read_json() dtype should support type[str]

[RobertCraigie]

Describe the bug

Currently the stub files for read_json() only support bool | Mapping[HashableT, DtypeArg] | None but at runtime, dtype=str works.

To Reproduce

import pandas as pd

df = pd.read_json('data.jsonl', lines=True, dtype=str).fillna("")

Where data.jsonl contains

{"prompt": "Translate the following English text to French: 'Hello, how are you?'", "completion": "Bonjour, comment ça va ?"}
{"prompt": "What is the capital of France?", "completion": "Paris"}
{"prompt": "Solve the following math problem: 2 + 2 = ?", "completion": "4"}

Please complete the following information:

• OS: [e.g. Windows, Linux, MacOS] MacOS
• OS Version [e.g. 22]
• python version 3.9
• version of type checker Pyright 1.1.326
• version of installed pandas-stubs latest

Additional context

Apologies if this is incorrect, I personally have very limited pandas experience, I simply ran into this while porting someone elses code from an untyped codebase to a typed codebase.

[twoertwein]

Based on the documentation, it has to be a bool or a dict of bools. I assume pandas interprets str as True'ish

[RobertCraigie]

Ah sorry I should've just read the docs, thanks!