Allow parsing of partial CSV strings

There are some cool things going on with the `influxdb-client-js` library (namely utilizing the streaming interface it exposed from `fetch`), that we can't support, because we can't parse partial CSVs in the fromFlux parser.

There have been talks about appending / truncating the data based on a shifting time frame (ie: progressive updates for dashboards). This is not [that](https://github.com/influxdata/influxdb/issues/19793). There's also talks about adding a pagination interface to expose partial views of the dataset. This is not that.

Instead of waiting for the full CSV to come in from the API before parsing, we want to be able to generate an instance of a parser, and append partial CSV strings like:
```
#group,false,false,tr
```

or

```
true,true,false,false
#datat
```

or

```
,,2,2020-10-23T15:14:52.865415455Z,2020-10-23T16:14:52.865415455Z,usage_percent,docker_container_cpu,plexinc/pms-docker,plex,runnin
```

and have that append rows to the data set as they come in

Start with the fromFlux parser entry point named in the issue and review how it currently consumes a complete CSV string. Define the incremental parser behavior using the partial CSV examples, then verify that appended fragments produce the corresponding rows as data arrives.

Written by the indexing model from the issue text.