Port pydocstyle code 401 (ImperativeMood) (#1999)

This adds support for pydocstyle code D401 using the `imperative` crate.

Cargo.lock

@@ -926,6 +926,16 @@ dependencies = [
  "winapi-util",
 ]
 
+[[package]]
+name = "imperative"
+version = "1.0.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9f92123bf2fe0d9f1b5df1964727b970ca3b2d0203d47cf97fb1f36d856b6398"
+dependencies = [
+ "phf 0.11.1",
+ "rust-stemmers",
+]
+
 [[package]]
 name = "indexmap"
 version = "1.9.2"
@@ -1511,6 +1521,15 @@ dependencies = [
  "phf_shared 0.10.0",
 ]
 
+[[package]]
+name = "phf"
+version = "0.11.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "928c6535de93548188ef63bb7c4036bd415cd8f36ad25af44b9789b2ee72a48c"
+dependencies = [
+ "phf_shared 0.11.1",
+]
+
 [[package]]
 name = "phf_codegen"
 version = "0.8.0"
@@ -1569,6 +1588,15 @@ dependencies = [
  "siphasher",
 ]
 
+[[package]]
+name = "phf_shared"
+version = "0.11.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e1fb5f6f826b772a8d4c0394209441e7d37cbbb967ae9c7e0e8134365c9ee676"
+dependencies = [
+ "siphasher",
+]
+
 [[package]]
 name = "pico-args"
 version = "0.4.2"
@@ -1923,6 +1951,7 @@ dependencies = [
  "glob",
  "globset",
  "ignore",
+ "imperative",
  "insta",
  "itertools",
  "js-sys",
@@ -2027,6 +2056,16 @@ dependencies = [
  "textwrap",
 ]
 
+[[package]]
+name = "rust-stemmers"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e46a2036019fdb888131db7a4c847a1063a7493f971ed94ea82c67eada63ca54"
+dependencies = [
+ "serde",
+ "serde_derive",
+]
+
 [[package]]
 name = "rustc-hash"
 version = "1.1.0"

Cargo.toml

@@ -35,6 +35,7 @@ fern = { version = "0.6.1" }
 glob = { version = "0.3.0" }
 globset = { version = "0.4.9" }
 ignore = { version = "0.4.18" }
+imperative = { version = "1.0.3" }
 itertools = { version = "0.10.5" }
 libcst = { git = "https://github.com/charliermarsh/LibCST", rev = "f2f0b7a487a8725d161fe8b3ed73a6758b21e177" }
 log = { version = "0.4.17" }
@@ -59,9 +60,9 @@ smallvec = { version = "1.10.0" }
 strum = { version = "0.24.1", features = ["strum_macros"] }
 strum_macros = { version = "0.24.3" }
 textwrap = { version = "0.16.0" }
+thiserror = { version = "1.0" }
 titlecase = { version = "2.2.1" }
 toml_edit = { version = "0.17.1", features = ["easy"] }
-thiserror = { version = "1.0" }
 
 # https://docs.rs/getrandom/0.2.7/getrandom/#webassembly-support
 # For (future) wasm-pack support

README.md

@@ -672,6 +672,7 @@ For more, see [pydocstyle](https://pypi.org/project/pydocstyle/6.1.1/) on PyPI.
 | D300 | uses-triple-quotes | Use """triple double quotes""" |  |
 | D301 | uses-r-prefix-for-backslashed-content | Use r""" if any backslashes in a docstring |  |
 | D400 | ends-in-period | First line should end with a period | 🛠 |
+| D401 | non-imperative-mood | First line of docstring should be in imperative mood: "{first_line}" |  |
 | D402 | no-signature | First line should not be the function's signature |  |
 | D403 | first-line-capitalized | First word of the first line should be properly capitalized |  |
 | D404 | no-this-prefix | First word of the docstring should not be "This" |  |

