event.wait()
+            self.ws_sync_event.clear()
+            if self.ws_event_loop and self.ws_async_event:


The bridge_event thread can hang indefinitely on shutdown. The self.ws_sync_event.wait() call at line 309 has no timeout, which means if api_stop is set but the sync event is never triggered, the thread will block forever. Consider adding a timeout to the wait call (e.g., self.ws_sync_event.wait(timeout=1.0)) to allow periodic checks of the api_stop flag.

Suggested change

self.ws_sync_event.wait()

self.ws_sync_event.clear()

if self.ws_event_loop and self.ws_async_event:

# Use a timeout so we can periodically check api_stop and avoid hanging on shutdown

triggered = self.ws_sync_event.wait(timeout=1.0)

if not triggered:

continue

self.ws_sync_event.clear()

if not self.api_stop and self.ws_event_loop and self.ws_async_event:

Copilot · 2026-01-02T12:13:12Z

apps/predbat/fox.py

-
-        # Refresh HA entities every minute
-        if first or (seconds % (60) == 0):
+            # Refresh HA entities


The comment on line 305 states "Refresh HA entities every minute" but was changed to run every 5 minutes (based on the condition seconds % (5 * 60) == 0 at line 300). The comment should be updated to match the actual behavior, or clarify that this is intentional. The comment has been moved but not updated to reflect the timing change.

Suggested change

# Refresh HA entities

# Refresh HA entities every 5 minutes

Copilot · 2026-01-02T12:13:12Z

apps/predbat/ha.py

+        # Extract result
+        if result_holder.get("error"):
+            self.log("Warn: Service call {}/{} failed: {}".format(domain, service, result_holder["error"]))
+            return None
+
+        success = result_holder.get("success", False)
+        if not success:
+            self.log("Warn: Service call {}/{} data {} failed".format(domain, service, service_data))
+            return None


When event.wait() times out at line 340, the result_holder will still have success=None and error=None. The check at line 347 treats None as False, logging "Service call failed" without indicating that it was a timeout. This makes it difficult to distinguish between an actual service call failure and a timeout. Consider checking explicitly for timeout: if result_holder.get("success") is None and not result_holder.get("error"), log a specific timeout message.

Copilot · 2026-01-02T12:13:12Z

apps/predbat/ha.py

+                                # Send command with current sid
+                                await websocket.send_json({"id": sid, "type": "call_service", "domain": domain, "service": service, "service_data": service_data, "return_response": return_response})
+
+                                # Track pending request
+                                with self.ws_pending_lock:
+                                    self.ws_pending_requests[sid] = {"event": event, "result_holder": result_holder, "timestamp": time.time()}
+
+                                sid += 1


If websocket.send_json() at line 495 raises an exception, the command is removed from the queue (line 487) but never tracked in ws_pending_requests. The caller thread will wait for 30 seconds at line 340 before timing out, but the error won't be properly propagated. Consider wrapping the send_json call in a try-except block to set an error in result_holder if the send fails, then signal the event to unblock the caller immediately.

Suggested change

# Send command with current sid

await websocket.send_json({"id": sid, "type": "call_service", "domain": domain, "service": service, "service_data": service_data, "return_response": return_response})

# Track pending request

with self.ws_pending_lock:

self.ws_pending_requests[sid] = {"event": event, "result_holder": result_holder, "timestamp": time.time()}

sid += 1

try:

# Send command with current sid

await websocket.send_json({"id": sid, "type": "call_service", "domain": domain, "service": service, "service_data": service_data, "return_response": return_response})

# Track pending request

with self.ws_pending_lock:

self.ws_pending_requests[sid] = {"event": event, "result_holder": result_holder, "timestamp": time.time()}

sid += 1

except Exception as e:

# Propagate send failure to waiting caller immediately

try:

result_holder["error"] = "send_failed: {}".format(e)

result_holder["success"] = False

except Exception:

# Defensive: avoid crashing if result_holder has unexpected structure

pass

try:

event.set()

except Exception:

pass

self.log("Error: WebSocket send_json failed for service {}.{}: {}".format(domain, service, e))

Copilot · 2026-01-02T12:13:13Z

apps/predbat/tests/test_hainterface_service.py

+
+    async def test_command():
+        with patch.object(threading.Event, 'wait', mock_wait):
+            # This will timeout immediately


Variable original_wait is not used.

