Call chain formatting in fluent style (#6151)

Implement fluent style/call chains. See the `call_chains.py` formatting
for examples.

This isn't fully like black because in `raise A from B` they allow `A`
breaking can influence the formatting of `B` even if it is already
multiline.

Similarity index:

| project      | main  | PR    |
|--------------|-------|-------|
| build        | ???   | 0.753 |
| django       | 0.991 | 0.998 |
| transformers | 0.993 | 0.994 |
| typeshed     | 0.723 | 0.723 |
| warehouse    | 0.978 | 0.994 |
| zulip        | 0.992 | 0.994 |

Call chain formatting is affected by
https://github.com/astral-sh/ruff/issues/627, but i'm cutting scope
here.

Closes #5343

**Test Plan**:
 * Added a dedicated call chains test file
 * The ecosystem checks found some bugs
 * I manually check django and zulip formatting

---------

Co-authored-by: Micha Reiser <micha@reiser.io>

crates/ruff_python_formatter/resources/test/fixtures/ruff/call_chains.py

@@ -0,0 +1,157 @@
+# Test cases for call chains and optional parentheses, with and without fluent style
+
+raise OsError("") from a.aaaaa(
+    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
+).a(aaaa)
+
+raise OsError(
+    "sökdjffffsldkfjlhsakfjhalsökafhsöfdahsödfjösaaksjdllllllllllllll"
+) from a.aaaaa(
+    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
+).a(
+    aaaa
+)
+
+a1 = Blog.objects.filter(entry__headline__contains="Lennon").filter(
+    entry__pub_date__year=2008
+)
+
+a2 = Blog.objects.filter(
+    entry__headline__contains="Lennon",
+).filter(
+    entry__pub_date__year=2008,
+)
+
+raise OsError("") from (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+)
+
+raise OsError("sökdjffffsldkfjlhsakfjhalsökafhsöfdahsödfjösaaksjdllllllllllllll") from (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+)
+
+# Break only after calls and indexing
+b1 = (
+    session.query(models.Customer.id)
+    .filter(
+        models.Customer.account_id == account_id, models.Customer.email == email_address
+    )
+    .count()
+)
+
+b2 = (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .limit_results[:10]
+    .filter(
+        entry__pub_date__month=10,
+    )
+)
+
+# Nested call chains
+c1 = (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    ).filter(
+        entry__pub_date__year=2008,
+    )
+    + Blog.objects.filter(
+        entry__headline__contains="McCartney",
+    )
+    .limit_results[:10]
+    .filter(
+        entry__pub_date__year=2010,
+    )
+).all()
+
+# Test different cases with trailing end of line comments:
+# * fluent style, fits: no parentheses -> ignore the expand_parent
+# * fluent style, doesn't fit: break all soft line breaks
+# * default, fits: no parentheses
+# * default, doesn't fit: parentheses but no soft line breaks
+
+# Fits, either style
+d11 = x.e().e().e() #
+d12 = (x.e().e().e()) #
+d13 = (
+    x.e() #
+    .e()
+    .e()
+)
+
+# Doesn't fit, default
+d2 = (
+    x.e().esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkfsdddd()  #
+)
+
+# Doesn't fit, fluent style
+d3 = (
+    x.e()  #
+    .esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk()
+    .esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk()
+)
+
+# Don't drop the bin op parentheses
+e1 = (1 + 2).w().t()
+e2 = (1 + 2)().w().t()
+e3 = (1 + 2)[1].w().t()
+
+# Treat preserved parentheses correctly
+f1 = (b().c()).d(1,)
+f2 = b().c().d(1,)
+f3 = (b).c().d(1,)
+f4 = (a)(b).c(1,)
+f5 = (a.b()).c(1,)
+
+# Indent in the parentheses without breaking
+g1 = (
+    queryset.distinct().order_by(field.name).values_list(field_name_flat_long_long=True)
+)
+
+# Fluent style in subexpressions
+if (
+    not a()
+    .b()
+    .cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc()
+):
+    pass
+h2 = (
+    bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
+    + ccccccccccccccccccccccccc()
+    .dddddddddddddddddddddd()
+    .eeeeeeeeee()
+    .ffffffffffffffffffffff()
+)
+
+# Parentheses aren't allowed on statement level, don't use fluent style here
+if True:
+    (alias).filter(content_typeold_content_type).update(
+        content_typenew_contesadfasfdant_type
+    )
+
+zero(
+    one,
+).two(
+    three,
+).four(
+    five,
+)
+
+

crates/ruff_python_formatter/resources/test/fixtures/ruff/expression/call.py

@@ -52,7 +52,7 @@ f(
     hey_this_is_a_very_long_call=1, it_has_funny_attributes_asdf_asdf=1, too_long_for_the_line=1, really=True
 )
 
