Add querying for KNN

See merge request jjerphan/cython_plus_experiments!1

This reverts commit 27426854.

Reverting to experiment with the original runtime again,
to see if removing the logging helps.

This reverts commit 39363f1c.

Reverting to experiment with the original runtime again,
to see if removing the logging helps.

The target commit removes the logging on contention
which slows the execution of the code.

Round robin is not thread safe.

It would require putting an atomic on ``next_worker``
which would probably slow dispatching queues.

Hence prefering random dispath.

One can now use `taskset(1)` and limit the number of workers
accordingly.

With some uggly duplication because we just want a quick test.

Simplify logic.
Add comments to explain motives.

Taken and adapted from:
https://github.com/scikit-learn/scikit-learn/blob/e4bb9fa86b0df873ad750b6d59090843d9d23d50/sklearn/neighbors/_binary_tree.pxi#L513

Adapt Makefile for python3.9

Update dependencies

This makes use of a class of actors wrapping a numpy array of
the nearest neighbors indices.

Also:
 - Replace depth for leaf_size
 - Make the counter an argument
 - Correctly manage the scheduler lifecycle

The bug was due to a pointer offset

We still have the prom on the secondf n the array.

```
 → p main.py 2> /dev/null
D 0x5609e0520450, 0x5609e05204b0, 0x5609e0520510
I 0x5609e064a580, 0x5609e064a5b0, 0x5609e064a5e0
Depth 10 on dim 0: [0, 24) med: 12 -531495856 -531495760 -531495664     [1, 0, 7, 3, 4, 5, 6, 2, 10, 9, 8, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]
               [1.000000, 0.000000, 7.000000, 3.000000, 4.000000, 5.000000, 6.000000, 2.000000, 10.000000, 9.000000, 8.000000, 11.000000]

D 0x5609e05204b0, 0x5609e05204e0, 0x5609e0520510
I 0x5609e064a5b0, 0x5609e064a5c8, 0x5609e064a5e0
Depth 9 on dim 1: [12, 24) med: 18 -531495760 -531495712 -531495664     [1, 0, 7, 3, 4, 5, 6, 2, 10, 9, 8, 11, 23, 22, 21, 20, 19, 18, 12, 13, 14, 17, 15, 16]
               [1.000000, 0.000000, 7.000000, 3.000000, 4.000000, 5.000000, 6.000000, 2.000000, 10.000000, 9.000000, 8.000000, 11.000000]

D 0x5609e0520450, 0x5609e0520480, 0x5609e05204b0
I 0x5609e064a580, 0x5609e064a598, 0x5609e064a5b0
Depth 9 on dim 1: [0, 12) med: 6 -531495856 -531495808 -531495760       [1, 0, 4, 3, 2, 5, 6, 7, 10, 9, 8, 11, 23, 22, 21, 20, 19, 18, 12, 13, 14, 17, 15, 16]
               [1.000000, 0.000000, 4.000000, 3.000000, 2.000000, 5.000000, 6.000000, 7.000000, 10.000000, 9.000000, 8.000000, 11.000000]

D 0x5609e05204b0, 0x5609e05204c8, 0x5609e05204e0
I 0x5609e064a5b0, 0x5609e064a5bc, 0x5609e064a5c8
Depth 8 on dim 0: [12, 18) med: 15 -531495760 -531495736 -531495712     [1, 0, 4, 3, 2, 5, 6, 7, 10, 9, 8, 11, 23, 21, 19, 20, 18, 22, 12, 13, 14, 17, 15, 16]
               [1.000000, 0.000000, 4.000000, 3.000000, 2.000000, 5.000000, 6.000000, 7.000000, 10.000000, 9.000000, 8.000000, 11.000000]

...
...
...
```

There's a problem with the third partial sort:
```
Depth 9 on dim 1: [6, 12) med: 9 27141056 27141080 27141104     [1, 0, 2, 3, 4, 5, 11, 10, 9, 6, 7, 8]
```

The second hald of the array i.e of indices in [6, 12) is wrong. It looks like it's actually reverted.
It should not be:                                                                 [11, 10, 9, 6, 7, 8]
But something with the pivot (9) in the middle:                                   [ 7,  6, 8, 9,11,10]

