bug/fallback-encoding-detection-for-file-like-objects

[ujjwalredd]

Describe the bug
Currently, when trying to determine a file type from its initial text (text_head), the parser attempts fallback character-set detection only for physical file paths. For file-like objects (e.g., SpooledTemporaryFile, BytesIO streams), it blindly decodes the first 4096 bytes using the default encoding (usually UTF-8) with errors="ignore". This can strip characters and lead to corrupted text or incorrect filetype classification when ingesting documents from cloud storage (S3/GCS) or APIs.

To Reproduce
Pass a non-UTF-8 encoded BytesIO object to detect_filetype.

Expected behavior
File-like objects should use the same detect_file_encoding fallback mechanism as physical file paths to avoid data loss on decoding errors.

Screenshots
N/A

Environment Info
N/A - General bug in filetype.py

Additional context
Resolves the TODO in filetype.py:text_head.

[feiiiiii5]

Opened PR #4438: text_head() now applies the same detect_file_encoding() fallback to file-like objects as it does to file paths, replacing the errors="ignore" decode that silently stripped characters. Two tests cover it (wrong-encoding-arg correction + latin-1 regression).