-# TODO(konstin): Call chains/fluent interface (https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#call-chains)
+# Call chains/fluent interface (https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#call-chains)
 result = (
     session.query(models.Customer.id)
     .filter(

crates/ruff_python_formatter/src/comments/placement.rs

@@ -1016,7 +1016,7 @@ fn handle_attribute_comment<'a>(
             .contains(comment.slice().start())
     );
     if comment.line_position().is_end_of_line() {
-        // Attach to node with b
+        // Attach as trailing comment to a. The specific placement is only relevant for fluent style
         // ```python
         // x322 = (
         //     a
@@ -1024,7 +1024,7 @@ fn handle_attribute_comment<'a>(
         //     b
         // )
         // ```
-        CommentPlacement::trailing(comment.enclosing_node(), comment)
+        CommentPlacement::trailing(attribute.value.as_ref(), comment)
     } else {
         CommentPlacement::dangling(attribute, comment)
     }

crates/ruff_python_formatter/src/expression/expr_attribute.rs

@@ -1,15 +1,28 @@
+use ruff_formatter::{write, FormatRuleWithOptions};
+use ruff_python_ast::node::AnyNodeRef;
 use ruff_python_ast::{Constant, Expr, ExprAttribute, ExprConstant};
 
-use ruff_formatter::write;
-use ruff_python_ast::node::AnyNodeRef;
-
 use crate::comments::{leading_comments, trailing_comments};
-use crate::expression::parentheses::{NeedsParentheses, OptionalParentheses, Parentheses};
+use crate::expression::parentheses::{
+    is_expression_parenthesized, NeedsParentheses, OptionalParentheses, Parentheses,
+};
+use crate::expression::CallChainLayout;
 use crate::prelude::*;
 use crate::FormatNodeRule;
 
 #[derive(Default)]
-pub struct FormatExprAttribute;
+pub struct FormatExprAttribute {
+    call_chain_layout: CallChainLayout,
+}
+
+impl FormatRuleWithOptions<ExprAttribute, PyFormatContext<'_>> for FormatExprAttribute {
+    type Options = CallChainLayout;
+
+    fn with_options(mut self, options: Self::Options) -> Self {
+        self.call_chain_layout = options;
+        self
+    }
+}
 
 impl FormatNodeRule<ExprAttribute> for FormatExprAttribute {
     fn fmt_fields(&self, item: &ExprAttribute, f: &mut PyFormatter) -> FormatResult<()> {
@@ -20,6 +33,8 @@ impl FormatNodeRule<ExprAttribute> for FormatExprAttribute {
             ctx: _,
         } = item;
 
+        let call_chain_layout = self.call_chain_layout.apply_in_node(item, f);
+
         let needs_parentheses = matches!(
             value.as_ref(),
             Expr::Constant(ExprConstant {
@@ -37,11 +52,36 @@ impl FormatNodeRule<ExprAttribute> for FormatExprAttribute {
 
         if needs_parentheses {
             value.format().with_options(Parentheses::Always).fmt(f)?;
-        } else if let Expr::Attribute(expr_attribute) = value.as_ref() {
+        } else if call_chain_layout == CallChainLayout::Fluent {
-            // We're in a attribute chain (`a.b.c`). The outermost node adds parentheses if
+            match value.as_ref() {
-            // required, the inner ones don't need them so we skip the `Expr` formatting that
+                Expr::Attribute(expr) => {
-            // normally adds the parentheses.
+                    expr.format().with_options(call_chain_layout).fmt(f)?;
-            expr_attribute.format().fmt(f)?;
+                }
+                Expr::Call(expr) => {
+                    expr.format().with_options(call_chain_layout).fmt(f)?;
+                    if call_chain_layout == CallChainLayout::Fluent {
+                        // Format the dot on its own line
+                        soft_line_break().fmt(f)?;
+                    }
+                }
+                Expr::Subscript(expr) => {
+                    expr.format().with_options(call_chain_layout).fmt(f)?;
+                    if call_chain_layout == CallChainLayout::Fluent {
+                        // Format the dot on its own line
+                        soft_line_break().fmt(f)?;
+                    }
+                }
+                _ => {
+                    // This matches [`CallChainLayout::from_expression`]
+                    if is_expression_parenthesized(value.as_ref().into(), f.context().source()) {
+                        value.format().with_options(Parentheses::Always).fmt(f)?;
+                        // Format the dot on its own line
+                        soft_line_break().fmt(f)?;
+                    } else {
+                        value.format().fmt(f)?;
+                    }
+                }
+            }
         } else {
             value.format().fmt(f)?;
         }
@@ -50,16 +90,51 @@ impl FormatNodeRule<ExprAttribute> for FormatExprAttribute {
             hard_line_break().fmt(f)?;
         }
 
