feat(cmd): add unified diff output option for formatting

* Implemented `-d`/`--diff` flag to print unified diffs.
* Added `unifiedDiff` function for generating diffs.
* Config discovery now merges fields from `.pgtidy.yaml`.

docs/config/default.pgtidy.yaml

@@ -0,0 +1,26 @@
+# PgTidy house style — all fields shown with their default values.
+# Place as .pgtidy.yaml in your project root (or any parent directory).
+# Any field you omit keeps its default.
+
+# Indentation string for one level (two spaces).
+indent: "  "
+
+# Line terminator written by the formatter.
+newline: "\n"
+
+# Casing for SQL keywords (SELECT, FROM, WHERE, …).
+# upper | lower | preserve
+keyword_case: upper
+
+# Casing for unquoted identifiers (column names, variable names, …).
+# upper | lower | preserve
+ident_case: lower
+
+# Casing for built-in type names (text, integer, boolean, …).
+# upper | lower | preserve
+type_case: lower
+
+# Comma placement in multi-line parameter / column lists.
+# leading  → comma at the start of the continuation line  (,col)
+# trailing → comma at the end of the preceding line       (col,)
+commas: leading

docs/config/lowercase.pgtidy.yaml

@@ -0,0 +1,7 @@
+# Fully lowercase style — keywords, types, and identifiers all lowercased.
+# Common in teams that prefer minimal visual noise.
+
+keyword_case: lower
+ident_case: lower
+type_case: lower
+commas: leading

docs/config/preserve-case.pgtidy.yaml

@@ -0,0 +1,8 @@
+# Preserve-case style — no casing changes applied.
+# Useful for codebases with mixed-convention legacy SQL that you want to
+# reformat structurally (indentation, commas) without touching casing.
+
+keyword_case: preserve
+ident_case: preserve
+type_case: preserve
+commas: leading

docs/config/trailing-commas.pgtidy.yaml

@@ -0,0 +1,7 @@
+# Trailing-comma style — comma at the end of each line rather than the start.
+# Matches the SQL style common in tools like dbt and some BI platforms.
+
+keyword_case: upper
+ident_case: lower
+type_case: lower
+commas: trailing