Add support for variadic arguments to SCRIPT

[lantiga]

This PR addresses #392.
A slightly different command API was chosen to keep things efficient.

We use the $ sign to signal that INPUTS following that sign will be passed to the script as a list:

redis-cli -x AI.SCRIPTSET foo CPU SOURCE < foo.py
redis-cli AI.TENSORSET a FLOAT 1 1 VALUES 1
redis-cli AI.TENSORSET b FLOAT 1 1 VALUES 1
redis-cli AI.TENSORSET c FLOAT 1 1 VALUES 1
redis-cli AI.SCRIPTRUN foo foo INPUTS a $ b c OUTPUTS d
redis-cli AI.TENSORGET d META VALUES

The corresponding script will need to be type-annotated with List[Tensor] for the last argument

def foo(a : Tensor, args: List[Tensor]):
    return a + torch.stack(args).sum()

Within the script, args will act as a standard Python list containing tensors.

Test has been added and docs have been updated.

[codecov]

Codecov Report

Merging #395 into master will decrease coverage by 0.07%.
The diff coverage is 90.00%.

@@            Coverage Diff             @@
##           master     #395      +/-   ##
==========================================
- Coverage   72.79%   72.72%   -0.08%     
==========================================
  Files          21       21              
  Lines        4345     4359      +14     
==========================================
+ Hits         3163     3170       +7     
- Misses       1182     1189       +7     

Impacted Files	Coverage Δ
src/backends/torch.c	84.37% <ø> (ø)
src/script.c	66.86% <66.66%> (+0.19%)	⬆️
src/libtorch_c/torch_c.cpp	56.77% <100.00%> (+1.64%)	⬆️
src/redisai.c	76.93% <0.00%> (-1.02%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0557118...b7388a8. Read the comment docs.

[filipecosta90]

LGTM @lantiga . WDYT of adding some more negative tests with variadic, just to ensure that the parsing, and run will always return the expected errors,etc... ?