feat/primekg loader #49
Conversation
There was a problem hiding this comment.
Nice work 💪. The notebooks were very easy to follow and demonstrate the functionality well which will hopefully come in handy for our hackathon participants.
Please take a look at my questions/comments and let me know if there is anything that needs to be clarified. In principle, I think everything looks good pending my assumptions I note are correct 😉.
aiagents4pharma/talk2knowledgegraphs/tests/test_biobridge_primekg_loader.py
Show resolved
Hide resolved
aiagents4pharma/talk2knowledgegraphs/tests/test_biobridge_primekg_loader.py
Show resolved
Hide resolved
There was a problem hiding this comment.
Are there any pre-trained embeddings for stark?
There was a problem hiding this comment.
I have rechecked the repository of Stark about this, and apparently they provided the query and node embeddings using 'text-embedding-ada-002'. However, there is no information on edge embeddings. I have updated the class of StarkQAPrimeKG by including pre-loaded embeddings of queries and nodes. However, we also need to embed the edges (~8M) using the 'text-embedding-ada-002' in the next step (KG construction). Alternatively, I am preparing codes to embed both nodes and edges using Ollama's nomic-embed-text to be included in the next PR.
Ref:
https://github.com/snap-stanford/stark/blob/main/emb_download.py
There was a problem hiding this comment.
I see. I do not believe they used edge embeddings in their manuscript (?).
I wouldn't worry about embedding the edges using text-embedding-ada-002 if time is of the essence. It does not seem that it would help us reproduce their work. Please correct me if I am wrong.
There was a problem hiding this comment.
Correct, in their manuscript and code repository, they didn't mention edge/relation embeddings. Thus, the methods used in their benchmark most likely didn't incorporate these features.
| self.starkqa_node_info: dict = None | ||
| self.query_emb_dict: dict = None | ||
| self.node_emb_dict: dict = None | ||
|
|
There was a problem hiding this comment.
Fine to include the relation embeddings in a subsequent PR as mentioned in another comment.
|
|
||
| return starkqa, starkqa_split_idx, starkqa_node_info | ||
|
|
||
| def _load_stark_embeddings(self) -> tuple: |
|
🎉 This PR is included in version 1.5.0 🎉 The release is available on GitHub release Your semantic-release bot 📦🚀 |
For authors
Description
Please:
I have developed a set of codes for loading a series of biomedical knowledge graph-based datasets, i.e., PrimeKG, StarkQA-PrimeKG, and BioBridge-PrimeKG. Following are the detailed key updates:
aiagents4pharma/talk2knowledgegraphs/datasets, I included an abstract class ofDatasetfollowed by a set of implementation classes ofPrimeKG,StarkQAPrimeKG, andBioBridgePrimeKGfor loading the corresponding datasets.aiagents4pharma/talk2knowledgegraphs/tests/*docs/notebooks/talk2knowledgegraphs/*.docs/talk2knowledgegraphs/*.Fixes #32
Type of change
Please delete options that are not relevant.
How Has This Been Tested?
Please describe the tests you conducted to verify your changes. These may involve creating new test scripts or updating existing ones.
testsfolderaiagents4pharma/talk2knowledgegraphs/tests/*)Checklist
testsfolder) that prove my fix is effective or that my feature worksFor reviewers
Checklist pre-approval
Checklist post-approval
developintomain? If so, please make sure to add a prefix (feat/fix/chore) and/or a suffix BREAKING CHANGE (if it's a major release) to your commit message.Checklist post-merge
developintomainand is it suppose to run an automated release workflow (if applicable)? If so, please make sure to check under the "Actions" tab to see if the workflow has been initiated, and return later to verify that it has completed successfully.