fix: prevent duplicate CSV header on append

[Sidu9]

Description

Fixes an issue where the CSV header was duplicated every time results were appended to an existing output file.

Previously, the CSV formatter always included the header, causing repeated headers when the CLI was run multiple times against the same output file.

Changes

• Ensure the CSV header is written only when the output file is new or empty.
• Keep CSV formatting focused on generating result rows.
• Preserve the existing append behavior for subsequent runs.

Testing

• Verified that the header is written on the first run.
• Verified that subsequent runs append results without duplicating the header.
• Ran the existing test suite successfully.

Closes #648

[kaifcodec]

Thanks for working on this @Sidu9!

The duplicate header issue when appending is definitely real. However, removing CSV_HEADER from into_csv() entirely breaks library usage for callers importing into_csv() from Python (as documented in docs/USAGE.md), who expect a valid CSV with headers.

Could we update this to:

1. Make include_header: bool = True an optional parameter in into_csv(results, include_header=True).
2. In __main__.py, determine has_content first and call formatter.into_csv(t_results, include_header=not has_content).
3. Add a test in tests/test_formatter.py verifying both include_header=True and include_header=False.

[A-S-Manoj]

LGTM. Tested locally, ran the append logic across multiple writes to the same file, and confirmed the header appears only once with no duplication. Output parses cleanly with csv.DictReader.

A regression test asserting the header shows up only once across repeated writes would be a nice add-on, but it's not blocking.

Approving.

[A-S-Manoj]

@kaifcodec Ah, missed your comment above, after checking, I agree with you. My approval was based on testing the CLI append path, which does work, but you're right that hardcoding into_csv() to drop the header breaks the documented library usage. +1 on the include_header param approach.

[kaifcodec]

Yeah no worries @A-S-Manoj, it happens.
I know the core better so I caught that up.
However thank you for opening the issue.

[Sidu9]

Addressed your review @kaifcodec and added regression tests covering all the scenarios @A-S-Manoj. Let me know if this is Okay. Do we also need to make any changes to the docs to incorporate this new param?

[A-S-Manoj]

Nice work @Sidu9, this addresses the concern.
Could you also update USAGE.md to mention the new include_header param?

[Sidu9]

I've updated the documentation @A-S-Manoj. Let me know if its ok