[pydocstyle] Skip leading whitespace for D403 (#14963)

This PR introduces three changes to `D403`, which has to do with
capitalizing the first word in a docstring.

1. The diagnostic and fix now skip leading whitespace when determining
what counts as "the first word".
2. The name has been changed to `first-word-uncapitalized` from
`first-line-capitalized`, for both clarity and compliance with our rule
naming policy.
3. The diagnostic message and documentation has been modified slightly
to reflect this.

Closes #14890

crates/ruff_linter/resources/test/fixtures/pydocstyle/D403.py

@@ -31,3 +31,13 @@ def single_word():
 
 def single_word_no_dot():
     """singleword"""
+
+def first_word_lots_of_whitespace():
+    """
+
+
+
+    here is the start of my docstring!
+
+    What do you think?    
+    """

crates/ruff_linter/src/checkers/ast/analyze/definitions.rs

@@ -46,7 +46,7 @@ pub(crate) fn definitions(checker: &mut Checker) {
         Rule::EndsInPeriod,
         Rule::EndsInPunctuation,
         Rule::EscapeSequenceInDocstring,
-        Rule::FirstLineCapitalized,
+        Rule::FirstWordUncapitalized,
         Rule::FitsOnOneLine,
         Rule::IndentWithSpaces,
         Rule::MultiLineSummaryFirstLine,
@@ -277,7 +277,7 @@ pub(crate) fn definitions(checker: &mut Checker) {
             if checker.enabled(Rule::NoSignature) {
                 pydocstyle::rules::no_signature(checker, &docstring);
             }
-            if checker.enabled(Rule::FirstLineCapitalized) {
+            if checker.enabled(Rule::FirstWordUncapitalized) {
                 pydocstyle::rules::capitalized(checker, &docstring);
             }
             if checker.enabled(Rule::DocstringStartsWithThis) {

crates/ruff_linter/src/codes.rs

@@ -566,7 +566,7 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<(RuleGroup, Rule)> {
         (Pydocstyle, "400") => (RuleGroup::Stable, rules::pydocstyle::rules::EndsInPeriod),
         (Pydocstyle, "401") => (RuleGroup::Stable, rules::pydocstyle::rules::NonImperativeMood),
         (Pydocstyle, "402") => (RuleGroup::Stable, rules::pydocstyle::rules::NoSignature),
-        (Pydocstyle, "403") => (RuleGroup::Stable, rules::pydocstyle::rules::FirstLineCapitalized),
+        (Pydocstyle, "403") => (RuleGroup::Stable, rules::pydocstyle::rules::FirstWordUncapitalized),
         (Pydocstyle, "404") => (RuleGroup::Stable, rules::pydocstyle::rules::DocstringStartsWithThis),
         (Pydocstyle, "405") => (RuleGroup::Stable, rules::pydocstyle::rules::CapitalizeSectionName),
         (Pydocstyle, "406") => (RuleGroup::Stable, rules::pydocstyle::rules::NewLineAfterSectionName),

crates/ruff_linter/src/rules/pydocstyle/mod.rs

@@ -32,8 +32,8 @@ mod tests {
     #[test_case(Rule::EndsInPeriod, Path::new("D400_415.py"))]
     #[test_case(Rule::EndsInPunctuation, Path::new("D.py"))]
     #[test_case(Rule::EndsInPunctuation, Path::new("D400_415.py"))]
-    #[test_case(Rule::FirstLineCapitalized, Path::new("D.py"))]
+    #[test_case(Rule::FirstWordUncapitalized, Path::new("D.py"))]
-    #[test_case(Rule::FirstLineCapitalized, Path::new("D403.py"))]
+    #[test_case(Rule::FirstWordUncapitalized, Path::new("D403.py"))]
     #[test_case(Rule::FitsOnOneLine, Path::new("D.py"))]
     #[test_case(Rule::IndentWithSpaces, Path::new("D.py"))]
     #[test_case(Rule::UndocumentedMagicMethod, Path::new("D.py"))]

crates/ruff_linter/src/rules/pydocstyle/rules/capitalized.rs

@@ -10,8 +10,8 @@ use crate::docstrings::Docstring;
 /// Checks for docstrings that do not start with a capital letter.
 ///
 /// ## Why is this bad?
-/// The first character in a docstring should be capitalized for, grammatical
+/// The first non-whitespace character in a docstring should be
-/// correctness and consistency.
+/// capitalized for grammatical correctness and consistency.
 ///
 /// ## Example
 /// ```python
@@ -30,16 +30,16 @@ use crate::docstrings::Docstring;
 /// - [NumPy Style Guide](https://numpydoc.readthedocs.io/en/latest/format.html)
 /// - [Google Python Style Guide - Docstrings](https://google.github.io/styleguide/pyguide.html#38-comments-and-docstrings)
 #[derive(ViolationMetadata)]
-pub(crate) struct FirstLineCapitalized {
+pub(crate) struct FirstWordUncapitalized {
     first_word: String,
     capitalized_word: String,
 }
 
