Update to Elastic7, all crossref dump format supported

[kermitt2]

Updates:

• rename node.js module "matching" as "indexing" because it is indexing and not matching anything :)
• migrate indexer and lookup to ElasticSearch 7
• support of CrossRef dump as single file .xz (GreeneLab), gzip or uncompressed, or directory of gzip or uncompressed json file (academic torrent or Metadata Plus) ; files can be jsonl or json in an array (indexer and lookup)
• fix potential index refresh bug
• update to gradle 7
• add a nicer progress indexing progress "bar" with colors :)
• update readme

Indexing with CrossRef Torrent Academic dump (January, 7, 2021, 120M records):

• 115,972,356 indexed records
• around 6:30 to index (working on the same time on the computer), 4797 records/s
• 25.94GB index size

I didn't find the courage to integrate the other changes in PR #50 #58 but I will try.

[kermitt2]

Note: I realized that with ES 7, to get the size of the ES index, we need to use the count API. The search API only provides an approximation which is useless (like "more than 10K" for the 160M documents). -> this is fixed in PR #62.

[kermitt2]

follow up in #66