Annotations lists in dump JSONs - why string formatting?

I am trying to parse the bulk download files for the dataset and I'm having trouble with the annotations for each article. The values in each type of annotations field are lists, but they're formatted as strings. This causes issues for all of the tools I've been using to parse them quickly and efficiently, as these string lists must be converted one-by-one. This is time-prohibitive at scale. I am curious why the values in each annotation sub-field are formatted as strings when no other list-type values in each article's JSON are formatted this way. Is there any chance you could remove the "" quotes from the values in these fields?

Example annotations fields. The lists have double quotes around them. This is in the JSON directly out of one of the dump zip files, and all of the vectorized JSON parsers (pandas, json, duckdb) I've been using interpret these as strings and not as lists.

```
"annotations": {
"abstract": "[{\"end\":1927,\"start\":1259}]",
"author": "[{\"end\":166,\"start\":122},{\"end\":179,\"start\":167},{\"end\":212,\"start\":180},{\"end\":248,\"start\":213},{\"end\":432,\"start\":249},{\"end\":517,\"start\":433},{\"end\":570,\"start\":518},{\"end\":599,\"start\":571},{\"end\":684,\"start\":600},{\"end\":737,\"start\":685},{\"end\":907,\"start\":738},{\"end\":961,\"start\":908}]"
}
```
In contrast, here is another list-type field in the same article's JSON that is not surrounded by quotes:

```
"pdfurls": [
"https://cienciadigital.org/revistacienciadigital2/index.php/AnatomiaDigital/article/download/2791/7024/"
],
```

Why the difference?