-        write!(
+        if call_chain_layout == CallChainLayout::Fluent {
-            f,
+            // Fluent style has line breaks before the dot
-            [
+            // ```python
-                text("."),
+            // blogs3 = (
-                trailing_comments(trailing_dot_comments),
+            //     Blog.objects.filter(
-                (!leading_attribute_comments.is_empty()).then_some(hard_line_break()),
+            //         entry__headline__contains="Lennon",
-                leading_comments(leading_attribute_comments),
+            //     )
-                attr.format()
+            //     .filter(
-            ]
+            //         entry__pub_date__year=2008,
-        )
+            //     )
+            //     .filter(
+            //         entry__pub_date__year=2008,
+            //     )
+            // )
+            // ```
+            write!(
+                f,
+                [
+                    (!leading_attribute_comments.is_empty()).then_some(hard_line_break()),
+                    leading_comments(leading_attribute_comments),
+                    text("."),
+                    trailing_comments(trailing_dot_comments),
+                    attr.format()
+                ]
+            )
+        } else {
+            // Regular style
+            // ```python
+            // blogs2 = Blog.objects.filter(
+            //     entry__headline__contains="Lennon",
+            // ).filter(
+            //     entry__pub_date__year=2008,
+            // )
+            // ```
+            write!(
+                f,
+                [
+                    text("."),
+                    trailing_comments(trailing_dot_comments),
+                    (!leading_attribute_comments.is_empty()).then_some(hard_line_break()),
+                    leading_comments(leading_attribute_comments),
+                    attr.format()
+                ]
+            )
+        }
     }
 
     fn fmt_dangling_comments(
@@ -79,7 +154,11 @@ impl NeedsParentheses for ExprAttribute {
         context: &PyFormatContext,
     ) -> OptionalParentheses {
         // Checks if there are any own line comments in an attribute chain (a.b.c).
-        if context
+        if CallChainLayout::from_expression(self.into(), context.source())
+            == CallChainLayout::Fluent
+        {
+            OptionalParentheses::Multiline
+        } else if context
             .comments()
             .dangling_comments(self)
             .iter()

crates/ruff_python_formatter/src/expression/expr_bin_op.rs

@@ -5,9 +5,7 @@ use ruff_python_ast::{
 };
 use smallvec::SmallVec;
 
-use ruff_formatter::{
+use ruff_formatter::{format_args, write, FormatOwnedWithRule, FormatRefWithRule};
-    format_args, write, FormatOwnedWithRule, FormatRefWithRule, FormatRuleWithOptions,
-};
 use ruff_python_ast::node::{AnyNodeRef, AstNode};
 use ruff_python_ast::str::is_implicit_concatenation;
 
@@ -19,23 +17,11 @@ use crate::expression::parentheses::{
     NeedsParentheses, OptionalParentheses,
 };
 use crate::expression::string::StringLayout;
-use crate::expression::Parentheses;
 use crate::prelude::*;
 use crate::FormatNodeRule;
 
 #[derive(Default)]
-pub struct FormatExprBinOp {
+pub struct FormatExprBinOp;
-    parentheses: Option<Parentheses>,
-}
-
-impl FormatRuleWithOptions<ExprBinOp, PyFormatContext<'_>> for FormatExprBinOp {
-    type Options = Option<Parentheses>;
-
-    fn with_options(mut self, options: Self::Options) -> Self {
-        self.parentheses = options;
-        self
-    }
-}
 
 impl FormatNodeRule<ExprBinOp> for FormatExprBinOp {
     fn fmt_fields(&self, item: &ExprBinOp, f: &mut PyFormatter) -> FormatResult<()> {

crates/ruff_python_formatter/src/expression/expr_call.rs

@@ -1,13 +1,25 @@
-use ruff_formatter::write;
+use crate::expression::CallChainLayout;
+use ruff_formatter::FormatRuleWithOptions;
 use ruff_python_ast::node::AnyNodeRef;
-use ruff_python_ast::ExprCall;
+use ruff_python_ast::{Expr, ExprCall};
 
 use crate::expression::parentheses::{NeedsParentheses, OptionalParentheses};
 use crate::prelude::*;
 use crate::FormatNodeRule;
 
 #[derive(Default)]
-pub struct FormatExprCall;
+pub struct FormatExprCall {
+    call_chain_layout: CallChainLayout,
+}
+
+impl FormatRuleWithOptions<ExprCall, PyFormatContext<'_>> for FormatExprCall {
+    type Options = CallChainLayout;
+
+    fn with_options(mut self, options: Self::Options) -> Self {
+        self.call_chain_layout = options;
+        self
+    }
+}
 
 impl FormatNodeRule<ExprCall> for FormatExprCall {
     fn fmt_fields(&self, item: &ExprCall, f: &mut PyFormatter) -> FormatResult<()> {
@@ -17,7 +29,32 @@ impl FormatNodeRule<ExprCall> for FormatExprCall {
             arguments,
         } = item;
 
