ROB: Ignore EOF in NumberObject.read_from_stream

Use `ignore_eof=True` just like `NameObject` does, which is the only other
caller to `read_until_regex` in this module.

This helps prevent arcane exceptions when trying to parse a number:

```
PyPDF2.errors.PdfStreamError: Stream has ended unexpectedly
```

The motivation is essentially identical to the change that introduced
`ignore_eof=True` on `NameObject`s:
431ba70

PyPDF2/generic/_base.py

@@ -402,7 +402,7 @@ def writeToStream(
 
  @staticmethod
  def read_from_stream(stream: StreamType) -> Union["NumberObject", "FloatObject"]:
- num = read_until_regex(stream, NumberObject.NumberPattern)
+ num = read_until_regex(stream, NumberObject.NumberPattern, ignore_eof=True)
  if num.find(b".") != -1:
  return FloatObject(num)
  return NumberObject(num)