rgu-computing / clood Goto Github PK

Reuse strategies for Boolean data type should be NN value (nearest neighbour), Majority, Minority.

We need to secure the API layer and also create a mechanism to protect the dashboard functions

@ike01 Need to fix or add in the readme on how to solve the issue. Is it directly from dockerZip: false?

This should support the use word embeddings generated by some model as attribute values

1. New paper at ICCBR add link and add poster link
2. Add key features of clood
3. Add usecases of using clood (optional)
4. Add the new SBert method details
5. Update new architecture diagram
6. Create a navigatable readme

• Filter Step and UI Changes @ike01
• Array UI Fixes @Andrew-McLeman
• Array Explanation @Andrew-McLeman
• Array Data Type Changes @ike01
• Project UI Improvement @chamathpali
• Example pre-loaded case @chamathpali
• Settings page improvements and dropdown @chamathpali

1. Filtering step for Retrieval Queries
2. Local Similarity values for retrieved cases need to be in the output
3. Clood similarity visualiser image (linked to 2)

Configuration not created for freshly running docker setups,
Can we find a way to automate the configuration process for new installations or are we depending on the API below?
http://localhost:3000/dev/config/create

In order to prevent local mismatches

we need better api validation and responses for invalid API calls

When we get 403 log out the user and remove from local storage
{"statusCode":403,"error":"Forbidden","message":"User is not authorized to access this resource"}

@Andrew-McLeman

E.g. when a user tries to retain a vector (array) attribute value with a dimension that is greater than was specified in the configuration

Objects needs to parsed maybe JSON.stringify?

Displays that show that a request is awaiting a response on the dashboard

[Object] would be returned in the CSV. better do a JSON stringify or a similar way to export the json structure?
@Andrew-McLeman

If large text we need to show hide text e.g. https://www.w3schools.com/howto/howto_js_read_more.asp
@Andrew-McLeman

@Andrew-McLeman

https://www.sbert.net/docs/installation.html

• This will be under other-sim

If we are to migrate to a new UI framework, following pages would be required

• Login
• Projects
  • Create Project
  • List Projects
• Inner Project
  • Attribute Configuration
  • Add bulk Cases (CSV)
  • Add single case
• CBR Tasks
  • Retrieve
  • Reuse
  • Retain
• Tokens (Generate and List)
• Settings/Config

Determining min, max and range parameters for similarity metrics like McSherry and Interval will reduce the burden on users to keep these values up to date. Parameter values should be checked/updated on data insertion. Handle situations where the range is zero which could lead to division by zero errors.

E.g. in CSV upload method can we parse before the POST request? or validate on the backend if the JSON is valid for Object type?

Ontology concept attribute configuration uses ontology id of the exported project leading to incorrect reference when trying to access the similarity index

This feature will be very useful to share same case structure among teams.

A root node can be specified in the configuration of path-based similarity such that similarity computation is limited to nodes in a sub-tree of the ontology. When determining the depth of nodes in the Wu & Palmer similarity computation, the distance to the actual root node of the ontology should be used instead of the distance to the user-specified root.