Implement copyright notice detection (#4701)

## Summary

Add copyright notice detection to enforce the presence of copyright
headers in Python files.

Configurable settings include: the relevant regular expression, the
author name, and the minimum file size, similar to
[flake8-copyright](https://github.com/savoirfairelinux/flake8-copyright).

Closes #3579

---------

Signed-off-by: ryan <ryang@waabi.ai>
Co-authored-by: Charlie Marsh <charlie.r.marsh@gmail.com>

README.md

@@ -261,6 +261,7 @@ quality tools, including:
 - [flake8-builtins](https://pypi.org/project/flake8-builtins/)
 - [flake8-commas](https://pypi.org/project/flake8-commas/)
 - [flake8-comprehensions](https://pypi.org/project/flake8-comprehensions/)
+- [flake8-copyright](https://pypi.org/project/flake8-copyright/)
 - [flake8-datetimez](https://pypi.org/project/flake8-datetimez/)
 - [flake8-debugger](https://pypi.org/project/flake8-debugger/)
 - [flake8-django](https://pypi.org/project/flake8-django/)

crates/ruff/src/checkers/physical_lines.rs

@@ -8,6 +8,7 @@ use ruff_python_ast::source_code::{Indexer, Locator, Stylist};
 use ruff_python_whitespace::UniversalNewlines;
 
 use crate::registry::Rule;
+use crate::rules::copyright::rules::missing_copyright_notice;
 use crate::rules::flake8_executable::helpers::{extract_shebang, ShebangDirective};
 use crate::rules::flake8_executable::rules::{
     shebang_missing, shebang_newline, shebang_not_executable, shebang_python, shebang_whitespace,
@@ -49,6 +50,7 @@ pub(crate) fn check_physical_lines(
     let enforce_blank_line_contains_whitespace =
         settings.rules.enabled(Rule::BlankLineWithWhitespace);
     let enforce_tab_indentation = settings.rules.enabled(Rule::TabIndentation);
+    let enforce_copyright_notice = settings.rules.enabled(Rule::MissingCopyrightNotice);
 
     let fix_unnecessary_coding_comment = settings.rules.should_fix(Rule::UTF8EncodingDeclaration);
     let fix_shebang_whitespace = settings.rules.should_fix(Rule::ShebangLeadingWhitespace);
@@ -172,6 +174,12 @@ pub(crate) fn check_physical_lines(
         }
     }
 
+    if enforce_copyright_notice {
+        if let Some(diagnostic) = missing_copyright_notice(locator, settings) {
+            diagnostics.push(diagnostic);
+        }
+    }
+
     diagnostics
 }
 

crates/ruff/src/codes.rs

@@ -374,6 +374,9 @@ pub fn code_to_rule(linter: Linter, code: &str) -> Option<(RuleGroup, Rule)> {
         (Flake8Simplify, "401") => (RuleGroup::Unspecified, rules::flake8_simplify::rules::IfElseBlockInsteadOfDictGet),
         (Flake8Simplify, "910") => (RuleGroup::Unspecified, rules::flake8_simplify::rules::DictGetWithNoneDefault),
 
+        // copyright
+        (Copyright, "001") => (RuleGroup::Unspecified, rules::copyright::rules::MissingCopyrightNotice),
+
         // pyupgrade
         (Pyupgrade, "001") => (RuleGroup::Unspecified, rules::pyupgrade::rules::UselessMetaclassType),
         (Pyupgrade, "003") => (RuleGroup::Unspecified, rules::pyupgrade::rules::TypeOfPrimitive),

crates/ruff/src/registry.rs

@@ -80,6 +80,9 @@ pub enum Linter {
     /// [flake8-commas](https://pypi.org/project/flake8-commas/)
     #[prefix = "COM"]
     Flake8Commas,
+    /// Copyright-related rules
+    #[prefix = "CPY"]
+    Copyright,
     /// [flake8-comprehensions](https://pypi.org/project/flake8-comprehensions/)
     #[prefix = "C4"]
     Flake8Comprehensions,
@@ -267,6 +270,7 @@ impl Rule {
             | Rule::ShebangLeadingWhitespace
             | Rule::TrailingWhitespace
             | Rule::TabIndentation
+            | Rule::MissingCopyrightNotice
             | Rule::BlankLineWithWhitespace => LintSource::PhysicalLines,
             Rule::AmbiguousUnicodeCharacterComment
             | Rule::AmbiguousUnicodeCharacterDocstring

