Resolve m types #351
Resolve m types #351
Conversation
|
I am wondering whether we should not modify some (or all) of the However, please feel free to challenge the above. |
| if m_types_response["pagination"]["total_items"] > page_size: | ||
| raise ValueError( | ||
| "Not all m-types were retreived, please increase the page size." | ||
| ) | ||
|
|
There was a problem hiding this comment.
I suggest we paginate instead of raising an error. For instance, we can remove the -p argument from the argparse, set manually a page_size and do something along the lines of:
retrieved_items = []
page = 1
while m_types_response["pagination"]["total_items"] > len(retrieved_items):
response = await httpx_client.get(..., params={"page_size": page_size, "page": page}
retrieved_items.extend(...)
page += 1| # Gather the names | ||
| pref_labels = [m_types.pref_label for m_types in m_types.mtypes] |
There was a problem hiding this comment.
Did we ever get an answer from Daniela about what field scientists will use to search for an mtype ?
There was a problem hiding this comment.
Not really, I guess there are more important things to work on for CNS
| ] = """Resolve the mtype pref label from natural english to its corresponding ID (formated ad UUID) using semantic search. | ||
| Accepts natural language inputs containing the full or partial name. If no exact match is found, returns the best scored candidates.""" |
There was a problem hiding this comment.
I am not sure if the LLM knows what the pref_label is. Not sure how I would say it better though.
|
We discussed this privately and we decided not to move forward with resolution tools. So @BoBer78 sorry for the extra work but I think we can just close this PR. |
Sure, let's maybe wait for a discussion with JDC to know how we proceed with the resolving. Then we close this PR. |
closes #338