Update DistNeighborSampler for hetero

[kgajdamo]

The purpose of this PR is to improve distributed hetero sampling algorithm.
IMPORTANT INFO: This PR is complementary with #284 from pyg-lib. The pyg-lib one needs to be merged for this one to work properly.

Description: (sorry if too long)
Distributed hetero neighbor sampling is a procedure analogous to homo sampling, but more complicated due to the presence of different types of nodes and edges.
Sampling in distributed training imitates the hetero_neighbor_sample() function in pyg-lib. Therefore, the mechanism of action and the nomenclature of variables are similar.
Due to the fact that in distributed training, after sampling each layer, it is necessary to synchronize the results between machines, the loop iterating through the layers was implemented in Python.

The main two loops iterate sequentially: over layers and edge types. Inside the loop, the sample_one_hop() function is called, which performs sampling for one layer.
The input to the sample_one_hop() function is data of a specific type, so its execution is almost identical to homo.
The sample_one_hop() function, depending on whether the input nodes are located on a given partition or a remote one, performs sampling or sends an RPC request to the remote machine to do so. The dist_neighbor_sample()->neighbor_sample() function is used for sampling. Nodes are sampled with duplicates so that they can later be used to construct local to global node mappings.
When all machines have finished sampling, their outputs are merged and synchronized in the same way as for homo.
Then the results return to the node_sample() function where they are written to the output dictionaries and the src nodes for the next layer are calculated.
After going through all the layers, the global node indices are finally mapped to the local ones in the hetero_dist_relabel() function.

Information about some of the variables used in a node_sample() function:
node_dict - class storing information about nodes. It has three fields: src, with_dupl, out, which are described in more detail in the distributed/utils.py file.
batch_dict - class used when sampling with the disjoint option. It stores information about the affiliation of nodes to subgraphs. Just like node_dict, it has three fields: src, with_dupl, out.
sampled_nbrs_per_node_dict - a dictionary that stores information about the number of sampled neighbors by each src node. To facilitate subsequent operations, for each edge type is additionally divided into layers.
num_sampled_nodes_dict, num_sampled_edges_dict - needed for HGAM to work.

[codecov]

Codecov Report

Attention: 5 lines in your changes are missing coverage. Please review.

Comparison is base (331e5d1) 88.90% compared to head (441df26) 89.19%.

❗ Current head 441df26 differs from pull request most recent head ac9e9f8. Consider uploading reports for the commit ac9e9f8 to get more accurate results

Files	Patch %	Lines
...rch_geometric/distributed/dist_neighbor_sampler.py	90.19%	5 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #8503      +/-   ##
==========================================
+ Coverage   88.90%   89.19%   +0.29%     
==========================================
  Files         480      480              
  Lines       30563    30619      +56     
==========================================
+ Hits        27171    27310     +139     
+ Misses       3392     3309      -83     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[JakubPietrakIntel]

Thanks Kinga, looks amazing.
Depending if #8605 comes in before that, we might do some adjustments in test_dist_neighbor_loader.py to account for a different RPC shutdown method or I'll make corrections in the other PR.