feat(pipeline): support reading pretty printed json

[reubenmiller]

Improve pipeline parsing to also accept pretty printed json objects, and also mixing between json objects and non json objects.

The following pipeline formats are now supported:

• pretty printed json
• multiple json objects on the same line, e.g. {"name":"1"}{"name":"2"}
• mixing json and text input
• trailing newline is no longer required when reading files
• lines with double quotes surrounding the string

Example: Mixed input

Given the following file input.txt

{"name": "device01"}{"name": "device02"}
    {"name": "device03"}
    device04
"device05"
{
    "name": "device06"
}

You can pipe the above file to c8y:

cat input.txt | c8y devices create --dry

Which will create 5 devices:

| id           | name          |
|--------------|---------------|
| 6424206      | device01      |
| 4924207      | device02      |
| 7822239      | device03      |
| 3422240      | device04      |
| 5223225      | device05      |
| 5264003      | device06      |

Previously, the previously more strict parsing would of only created 3 devices (as invalid lines would of been ignored)

| id           | name          |
|--------------|---------------|
| 6424206      | device01      |
| 7822239      | device03      |
| 3422240      | device04      |