-impl AlwaysFixableViolation for FirstLineCapitalized {
+impl AlwaysFixableViolation for FirstWordUncapitalized {
     #[derive_message_formats]
     fn message(&self) -> String {
         format!(
-            "First word of the first line should be capitalized: `{}` -> `{}`",
+            "First word of the docstring should be capitalized: `{}` -> `{}`",
             self.first_word, self.capitalized_word
         )
     }
@@ -59,7 +59,8 @@ pub(crate) fn capitalized(checker: &mut Checker, docstring: &Docstring) {
     }
 
     let body = docstring.body();
-    let first_word = body.split_once(' ').map_or_else(
+    let trim_start_body = body.trim_start();
+    let first_word = trim_start_body.split_once(' ').map_or_else(
         || {
             // If the docstring is a single word, trim the punctuation marks because
             // it makes the ASCII test below fail.
@@ -91,8 +92,10 @@ pub(crate) fn capitalized(checker: &mut Checker, docstring: &Docstring) {
 
     let capitalized_word = uppercase_first_char.to_string() + first_word_chars.as_str();
 
+    let leading_whitespace_len = body.text_len() - trim_start_body.text_len();
+
     let mut diagnostic = Diagnostic::new(
-        FirstLineCapitalized {
+        FirstWordUncapitalized {
             first_word: first_word.to_string(),
             capitalized_word: capitalized_word.to_string(),
         },
@@ -101,7 +104,7 @@ pub(crate) fn capitalized(checker: &mut Checker, docstring: &Docstring) {
 
     diagnostic.set_fix(Fix::safe_edit(Edit::range_replacement(
         capitalized_word,
-        TextRange::at(body.start(), first_word.text_len()),
+        TextRange::at(body.start() + leading_whitespace_len, first_word.text_len()),
     )));
 
     checker.diagnostics.push(diagnostic);

crates/ruff_linter/src/rules/pydocstyle/snapshots/ruff_linter__rules__pydocstyle__tests__D403_D403.py.snap

@@ -1,8 +1,7 @@
 ---
 source: crates/ruff_linter/src/rules/pydocstyle/mod.rs
-snapshot_kind: text
 ---
-D403.py:2:5: D403 [*] First word of the first line should be capitalized: `this` -> `This`
+D403.py:2:5: D403 [*] First word of the docstring should be capitalized: `this` -> `This`
   |
 1 | def bad_function():
 2 |     """this docstring is not capitalized"""
@@ -20,7 +19,7 @@ D403.py:2:5: D403 [*] First word of the first line should be capitalized: `this`
 4 4 | def good_function():
 5 5 |     """This docstring is capitalized."""
 
-D403.py:30:5: D403 [*] First word of the first line should be capitalized: `singleword` -> `Singleword`
+D403.py:30:5: D403 [*] First word of the docstring should be capitalized: `singleword` -> `Singleword`
    |
 29 | def single_word():
 30 |     """singleword."""
@@ -40,11 +39,13 @@ D403.py:30:5: D403 [*] First word of the first line should be capitalized: `sing
 32 32 | def single_word_no_dot():
 33 33 |     """singleword"""
 
-D403.py:33:5: D403 [*] First word of the first line should be capitalized: `singleword` -> `Singleword`
+D403.py:33:5: D403 [*] First word of the docstring should be capitalized: `singleword` -> `Singleword`
    |
 32 | def single_word_no_dot():
 33 |     """singleword"""
    |     ^^^^^^^^^^^^^^^^ D403
+34 | 
+35 | def first_word_lots_of_whitespace():
    |
    = help: Capitalize `singleword` to `Singleword`
 
@@ -54,3 +55,32 @@ D403.py:33:5: D403 [*] First word of the first line should be capitalized: `sing
 32 32 | def single_word_no_dot():
 33    |-    """singleword"""
    33 |+    """Singleword"""
+34 34 | 
+35 35 | def first_word_lots_of_whitespace():
+36 36 |     """
+
+D403.py:36:5: D403 [*] First word of the docstring should be capitalized: `here` -> `Here`
+   |
+35 |   def first_word_lots_of_whitespace():
+36 |       """
+   |  _____^
+37 | | 
+38 | | 
+39 | | 
+40 | |     here is the start of my docstring!
+41 | | 
+42 | |     What do you think?    
+43 | |     """
+   | |_______^ D403
+   |
+   = help: Capitalize `here` to `Here`
+
+ℹ Safe fix
+37 37 | 
+38 38 | 
+39 39 | 
+40    |-    here is the start of my docstring!
+   40 |+    Here is the start of my docstring!
+41 41 | 
+42 42 |     What do you think?    
+43 43 |     """