Cannot append to REQUIRED field when using `client.load_table_from_file` without providing table schema

#### Environment details

- OS type and version: Ubuntu 20.04 (WSL)
- Python version: 3.11.8
- pip version: 24.1.2
- `google-cloud-bigquery` version: 3.25.0

#### Steps to reproduce

1. Create a table that has a single field with mode=REQUIRED
2. Attempt to append to the table using `client.load_table_from_file` with a parquet file written from memory to a `BytesIO` buffer. The library writing to the buffer either does not have an option of nullable/required fields (Polars), or `nullable=False` is provided to the field (PyArrow).

#### Issue details

I am unable to use `client.load_table_from_file` to append to an existing table with a REQUIRED field, without providing the table schema in the `LoadJobConfig`. The docs say that the schema does not need to be supplied if the table already exits.

This issue is similar to https://github.com/googleapis/google-cloud-python/issues/8093, but relates to `load_table_from_file` rather than `load_table_from_dataframe`. It is also somewhat related to https://github.com/googleapis/google-cloud-python/issues/8142 (as explicitly suppling the BigQuery table schema fixes the issue), but again this relates to `load_table_from_file` rather than `load_table_from_dataframe`.

As an aside, the fix should definitely not require PyArrow. The current Polars code functions without PyArrow if the table BigQuery schema is provided.

I am filing this as a bug rather than a feature request as the [docs for `schema` in `JobConfigurationLoad`](https://cloud.google.com/bigquery/docs/reference/rest/v2/Job#JobConfigurationLoad.FIELDS.schema) say
> The schema can be omitted if the destination table already exists

Which does not hold up in the below example.

#### Code example

Apologies, in advance that the example is a bit long.

It demonstrates Parquet files written to BytesIO buffers from both Polars and Pyarrow unable to be written to a BigQuery table with mode=REQUIRED.

Code example

```python
from io import BytesIO

import pyarrow as pa
import pyarrow.parquet as pq
from google.cloud import bigquery

import polars as pl

PROJECT = ""

def create_and_return_table(table_name: str, client: bigquery.Client) -> bigquery.Table:
schema = [bigquery.SchemaField("foo", "INTEGER", mode="REQUIRED")]
table = bigquery.Table(f"{PROJECT}.testing.{table_name}", schema=schema)

client.delete_table(table, not_found_ok=True)
return client.create_table(table)

def polars_way(table: bigquery.Table, client: bigquery.Client):
df = pl.DataFrame({"foo": [1, 2, 3]})

with BytesIO() as stream:
df.write_parquet(stream)

job_config = bigquery.LoadJobConfig(
source_format=bigquery.SourceFormat.PARQUET,
# Default option, but make it explicit
write_disposition=bigquery.WriteDisposition.WRITE_APPEND,
# The issue does not occur if the schema is explicitly provided.
# schema=table.schema,
)

job = client.load_table_from_file(
stream,
destination=table,
rewind=True,
job_config=job_config,
)

job.result()

def pyarrow_way(table: bigquery.Table, client: bigquery.Client):
# nullable=True is the default, but make it explicit
# This issue does not occur if nullable=False, but the docs imply that shouldn't be
# required
pyarrow_schema = pa.schema([pa.field("foo", pa.int64(), nullable=True)])
pyarrow_table = pa.Table.from_pydict({"foo": [1, 2, 3]}, schema=pyarrow_schema)

with BytesIO() as stream:
writer = pq.ParquetWriter(stream, pyarrow_schema)
writer.write(pyarrow_table)
writer.close()

job_config = bigquery.LoadJobConfig(
source_format=bigquery.SourceFormat.PARQUET,
# Default option, but make it explicit
write_disposition=bigquery.WriteDisposition.WRITE_APPEND,
# The issue does not occur if the schema is explicitly provided.
# schema=table.schema,
)
job = client.load_table_from_file(
stream,
destination=table,
rewind=True,
job_config=job_config,
)

job.result()

def main():
client = bigquery.Client()
table = create_and_return_table("test_pl", client)
polars_way(table, client)
table = create_and_return_table("test_pa", client)
pyarrow_way(table, client)

if __name__ == "__main__":
main()

```

#### Stack trace
Both the `polars_way` and the `pyarrow_way` raise with the error. Here they both are.
```
# polars_way
Traceback (most recent call last):
File "/home/henry/development/polars_bq/combined.py", line 93, in
main()
File "/home/henry/development/polars_bq/combined.py", line 80, in main
polars_way(table, client)
File "/home/henry/development/polars_bq/combined.py", line 41, in polars_way
job.result()
File "/home/henry/development/polars_bq/.venv/lib/python3.11/site-packages/google/cloud/bigquery/job/base.py", line 966, in result
return super(_AsyncJob, self).result(timeout=timeout, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/henry/development/polars_bq/.venv/lib/python3.11/site-packages/google/api_core/future/polling.py", line 261, in result
raise self._exception
google.api_core.exceptions.BadRequest: 400 Provided Schema does not match Table :testing.test_pl.
Field foo has changed mode from REQUIRED to NULLABLE; reason: invalid,
message: Provided Schema does not match Table :testing.test_pl. Field foo has changed mode from REQUIRED to NULLABLE

# pyarrow_way
Traceback (most recent call last):
File "/home/henry/development/polars_bq/combined.py", line 86, in
main()
File "/home/henry/development/polars_bq/combined.py", line 82, in main
pyarrow_way(table, client)
File "/home/henry/development/polars_bq/combined.py", line 74, in pyarrow_way
job.result()
File "/home/henry/development/polars_bq/.venv/lib/python3.11/site-packages/google/cloud/bigquery/job/base.py", line 966, in result
return super(_AsyncJob, self).result(timeout=timeout, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/henry/development/polars_bq/.venv/lib/python3.11/site-packages/google/api_core/future/polling.py", line 261, in result
raise self._exception
google.api_core.exceptions.BadRequest: 400 Provided Schema does not match Table :testing.test_pa.
Field foo has changed mode from REQUIRED to NULLABLE; reason: invalid,
message: Provided Schema does not match Table :testing.test_pa. Field foo has changed mode from REQUIRED to NULLABLE
```