resources/test/fixtures/pydocstyle/D401.py

@@ -0,0 +1,43 @@
+# Bad examples
+
+def bad_liouiwnlkjl():
+    """Returns foo."""
+
+
+def bad_sdgfsdg23245():
+    """Constructor for a foo."""
+
+
+def bad_sdgfsdg23245777():
+    """
+
+    Constructor for a boa.
+
+    """
+
+
+def bad_run_something():
+    """Runs something"""
+    pass
+
+
+def multi_line():
+    """Writes a logical line that
+    extends to two physical lines.
+    """
+
+
+# Good examples
+
+def good_run_something():
+    """Run away."""
+
+
+def good_construct():
+    """Construct a beautiful house."""
+
+
+def good_multi_line():
+    """Write a logical line that
+    extends to two physical lines.
+    """

ruff.schema.json

@@ -1277,6 +1277,7 @@
         "D4",
         "D40",
         "D400",
+        "D401",
         "D402",
         "D403",
         "D404",

src/checkers/ast.rs

@@ -4497,6 +4497,7 @@ impl<'a> Checker<'a> {
                 .rules
                 .enabled(&Rule::UsesRPrefixForBackslashedContent)
             || self.settings.rules.enabled(&Rule::EndsInPeriod)
+            || self.settings.rules.enabled(&Rule::NonImperativeMood)
             || self.settings.rules.enabled(&Rule::NoSignature)
             || self.settings.rules.enabled(&Rule::FirstLineCapitalized)
             || self.settings.rules.enabled(&Rule::NoThisPrefix)
@@ -4645,6 +4646,9 @@ impl<'a> Checker<'a> {
                 if self.settings.rules.enabled(&Rule::EndsInPeriod) {
                     pydocstyle::rules::ends_with_period(self, &docstring);
                 }
+                if self.settings.rules.enabled(&Rule::NonImperativeMood) {
+                    pydocstyle::rules::non_imperative_mood::non_imperative_mood(self, &docstring);
+                }
                 if self.settings.rules.enabled(&Rule::NoSignature) {
                     pydocstyle::rules::no_signature(self, &docstring);
                 }

src/registry.rs

@@ -281,6 +281,7 @@ ruff_macros::define_rule_mapping!(
     D300 => violations::UsesTripleQuotes,
     D301 => violations::UsesRPrefixForBackslashedContent,
     D400 => violations::EndsInPeriod,
+    D401 => crate::rules::pydocstyle::rules::non_imperative_mood::NonImperativeMood,
     D402 => violations::NoSignature,
     D403 => violations::FirstLineCapitalized,
     D404 => violations::NoThisPrefix,

src/rules/pydocstyle/helpers.rs

@@ -55,3 +55,11 @@ pub fn logical_line(content: &str) -> Option<usize> {
     }
     logical_line
 }
+
+/// Normalize a word by removing all non-alphanumeric characters
+/// and converting it to lowercase.
+pub fn normalize_word(first_word: &str) -> String {
+    first_word
+        .replace(|c: char| !c.is_alphanumeric(), "")
+        .to_lowercase()
+}

src/rules/pydocstyle/mod.rs

@@ -44,6 +44,7 @@ mod tests {
     #[test_case(Rule::UsesRPrefixForBackslashedContent, Path::new("D.py"); "D301")]
     #[test_case(Rule::EndsInPeriod, Path::new("D.py"); "D400_0")]
     #[test_case(Rule::EndsInPeriod, Path::new("D400.py"); "D400_1")]
+    #[test_case(Rule::NonImperativeMood, Path::new("D401.py"); "D401")]
     #[test_case(Rule::NoSignature, Path::new("D.py"); "D402")]
     #[test_case(Rule::FirstLineCapitalized, Path::new("D.py"); "D403")]
     #[test_case(Rule::NoThisPrefix, Path::new("D.py"); "D404")]

src/rules/pydocstyle/rules/mod.rs

@@ -31,6 +31,7 @@ mod multi_line_summary_start;
 mod newline_after_last_paragraph;
 mod no_signature;
 mod no_surrounding_whitespace;
+pub mod non_imperative_mood;
 mod not_empty;
 mod not_missing;
 mod one_liner;