crates/ruff/src/rules/copyright/mod.rs

@@ -0,0 +1,158 @@
+//! Rules related to copyright notices.
+pub(crate) mod rules;
+
+pub mod settings;
+
+#[cfg(test)]
+mod tests {
+    use crate::registry::Rule;
+    use crate::test::test_snippet;
+    use crate::{assert_messages, settings};
+
+    #[test]
+    fn notice() {
+        let diagnostics = test_snippet(
+            r#"
+# Copyright 2023
+
+import os
+"#
+            .trim(),
+            &settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice]),
+        );
+        assert_messages!(diagnostics);
+    }
+
+    #[test]
+    fn notice_with_c() {
+        let diagnostics = test_snippet(
+            r#"
+# Copyright (C) 2023
+
+import os
+"#
+            .trim(),
+            &settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice]),
+        );
+        assert_messages!(diagnostics);
+    }
+
+    #[test]
+    fn notice_with_caps() {
+        let diagnostics = test_snippet(
+            r#"
+# COPYRIGHT (C) 2023
+
+import os
+"#
+            .trim(),
+            &settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice]),
+        );
+        assert_messages!(diagnostics);
+    }
+
+    #[test]
+    fn notice_with_range() {
+        let diagnostics = test_snippet(
+            r#"
+# Copyright (C) 2021-2023
+
+import os
+"#
+            .trim(),
+            &settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice]),
+        );
+        assert_messages!(diagnostics);
+    }
+
+    #[test]
+    fn valid_author() {
+        let diagnostics = test_snippet(
+            r#"
+# Copyright (C) 2023 Ruff
+
+import os
+"#
+            .trim(),
+            &settings::Settings {
+                copyright: super::settings::Settings {
+                    author: Some("Ruff".to_string()),
+                    ..super::settings::Settings::default()
+                },
+                ..settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice])
+            },
+        );
+        assert_messages!(diagnostics);
+    }
+
+    #[test]
+    fn invalid_author() {
+        let diagnostics = test_snippet(
+            r#"
+# Copyright (C) 2023 Some Author
+
+import os
+"#
+            .trim(),
+            &settings::Settings {
+                copyright: super::settings::Settings {
+                    author: Some("Ruff".to_string()),
+                    ..super::settings::Settings::default()
+                },
+                ..settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice])
+            },
+        );
+        assert_messages!(diagnostics);
+    }
+
+    #[test]
+    fn small_file() {
+        let diagnostics = test_snippet(
+            r#"
+import os
+"#
+            .trim(),
+            &settings::Settings {
+                copyright: super::settings::Settings {
+                    min_file_size: 256,
+                    ..super::settings::Settings::default()
+                },
+                ..settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice])
+            },
+        );
+        assert_messages!(diagnostics);
+    }
+
+    #[test]
+    fn late_notice() {
+        let diagnostics = test_snippet(
+            r#"
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+# Content Content Content Content Content Content Content Content Content Content
+
+# Copyright 2023
+"#
+            .trim(),
+            &settings::Settings::for_rules(vec![Rule::MissingCopyrightNotice]),
+        );
+        assert_messages!(diagnostics);
+    }
+}

crates/ruff/src/rules/copyright/rules/missing_copyright_notice.rs

@@ -0,0 +1,59 @@
+use ruff_text_size::{TextRange, TextSize};
+
+use ruff_diagnostics::{Diagnostic, Violation};
+use ruff_macros::{derive_message_formats, violation};
+use ruff_python_ast::source_code::Locator;
+
+use crate::settings::Settings;
+
+/// ## What it does
+/// Checks for the absence of copyright notices within Python files.
+///
+/// ## Why is this bad?
+/// In some codebases, it's common to have a license header at the top of every
+/// file. This rule ensures that the license header is present.
+#[violation]
+pub struct MissingCopyrightNotice;
+
+impl Violation for MissingCopyrightNotice {
+    #[derive_message_formats]
+    fn message(&self) -> String {
+        format!("Missing copyright notice at top of file")
+    }
+}
+
+/// CPY001
+pub(crate) fn missing_copyright_notice(
+    locator: &Locator,
+    settings: &Settings,
+) -> Option<Diagnostic> {
+    // Ignore files that are too small to contain a copyright notice.
+    if locator.len() < settings.copyright.min_file_size {
+        return None;
+    }
+
+    // Only search the first 1024 bytes in the file.
+    let contents = if locator.len() < 1024 {
+        locator.contents()
+    } else {
+        locator.up_to(TextSize::from(1024))
+    };
+
+    // Locate the copyright notice.
+    if let Some(match_) = settings.copyright.notice_rgx.find(contents) {
+        match settings.copyright.author {
+            Some(ref author) => {
+                // Ensure that it's immediately followed by the author.
+                if contents[match_.end()..].trim_start().starts_with(author) {
+                    return None;
+                }
+            }
+            None => return None,
+        }
+    }
+
+    Some(Diagnostic::new(
+        MissingCopyrightNotice,
+        TextRange::default(),
+    ))
+}

