SegyFactory.create_textual_header() raises UnicodeEncodeError

[ajmalrasi]

Issue

SegyFactory.create_textual_header() throws a UnicodeDecodeError on some EBCDIC characters, even though SegyFile.text_header can already read and print the same header. The two code paths should behave the same way.

---

Steps to Reproduce

from mdio.segy import SegyFile, SegyFactory, get_segy_standard

segy_path = "gs://tgs-geophysical-test-samples/AP_Data/900_General/WeirdCharinEBCDIC.sgy"

# 1. Read the file
sreader = SegyFile(segy_path)          # Works
print(sreader.text_header)             # Works (shows weired chars)

factory = SegyFactory(
    spec=get_segy_standard(1.0),
    samples_per_trace=2001,
    sample_interval=4_000,
)

txt = factory.create_textual_header(sreader.text_header)  # Fails

Error:
UnicodeEncodeError: 'ascii' codec can't encode character '\ufffd' in position 475: ordinal not in range(128)

[amitpendharkar]

Here is the code that only uses functions from segy library

from segy import SegyFile
from segy.factory import SegyFactory
from segy.standards import get_segy_standard


segy_path = "gs://tgs-geophysical-test-samples/AP_Data/900_General/WeirdCharinEBCDIC.sgy"

# 1. Read the file
sreader = SegyFile(segy_path)           # Works
header = sreader.text_header
print(header)                           # Works (shows weird chars)

factory = SegyFactory(
    spec=get_segy_standard(1.0),
    samples_per_trace=2001,
    sample_interval=4_000,
)

txt = factory.create_textual_header(sreader.text_header)  # Fails
print(txt)

The problem is while reading, it uses 'ebcdic' encoding which accepts these weird characters. These weird characters can be seen on C 2 and C 6 when printed.
And while writing, in create_textual_header this hardcoded line treats it as ascii and it fails.