Closes #59329
Relates to #59352

We switch the format of the results download from CSV to line-separated JSON.
Each line corresponds to a JSON object containing chunk matches.
The JSON object has the same format as the matches served by the Stream API.

example.json

This is a breaking change motivated by customer feedback.
Search Jobs is still released as EAP so a breaking change is acceptable.

Pros:

• richer information (matches, content, positions)
• supports all result types
• same format as Stream API

Cons:

• Requires more storage
• Not as easy to parse by a human as a CSV

The commits can be reviewed separatedly.
The first 2 commits contain the core of this change.
In the third commit I delete code and update tests

Next: update documentation

Test plan

• updated and new units tests
• manual testing
  • I used this Python script to get the result counts from the downloaded json and compared them to the counts in the web app for a couple of examples.

"""
Usage:
    python main.py <path/to/json>
"""
import json
import argparse


def count_results(filename):
    return sum(len(chunks['ranges']) for line in open(filename) for chunks in json.loads(line)['chunkMatches'])


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('filename', type=str, help='Search Jobs result json')
    args = parser.parse_args()

    print(f'{count_results(args.filename)} results')