gr.Dataframe with datatype="auto" crashes on empty list or 1D numpy array

[ParamChordiya]

Describe the bug

`gr.Dataframe` with `datatype="auto"` crashes with `IndexError` when given an empty list or a 1D numpy array as the value.

Root cause: In `gradio/components/dataframe.py`, the `set_auto_datatype` method has two crash paths:

1. Empty list (line 583):

elif isinstance(value, list):
    self.datatype = [... for val in value[0]]  # IndexError: list index out of range

2. 1D numpy array (lines 568, 572):

elif isinstance(value, np.ndarray):
    self.datatype = [
        ... str(type(value[0, i]).__name__) ...  # IndexError: too many indices for array
        for i in range(value.shape[1])            # IndexError: tuple index out of range
    ]

A 1D array doesn't support 2D indexing (`value[0, i]`), and `value.shape[1]` fails because a 1D shape tuple has no second element.

Have you searched existing issues?

Yes — no existing issue covers this.

Reproduction

import gradio as gr
import numpy as np

# Crash 1: empty list
gr.Dataframe(value=[], datatype="auto")
# IndexError: list index out of range

# Crash 2: 1D numpy array
gr.Dataframe(value=np.array([1, 2, 3]), datatype="auto")
# IndexError: tuple index out of range

Expected behavior

Both cases should fall back gracefully to the default `"str"` datatype instead of crashing. The component already has an `is_empty` method that properly handles empty inputs — `set_auto_datatype` should use similar guards.

Suggested fix:

elif isinstance(value, np.ndarray):
    if value.ndim < 2 or value.size == 0:
        self.datatype = "str"
        return
    # ... existing logic

elif isinstance(value, list):
    if len(value) == 0:
        self.datatype = "str"
        return
    # ... existing logic

System info

Reproduced on `main` branch, `gradio/components/dataframe.py` lines 564-584.

[onthebed]

Looking into this — I’m tracing how Dataframe infers for empty and 1D inputs so I can see why that path crashes instead of normalizing the shape first. Happy to send a fix if I can reproduce it cleanly.