Copilot

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.

Copilot · 2026-01-02T13:37:42Z

apps/predbat/ha.py

+
+        # Check for timeout (neither success nor error was set)
+        if result_holder.get("success") is None and not result_holder.get("error"):
+            self.log("Warn: Service call {}/{} timed out....".format(domain, service))


The timeout message contains four dots ("timed out....") which appears to be a typo. This should be three dots ("timed out...") or remove the extra dots for consistency with other log messages in the codebase.

Suggested change

self.log("Warn: Service call {}/{} timed out....".format(domain, service))

self.log("Warn: Service call {}/{} timed out...".format(domain, service))

Copilot · 2026-01-02T13:37:42Z

apps/predbat/tests/test_hainterface_service.py

+    ha_interface.ws_pending_lock = threading.Lock()
+
+    # Patch threading.Event.wait to simulate instant timeout
+    original_wait = threading.Event.wait


Variable original_wait is not used.

Suggested change

original_wait = threading.Event.wait

springfall2008 added 4 commits January 1, 2026 19:37

Increase timeout to 10 minutes

ebbbce3

Cleanup timeout

db0d86a

Websocket changes

905670b

Catch missing rates

f30b64a

Copilot AI review requested due to automatic review settings January 2, 2026 12:09

Copilot started reviewing on behalf of springfall2008 January 2, 2026 12:09 View session

[pre-commit.ci lite] apply automatic fixes

31ca817

Copilot AI reviewed Jan 2, 2026

View reviewed changes

springfall2008 added 2 commits January 2, 2026 13:31

HA I/F fixes

e2708bd

Merge branch 'fixes5' of github.com:springfall2008/batpred into fixes5

2dcf734

springfall2008 requested a review from Copilot January 2, 2026 13:33

Copilot started reviewing on behalf of springfall2008 January 2, 2026 13:34 View session

Copilot AI reviewed Jan 2, 2026

View reviewed changes

springfall2008 merged commit b1bbbf9 into main Jan 2, 2026
7 checks passed

springfall2008 deleted the fixes5 branch January 2, 2026 13:38

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

HA Websocket for service calls, catch zero rates #3168

HA Websocket for service calls, catch zero rates #3168

Uh oh!

springfall2008 commented Jan 2, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Jan 2, 2026

Uh oh!

Copilot AI Jan 2, 2026

Uh oh!

Copilot AI Jan 2, 2026

Uh oh!

Copilot AI Jan 2, 2026

Uh oh!

Copilot AI Jan 2, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Jan 2, 2026

Uh oh!

Copilot AI Jan 2, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

-            self.ws_sync_event.wait()
-            self.ws_sync_event.clear()
-            if self.ws_event_loop and self.ws_async_event:
+            # Use a timeout so we can periodically check api_stop and avoid hanging on shutdown
+            triggered = self.ws_sync_event.wait(timeout=1.0)
+            if not triggered:
+                continue
+            self.ws_sync_event.clear()
+            if not self.api_stop and self.ws_event_loop and self.ws_async_event:

	self.log("Warn: Service call {}/{} timed out....".format(domain, service))
	self.log("Warn: Service call {}/{} timed out...".format(domain, service))

Uh oh!

HA Websocket for service calls, catch zero rates #3168

HA Websocket for service calls, catch zero rates #3168

Uh oh!

Conversation

springfall2008 commented Jan 2, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Jan 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Jan 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 2, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants