-
Notifications
You must be signed in to change notification settings - Fork 16
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
Gml 1824 regression test with cypher support and no hallucination and usefulness check #248
Gml 1824 regression test with cypher support and no hallucination and usefulness check #248
Conversation
Gml 1793 UI dialog
Minor UI fixes
url sources as links
…-in-docker-compose GML-1796 dependency cycle detecte
Render markdown
…ementation + UI styling updates (#229) * [GML-1779-show-response-as-table-or-graph] graph interface added - mock data * GML-1797 - Investigate ECC Updated ECC to let workers work take a processing node after the expire window. Added more logging and options during configuration * Lots of cleanup of the ECC * [GML-1779-show-response-as-table-or-graph] updates to graph and table, styling updates * [GML-1779-show-response-as-table-or-graph] table vis advancements * [GML-1779-show-response-as-table-or-graph] re-do updates --------- Co-authored-by: Ahmed Albadri <[email protected]> Co-authored-by: RobRossmiller-TG <[email protected]>
* condition tweak for table vis * cleanup, styling tweaks
* condition tweak for table vis * cleanup, styling tweaks * caret typewriter animation edit
…umber-of-transactions
…umber-of-transactions
@@ -19,6 +19,10 @@ def __init__(self, config): | |||
from langchain.chat_models import ChatOpenAI | |||
|
|||
model_name = config["llm_model"] | |||
if graphname == "Transaction_Fraud": |
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.
This shouldn't be here
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.
If I remember correctly, @RobRossmiller-TG added it to ensure better performance for demo. since gpt-4 works better on transaction fraud and inquiryAI.
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.
Yeah I don't think that we should ship that though. For the demo it worked well.
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.
sounds good. got it removed now
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.
It was removed in the dev branch unless I missed it. Is this based on the latest dev?
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.
Yes. I also rebase it. It should be based on the latest dev branch now.
copilot/app/agent/agent.py
Outdated
@@ -156,7 +165,7 @@ def question_for_agent( | |||
def make_agent(graphname, conn, use_cypher, ws: WebSocket = None, supportai_retriever="hnsw_overlap") -> TigerGraphAgent: | |||
if llm_config["completion_service"]["llm_service"].lower() == "openai": | |||
llm_service_name = "openai" | |||
llm_provider = OpenAI(llm_config["completion_service"]) | |||
llm_provider = OpenAI(llm_config["completion_service"], graphname=graphname) |
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.
Remove the graphname info here too
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.
removed it now.
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 want to optionally enable cypher generation
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 shouldn't completely remove cypher
Sure. I have added cypher support back as optional now. |
copilot/app/agent/agent.py
Outdated
@@ -90,7 +91,7 @@ def __init__( | |||
self.embedding_store, | |||
self.mq2s, | |||
self.gen_func, | |||
cypher_gen_tool=self.cypher_tool, | |||
# cypher_gen_tool=self.cypher_tool, |
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.
Just double check, is cypher support still available and optional?
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.
good catch. just added it back. should be good to go.
…hen asking how many
for key, value in output.items(): | ||
logger.info(f"testing steps {key}: {value}") |
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.
Missed this line. Do we need to log all the questions and generated answer?
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 can comment it out. It was trying to understand which step went wrong during testing. I will remove it in next PR.
regression test with cypher support only but not hallucination and usefulness check