From there, it just mess with the second half of the indices' array.

```
C 0x19e2460, 0x19e2478, 0x19e2490
Depth 10 on dim 0: [0, 12) med: 6 27141008 27141056 27141104    [1, 0, 4, 3, 2, 5, 6, 7, 8, 9, 10, 11]
             [1.000000, 0.000000, 4.000000, 3.000000, 2.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0x19e2460, 0x19e246c, 0x19e2478
Depth 9 on dim 1: [0, 6) med: 3 27141008 27141032 27141056      [1, 0, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0x19e2478, 0x19e2484, 0x19e2490
Depth 9 on dim 1: [6, 12) med: 9 27141056 27141080 27141104     [1, 0, 2, 3, 4, 5, 11, 10, 9, 6, 7, 8]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 11.000000, 10.000000, 9.000000, 6.000000, 7.000000, 8.000000]

C 0x19e246c, 0x19e2470, 0x19e2478
Depth 8 on dim 0: [3, 6) med: 4 27141032 27141040 27141056      [1, 0, 2, 3, 4, 5, 11, 10, 9, 6, 7, 8]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 11.000000, 10.000000, 9.000000, 6.000000, 7.000000, 8.000000]

C 0x19e2484, 0x19e2488, 0x19e2490
Depth 8 on dim 0: [9, 12) med: 10 27141080 27141088 27141104    [1, 0, 2, 3, 4, 5, 11, 10, 9, 8, 6, 7]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 11.000000, 10.000000, 9.000000, 8.000000, 6.000000, 7.000000]

C 0x19e2470, 0x19e2474, 0x19e2478
Depth 7 on dim 1: [4, 6) med: 5 27141040 27141048 27141056      [1, 0, 2, 3, 4, 5, 11, 10, 9, 8, 6, 7]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 11.000000, 10.000000, 9.000000, 8.000000, 6.000000, 7.000000]

C 0x19e2488, 0x19e248c, 0x19e2490
Depth 7 on dim 1: [10, 12) med: 11 27141088 27141096 27141104   [1, 0, 2, 3, 4, 5, 11, 10, 9, 8, 7, 6]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 11.000000, 10.000000, 9.000000, 8.000000, 7.000000, 6.000000]

C 0x19e2478, 0x19e247c, 0x19e2484
Depth 8 on dim 0: [6, 9) med: 7 27141056 27141064 27141080      [1, 0, 2, 3, 4, 5, 9, 10, 11, 8, 7, 6]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 9.000000, 10.000000, 11.000000, 8.000000, 7.000000, 6.000000]

C 0x19e247c, 0x19e2480, 0x19e2484
Depth 7 on dim 1: [7, 9) med: 8 27141064 27141072 27141080      [1, 0, 2, 3, 4, 5, 9, 10, 11, 8, 7, 6]
             [1.000000, 0.000000, 2.000000, 3.000000, 4.000000, 5.000000, 9.000000, 10.000000, 11.000000, 8.000000, 7.000000, 6.000000]

C 0x19e2460, 0x19e2464, 0x19e246c
Depth 8 on dim 0: [0, 3) med: 1 27141008 27141016 27141032      [0, 1, 2, 3, 4, 5, 9, 10, 11, 8, 7, 6]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 9.000000, 10.000000, 11.000000, 8.000000, 7.000000, 6.000000]

C 0x19e2464, 0x19e2468, 0x19e246c
Depth 7 on dim 1: [1, 3) med: 2 27141016 27141024 27141032      [0, 1, 2, 3, 4, 5, 9, 10, 11, 8, 7, 6]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 9.000000, 10.000000, 11.000000, 8.000000, 7.000000, 6.000000]

indices[0] = 0
indices[1] = 1
indices[2] = 2
indices[3] = 3
indices[4] = 4
indices[5] = 5
indices[6] = 9
indices[7] = 10
indices[8] = 11
indices[9] = 8
indices[10] = 7
indices[11] = 6
X[indices[0]]= 0.000000 0.000000
X[indices[1]]= 1.000000 1.000000
X[indices[2]]= 2.000000 2.000000
X[indices[3]]= 3.000000 3.000000
X[indices[4]]= 4.000000 4.000000
X[indices[5]]= 5.000000 5.000000
X[indices[6]]= 9.000000 9.000000
X[indices[7]]= 10.000000 10.000000
X[indices[8]]= 11.000000 11.000000
X[indices[9]]= 8.000000 8.000000
X[indices[10]]= 7.000000 7.000000
X[indices[11]]= 6.000000 6.000000
```

C 0xb44460, 0xb44478, 0xb44490
Depth 10 on dim 0: [0, 12) med: 6 11813776 11813824 11813872    [1, 0, 4, 3, 2, 5, 6, 7, 8, 9, 10, 11]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb44478, 0xb44484, 0xb44490
Depth 9 on dim 1: [6, 12) med: 9 11813824 11813848 11813872     [1, 0, 4, 3, 2, 5, 11, 10, 9, 6, 7, 8]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb44460, 0xb4446c, 0xb44478
Depth 9 on dim 1: [0, 6) med: 3 11813776 11813800 11813824      [1, 0, 2, 3, 4, 5, 11, 10, 9, 6, 7, 8]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb44484, 0xb44488, 0xb44490
Depth 8 on dim 0: [3, 6) med: 4 11813800 11813808 11813824      [1, 0, 2, 3, 4, 5, 11, 10, 9, 6, 7, 8]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb4446c, 0xb44470, 0xb44478
Depth 8 on dim 0: [9, 12) med: 10 11813848 11813856 11813872    [1, 0, 2, 3, 4, 5, 11, 10, 9, 8, 6, 7]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb44460, 0xb44464, 0xb4446c
Depth 8 on dim 0: [0, 3) med: 1 11813776 11813784 11813800      [0, 1, 2, 3, 4, 5, 11, 10, 9, 8, 6, 7]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb44488, 0xb4448c, 0xb44490
Depth 7 on dim 1: [10, 12) med: 11 11813856 11813864 11813872   [0, 1, 2, 3, 4, 5, 11, 10, 9, 8, 7, 6]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb44470, 0xb44474, 0xb44478
Depth 7 on dim 1: [4, 6) med: 5 11813808 11813816 11813824      [0, 1, 2, 3, 4, 5, 11, 10, 9, 8, 7, 6]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C C 0xb44478, 0xb4447c, 0xb44484
Depth 8 on dim 0: [6, 9) med: 7 11813824 11813832 11813848      [0, 1, 2, 3, 4, 5, 9, 10, 11, 8, 7, 6]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

0xb44464, 0xb44468, 0xb4446c
Depth 7 on dim 1: [1, 3) med: 2 11813784 11813792 11813800      [0, 1, 2, 3, 4, 5, 9, 10, 11, 8, 7, 6]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]

C 0xb4447c, 0xb44480, 0xb44484
Depth 7 on dim 1: [7, 9) med: 8 11813832 11813840 11813848      [0, 1, 2, 3, 4, 5, 9, 10, 11, 8, 7, 6]
             [0.000000, 1.000000, 2.000000, 3.000000, 4.000000, 5.000000, 6.000000, 7.000000, 8.000000, 9.000000, 10.000000, 11.000000]