#527: Update JSON generator to work with vt's LBDataHolder

[tlamonthezie]

Fixes #527
Fixes #531

This PR provides the following changes;

1. Add packed id support in entities
   1.2 The Object class -> __index renamed as seq_id, + addition of a packed_id field
2. Add collection_id field
   2.1 Object class: Add the collection_id field to the attributes of the Object class, add objgroup_id field to unused params.
   2.2 JSON Generator: Enable the collection_id field to be set by the JSON generator (JSONDataFilesMaker) (with interactive mode support)
   2.3 set collection_id automatically with some default value of 7 if object is added as migratable to a rank (collection_id required by vt for migratable objects so required for the LBAF Json writer).
   with the help of the new TaskSpecification typed dict defining both time (required) and collection_id (not required)
3. VTDataWriter: copy object fields to the entity node inside the communication from or to nodes
4. Adapt unit tests (code, config) to work with these source updates
5. Update data sets with small entity ids by replacing id by seq_id (considering small ids sequences are in fact seg_id's) regenerate to get data in communiaction endpoints nodes and add some generator config file
6. Replace shared_block_id by shared_id everywhere

Note: this PR is also needed to adapt to the new LB Data files schema defined by VT PR 2243

IMPORTANT: vt-tv must be updated as well to support seq_id as object identifiers.
Else we have to generate entity id values in LBAF if missing in input data

[ppebay]

Thanks for changes @tlamonthezie

[ppebay]

Requesting now a SNL review @nlslatt @lifflander due to connection with vt

[tlamonthezie]

#531 now included in the PR

[cwschilly]

The newly generated synthetic-blocks data works with vt 👍

[cwschilly]

LGTM (and works with vt tests)

[nlslatt]

Why is there a python script to "copy" json files by reading, decompressing, converting to a dict, converting back to a string, compressing, and writing? This seems very inefficient.

[tlamonthezie]

Hello @nlslatt this script was in fact very helpful to rename keys in some compressed datasets.
The script can be called for example to de-compress a dataset, then you can update the dataset using a text editor, then call the script again to compress.
It was particularly helpful in this issue to rename id to seq_id in some compressed datasets (in particular tests datasets)
For future schema updates, it might be useful to keep this script