-
Notifications
You must be signed in to change notification settings - Fork 4.7k
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.
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
Loading and predicting with graph based model. #9747
Loading and predicting with graph based model. #9747
Conversation
520d4aa
to
5885e5b
Compare
Wohooo! 🎉 Amazing work! Will have a look tomorrow 🔎 |
These are the public methods of the processor that end up with a graph run:
|
self.domain = domain | ||
self.tracker_store = tracker_store | ||
self.lock_store = lock_store | ||
self.max_number_of_predictions = max_number_of_predictions | ||
self.message_preprocessor = message_preprocessor |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We probably need to include this in the migration guide.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'll continue tomorrow morning. So far it's been a very pleasant read. Considering that we are avoiding refactorings for now just this state is already multiple factors better than what we had 💯
data/graph_schemas/config_pretrained_embeddings_mitie_predict_schema.yml
Outdated
Show resolved
Hide resolved
@@ -557,47 +431,42 @@ async def parse_message_using_nlu_interpreter( | |||
} | |||
|
|||
""" | |||
|
|||
processor = self.create_processor() | |||
if not self.is_ready(): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Does it make sense to make this a decorator? Somewhat duplicate with ensure_loaded_agent
in server
🤔
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Next batch of comments - I think I'm mostly missing the tests now
parsed_message = processor.interpreter.featurize_message( | ||
Message(data={TEXT: text}) | ||
) | ||
parsed_message = processor.parse_message(UserMessage(text=text)) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
don't we only do this when having e2e
which means that we already have the tokens from the time when we verified the NLU test result?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Amazing stuff!
…ure-revamp/9604/prediction
…604/prediction/cache-fix fix cache usage in tests
…604/prediction/clazz-impl-in-fingerprint include class implementation in fingerprint
…RasaHQ/rasa into 3.0-architecture-revamp/9604/prediction
…768/project_fingerprint add project fingerprint to model metadata
…RasaHQ/rasa into 3.0-architecture-revamp/9604/prediction
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great work! Requesting changes due to the raise
at the server start
parsed_message = processor.interpreter.featurize_message( | ||
Message(data={TEXT: text}) | ||
) | ||
parsed_message = processor.parse_message(UserMessage(text=text)) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
) | ||
new_agent.lock_store = app.agent.lock_store |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't really get why we have to do the lockstore separately but as it was done like that before 🤷🏻
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
💯 🎸 🚀
Proposed changes:
This implements most of the functionality for loading and predicting with the graph runner.
TODO:
yaml_store_reader
in shared is usingregex_message_handler
.nlu
/core
)?