src/rules/pydocstyle/rules/non_imperative_mood.rs

@@ -0,0 +1,50 @@
+use imperative::Mood;
+use once_cell::sync::Lazy;
+use ruff_macros::derive_message_formats;
+
+use crate::ast::types::Range;
+use crate::checkers::ast::Checker;
+use crate::define_violation;
+use crate::docstrings::definition::Docstring;
+use crate::registry::Diagnostic;
+use crate::rules::pydocstyle::helpers::normalize_word;
+use crate::violation::Violation;
+
+static MOOD: Lazy<Mood> = Lazy::new(Mood::new);
+
+/// D401
+pub fn non_imperative_mood(checker: &mut Checker, docstring: &Docstring) {
+    let body = docstring.body;
+
+    // Find first line, disregarding whitespace.
+    let line = match body.trim().lines().next() {
+        Some(line) => line.trim(),
+        None => return,
+    };
+    // Find the first word on that line and normalize it to lower-case.
+    let first_word_norm = match line.split_whitespace().next() {
+        Some(word) => normalize_word(word),
+        None => return,
+    };
+    if first_word_norm.is_empty() {
+        return;
+    }
+    if let Some(false) = MOOD.is_imperative(&first_word_norm) {
+        let diagnostic = Diagnostic::new(
+            NonImperativeMood(line.to_string()),
+            Range::from_located(docstring.expr),
+        );
+        checker.diagnostics.push(diagnostic);
+    }
+}
+
+define_violation!(
+    pub struct NonImperativeMood(pub String);
+);
+impl Violation for NonImperativeMood {
+    #[derive_message_formats]
+    fn message(&self) -> String {
+        let NonImperativeMood(first_line) = self;
+        format!("First line of docstring should be in imperative mood: \"{first_line}\"")
+    }
+}

src/rules/pydocstyle/rules/starts_with_this.rs

@@ -2,6 +2,7 @@ use crate::ast::types::Range;
 use crate::checkers::ast::Checker;
 use crate::docstrings::definition::Docstring;
 use crate::registry::Diagnostic;
+use crate::rules::pydocstyle::helpers::normalize_word;
 use crate::violations;
 
 /// D404
@@ -16,11 +17,7 @@ pub fn starts_with_this(checker: &mut Checker, docstring: &Docstring) {
     let Some(first_word) = body.split(' ').next() else {
         return
     };
-    if first_word
+    if normalize_word(first_word) != "this" {
-        .replace(|c: char| !c.is_alphanumeric(), "")
-        .to_lowercase()
-        != "this"
-    {
         return;
     }
     checker.diagnostics.push(Diagnostic::new(

src/rules/pydocstyle/snapshots/ruff__rules__pydocstyle__tests__D401_D401.py.snap

@@ -0,0 +1,55 @@
+---
+source: src/rules/pydocstyle/mod.rs
+expression: diagnostics
+---
+- kind:
+    NonImperativeMood: Returns foo.
+  location:
+    row: 4
+    column: 4
+  end_location:
+    row: 4
+    column: 22
+  fix: ~
+  parent: ~
+- kind:
+    NonImperativeMood: Constructor for a foo.
+  location:
+    row: 8
+    column: 4
+  end_location:
+    row: 8
+    column: 32
+  fix: ~
+  parent: ~
+- kind:
+    NonImperativeMood: Constructor for a boa.
+  location:
+    row: 12
+    column: 4
+  end_location:
+    row: 16
+    column: 7
+  fix: ~
+  parent: ~
+- kind:
+    NonImperativeMood: Runs something
+  location:
+    row: 20
+    column: 4
+  end_location:
+    row: 20
+    column: 24
+  fix: ~
+  parent: ~
+- kind:
+    NonImperativeMood: Writes a logical line that
+  location:
+    row: 25
+    column: 4
+  end_location:
+    row: 27
+    column: 7
+  fix: ~
+  parent: ~
+