.gitattributes

# ============================================================
# .gitattributes (Keep files consistent across operating systems)
# ============================================================

# REQ.UNIVERSAL: All professional GitHub project repositories MUST include .gitattributes.
# WHY: Ensure consistent line endings, diff behavior, and file classification
# across Windows, macOS, and Linux environments.
# ALT: Repository may omit .gitattributes ONLY if equivalent normalization is
# enforced reliably by tooling and CI (rare and fragile).
# CUSTOM: Update file-type rules only when introducing new languages,
# binary artifacts, or documentation formats.
# NOTE: Rules are ordered by impact and generality, not alphabetically.
# Git attributes are documented at https://git-scm.com/docs/gitattributes

# === Core defaults (always apply) ===

# WHY: Auto-detect text files and normalize line endings to avoid cross-platform drift.
* text=auto


# === Programming languages and scripts ===

# WHY: Python and shell scripts must use LF for CI/CD, Linux environments, and containers.
*.py    text eol=lf
*.sh    text eol=lf

# WHY: PowerShell convention on Windows uses CRLF.
*.ps1   text eol=crlf


# === Markup and documentation ===

# WHY: Documentation and markup files use LF; standard for cross-platform tooling.
*.md    text eol=lf


# === Configuration and structured text ===

# WHY: Configuration and structured text formats use LF for stable diffs.
*.json    text eol=lf
*.jsonc   text eol=lf
*.jsonl   text eol=lf
*.toml    text eol=lf
*.yaml    text eol=lf
*.yml     text eol=lf


# === Databases (binary) ===

# WHY: Database files are binary; prevent text normalization and meaningless diffs.
*.db       binary
*.sqlite   binary
*.sqlite3  binary


# === Compressed artifacts (binary) ===

# WHY: Compressed artifacts are binary; prevent text normalization.
*.gz    binary
*.tar   binary
*.tgz   binary
*.zip   binary
*.7z    binary


# === Log files (binary) ===

# WHY: Log files are typically large; prevent text normalization.
*.log         binary
project.log   binary


# === GitHub metadata and UI ===

# WHY: Exclude documentation and tests from GitHub language statistics.
docs/**         linguist-documentation
tests/**        linguist-documentation
evaluation/**   linguist-documentation