crates/ruff/src/rules/copyright/rules/mod.rs

@@ -0,0 +1,3 @@
+pub(crate) use missing_copyright_notice::{missing_copyright_notice, MissingCopyrightNotice};
+
+mod missing_copyright_notice;

crates/ruff/src/rules/copyright/settings.rs

@@ -0,0 +1,94 @@
+//! Settings for the `copyright` plugin.
+
+use once_cell::sync::Lazy;
+use regex::Regex;
+use serde::{Deserialize, Serialize};
+
+use ruff_macros::{CacheKey, CombineOptions, ConfigurationOptions};
+
+#[derive(
+    Debug, PartialEq, Eq, Serialize, Deserialize, Default, ConfigurationOptions, CombineOptions,
+)]
+#[serde(
+    deny_unknown_fields,
+    rename_all = "kebab-case",
+    rename = "CopyrightOptions"
+)]
+#[cfg_attr(feature = "schemars", derive(schemars::JsonSchema))]
+pub struct Options {
+    #[option(
+        default = r#"(?i)Copyright\s+(\(C\)\s+)?\d{4}([-,]\d{4})*"#,
+        value_type = "str",
+        example = r#"notice-rgx = "(?i)Copyright \\(C\\) \\d{4}""#
+    )]
+    /// The regular expression used to match the copyright notice, compiled
+    /// with the [`regex`](https://docs.rs/regex/latest/regex/) crate.
+    ///
+    /// Defaults to `(?i)Copyright\s+(\(C\)\s+)?\d{4}(-\d{4})*`, which matches
+    /// the following:
+    /// - `Copyright 2023`
+    /// - `Copyright (C) 2023`
+    /// - `Copyright 2021-2023`
+    /// - `Copyright (C) 2021-2023`
+    pub notice_rgx: Option<String>,
+    #[option(default = "None", value_type = "str", example = r#"author = "Ruff""#)]
+    /// Author to enforce within the copyright notice. If provided, the
+    /// author must be present immediately following the copyright notice.
+    pub author: Option<String>,
+    #[option(
+        default = r#"0"#,
+        value_type = "int",
+        example = r#"
+            # Avoid enforcing a header on files smaller than 1024 bytes.
+            min-file-size = 1024
+        "#
+    )]
+    /// A minimum file size (in bytes) required for a copyright notice to
+    /// be enforced. By default, all files are validated.
+    pub min_file_size: Option<usize>,
+}
+
+#[derive(Debug, CacheKey)]
+pub struct Settings {
+    pub notice_rgx: Regex,
+    pub author: Option<String>,
+    pub min_file_size: usize,
+}
+
+static COPYRIGHT: Lazy<Regex> =
+    Lazy::new(|| Regex::new(r"(?i)Copyright\s+(\(C\)\s+)?\d{4}(-\d{4})*").unwrap());
+
+impl Default for Settings {
+    fn default() -> Self {
+        Self {
+            notice_rgx: COPYRIGHT.clone(),
+            author: None,
+            min_file_size: 0,
+        }
+    }
+}
+
+impl From<Options> for Settings {
+    fn from(options: Options) -> Self {
+        Self {
+            notice_rgx: options
+                .notice_rgx
+                .map(|pattern| Regex::new(&pattern))
+                .transpose()
+                .expect("Invalid `notice-rgx`")
+                .unwrap_or_else(|| COPYRIGHT.clone()),
+            author: options.author,
+            min_file_size: options.min_file_size.unwrap_or_default(),
+        }
+    }
+}
+
+impl From<Settings> for Options {
+    fn from(settings: Settings) -> Self {
+        Self {
+            notice_rgx: Some(settings.notice_rgx.to_string()),
+            author: settings.author,
+            min_file_size: Some(settings.min_file_size),
+        }
+    }
+}

crates/ruff/src/rules/copyright/snapshots/ruff__rules__copyright__tests__invalid_author.snap

@@ -0,0 +1,12 @@
+---
+source: crates/ruff/src/rules/copyright/mod.rs
+---
+<filename>:1:1: CPY001 Missing copyright notice at top of file
+  |
+1 | # Copyright (C) 2023 Some Author
+  |  CPY001
+2 | 
+3 | import os
+  |
+
+