Fix socketio issue when using custom payload outputs (#144)

* do not unpack json payload if data key is not present

* add room arg to else branch

* prepared release of version 3.7.3.dev1 (#151)

* Prepare-release-3.7.3.dev2 (#164)

* prepared release of version 3.7.3.dev2

* allow dev releases without changelogs

* add changelog entry

---------

Co-authored-by: Shailendra Paliwal <hello@shailendra.me>

CHANGELOG.mdx

@@ -1,6 +1,6 @@
 ## [3.7.5] - 2024-01-24
-                       
-Rasa 3.7.5 (2024-01-24)                        
+
+Rasa 3.7.5 (2024-01-24)
 ### Improvements
 - [#193](https://github.com/rasahq/rasa/issues/193): Add new embedding types: `huggingface` and `huggingface_bge`. These new types import the `HuggingFaceEmbeddings` and `HuggingFaceBgeEmbeddings` embedding classes from Langchain.
 
@@ -12,8 +12,8 @@ Rasa 3.7.5 (2024-01-24)
 
 
 ## [3.7.4] - 2024-01-03
-                       
-Rasa 3.7.4 (2024-01-03)                        
+
+Rasa 3.7.4 (2024-01-03)
 ### Improvements
 - [#142](https://github.com/rasahq/rasa/issues/142): Add embeddings type `azure` to simplify azure configurations, particularly when using Enterprise Search Policy
 
@@ -22,8 +22,8 @@ Rasa 3.7.4 (2024-01-03)
 
 
 ## [3.7.3] - 2023-12-21
-                       
-Rasa 3.7.3 (2023-12-21)                        
+
+Rasa 3.7.3 (2023-12-21)
 ### Improvements
 - [#133](https://github.com/rasahq/rasa/issues/133): Persist prompt as part of the model and reread prompt from the model storage instead of original file path during loading. Impacts LLMCommandGenerator.
 - [#141](https://github.com/rasahq/rasa/issues/141): Replaced soon to be depracted text-davinci-003 model with gpt-3.5-turbo. Affects components - LLM Intent Classifier and Contextual Response Rephraser.

changelog/144.bugfix.md

@@ -0,0 +1,3 @@
+Fix reported issue, e.g. https://github.com/RasaHQ/rasa/issues/5461 in Rasa Pro:
+Do not unpack json payload if `data` key is not present in the response custom output payloads when using socketio channel.
+This allows assistants which use custom output payloads to work with the Rasa Inspector debugging tool.

rasa/core/channels/socketio.py

@@ -115,9 +115,11 @@ async def send_custom_json(
         self, recipient_id: Text, json_message: Dict[Text, Any], **kwargs: Any
     ) -> None:
         """Sends custom json to the output."""
-        json_message.setdefault("room", recipient_id)
-
-        await self.sio.emit(self.bot_message_evt, **json_message)
+        if "data" in json_message:
+            json_message.setdefault("room", recipient_id)
+            await self.sio.emit(self.bot_message_evt, **json_message)
+        else:
+            await self.sio.emit(self.bot_message_evt, json_message, room=recipient_id)
 
     async def send_attachment(  # type: ignore[override]
         self, recipient_id: Text, attachment: Dict[Text, Any], **kwargs: Any

scripts/publish_gh_release_notes.py

@@ -81,7 +81,7 @@ def main():
         return 1
 
     version = Version(tag_name)
-    if version.pre:
+    if version.pre or version.dev:
         md_body = "_Pre-release version_"
     else:
         md_body = parse_changelog(tag_name)