Add DotProduct Method and README Example for Embedding Similarity Search

[ealvar3z]

Describe the change

• Implement a DotProduct() method for the Embedding struct to calculate the dot product between two embeddings.
• Add a custom error type for vector length mismatch.

Describe your solution

• Finish implementing the following requested changes from @sashabaranov:

We don't need two huge examples [see](https://github.com/sashabaranov/go-openai/pull/274/files#r1232058195) — Please make one small example without encoding/gob etc.
I'm not sure this change is relevant [here](https://github.com/sashabaranov/go-openai/pull/274/files#diff-b335630551682c19a781afebcf4d07bf978fb1f8ac04c6bf87428ed5106870f5L596)
Let's make DotProduct a method of Embedding struct [see](https://github.com/sashabaranov/go-openai/pull/274/files#r1232242888)

Tests

• Add unit tests to validate the new DotProduct() method and error handling.

Additional context

• Update README.md with a complete example demonstrating how to perform an embedding similarity search for user queries.

Issue: #XXXX

[codecov]

Codecov Report

Merging #492 (4a7e7c0) into master (3589837) will increase coverage by 0.15%.
Report is 1 commits behind head on master.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #492      +/-   ##
==========================================
+ Coverage   97.30%   97.46%   +0.15%     
==========================================
  Files          18       18              
  Lines         780      829      +49     
==========================================
+ Hits          759      808      +49     
  Misses         15       15              
  Partials        6        6              

Files Changed	Coverage Δ
embeddings.go	100.00% <100.00%> (ø)

[sashabaranov]

Thank you so much, this was coming for a long time :D