[Vector sets]: redis-cli recall testing abilities

[antirez]

Vector sets have the ability to also ask for ground truth performing an O(N) scan. This allows to perform a recall test against any key holding a vector set, allowing users to verify what is the best EF value to use and how HNSW performs depending on the data set on a given key (the level of clustering changes significantly how vectors near/far a cluster will behave).

Example:

./redis-cli --vset-recall fasttext_i --vset-recall-ef 500 --vset-recall-ele 3 --vset-recall-count 100

# Testing recall for vector set: fasttext_i (dimension: 300)
# Mixing 3 random element vectors, top 100 results, EF=500

Queries: 7024 | Avg recall: 98.96%^C

====================================
       Recall Test Results
====================================

Total queries:   7158
Average recall:  98.95%
Mean recall:     99.03%
Median recall:   99.18%
StdDev:          1.85%
Min recall:      41.91%
Max recall:      100.46%

--- Recall Thresholds ---
At least    % of queries
--------    ------------
   0.0%         100.00%
  50.0%          99.90%
  60.0%          99.90%
  70.0%          99.90%
  80.0%          99.90%
  85.0%          99.80%
  90.0%          99.70%
  95.0%          98.40%
  99.0%          74.10%
 100.0%          48.90%

[jit-ci[bot]]

Hi, I’m Jit, a friendly security platform designed to help developers build secure applications from day zero with an MVS (Minimal viable security) mindset.

In case there are security findings, they will be communicated to you as a comment inside the PR.

Hope you’ll enjoy using Jit.

Questions? Comments? Want to learn more? Get in touch with us.

[antirez]

This "jit-ci" things looks like a false positive.

[antirez]

Would love to see this merged. Thanks.

[sundb]

@antirez The code has been frozen. How about making it in 8.6?