ucdenver-ccp / cartoomics Goto Github PK

With this involve creating a new input file

When cosine similarity is calculated for each node within a path, the index being used is based on the iGraph specification, rather than the entity map created by the embeddings generation step. This causes the incorrect node embedding to be used in the calculation.

• Create config file that maps --knowledge-graph arguments to resource online
• Download KG from online according to --knowledge-graph command
• Remove --knowledge-graph-dir command line option

1. add user input
2. allow users to revise input based on kg search results
3. add txt logs of user inputs
4. sort nodes by ontology
5. present exact matches first

It may be easy for a user to create an _example_input file that has extra fields (i.e., a blank column). Trim extra blank columns to avoid running into this error:

"pandas.errors.ParserError: Error tokenizing data. C error: Expected 3 fields in line 10, saw 4"

shortest path
cs
pdp

PFOCR gene mentions include identifiers (e.g., hgnc_symbol, entrez, chebi) to map onto the KG nodes.