-        write!(f, [func.format(), arguments.format()])
+        let call_chain_layout = self.call_chain_layout.apply_in_node(item, f);
+
+        let fmt_inner = format_with(|f| {
+            match func.as_ref() {
+                Expr::Attribute(expr) => expr.format().with_options(call_chain_layout).fmt(f)?,
+                Expr::Call(expr) => expr.format().with_options(call_chain_layout).fmt(f)?,
+                Expr::Subscript(expr) => expr.format().with_options(call_chain_layout).fmt(f)?,
+                _ => func.format().fmt(f)?,
+            }
+
+            arguments.format().fmt(f)
+        });
+
+        // Allow to indent the parentheses while
+        // ```python
+        // g1 = (
+        //     queryset.distinct().order_by(field.name).values_list(field_name_flat_long_long=True)
+        // )
+        // ```
+        if call_chain_layout == CallChainLayout::Fluent
+            && self.call_chain_layout == CallChainLayout::Default
+        {
+            group(&fmt_inner).fmt(f)
+        } else {
+            fmt_inner.fmt(f)
+        }
     }
 }
 
@@ -27,6 +64,12 @@ impl NeedsParentheses for ExprCall {
         _parent: AnyNodeRef,
         context: &PyFormatContext,
     ) -> OptionalParentheses {
-        self.func.needs_parentheses(self.into(), context)
+        if CallChainLayout::from_expression(self.into(), context.source())
+            == CallChainLayout::Fluent
+        {
+            OptionalParentheses::Multiline
+        } else {
+            self.func.needs_parentheses(self.into(), context)
+        }
     }
 }

crates/ruff_python_formatter/src/expression/expr_subscript.rs

@@ -1,18 +1,29 @@
-use ruff_python_ast::{Expr, ExprSubscript};
+use ruff_formatter::{format_args, write, FormatRuleWithOptions};
-
-use ruff_formatter::{format_args, write};
 use ruff_python_ast::node::{AnyNodeRef, AstNode};
+use ruff_python_ast::{Expr, ExprSubscript};
 
 use crate::comments::trailing_comments;
 use crate::context::PyFormatContext;
 use crate::context::{NodeLevel, WithNodeLevel};
 use crate::expression::expr_tuple::TupleParentheses;
 use crate::expression::parentheses::{NeedsParentheses, OptionalParentheses};
+use crate::expression::CallChainLayout;
 use crate::prelude::*;
 use crate::FormatNodeRule;
 
 #[derive(Default)]
-pub struct FormatExprSubscript;
+pub struct FormatExprSubscript {
+    call_chain_layout: CallChainLayout,
+}
+
+impl FormatRuleWithOptions<ExprSubscript, PyFormatContext<'_>> for FormatExprSubscript {
+    type Options = CallChainLayout;
+
+    fn with_options(mut self, options: Self::Options) -> Self {
+        self.call_chain_layout = options;
+        self
+    }
+}
 
 impl FormatNodeRule<ExprSubscript> for FormatExprSubscript {
     fn fmt_fields(&self, item: &ExprSubscript, f: &mut PyFormatter) -> FormatResult<()> {
@@ -23,6 +34,8 @@ impl FormatNodeRule<ExprSubscript> for FormatExprSubscript {
             ctx: _,
         } = item;
 
+        let call_chain_layout = self.call_chain_layout.apply_in_node(item, f);
+
         let comments = f.context().comments().clone();
         let dangling_comments = comments.dangling_comments(item.as_any_node_ref());
         debug_assert!(
@@ -30,12 +43,19 @@ impl FormatNodeRule<ExprSubscript> for FormatExprSubscript {
             "A subscript expression can only have a single dangling comment, the one after the bracket"
         );
 
+        let format_value = format_with(|f| match value.as_ref() {
+            Expr::Attribute(expr) => expr.format().with_options(call_chain_layout).fmt(f),
+            Expr::Call(expr) => expr.format().with_options(call_chain_layout).fmt(f),
+            Expr::Subscript(expr) => expr.format().with_options(call_chain_layout).fmt(f),
+            _ => value.format().fmt(f),
+        });
+
         if let NodeLevel::Expression(Some(_)) = f.context().node_level() {
             // Enforce the optional parentheses for parenthesized values.
             let mut f = WithNodeLevel::new(NodeLevel::Expression(None), f);
-            write!(f, [value.format()])?;
+            write!(f, [format_value])?;
         } else {
-            value.format().fmt(f)?;
+            format_value.fmt(f)?;
         }
 
         let format_slice = format_with(|f: &mut PyFormatter| {
@@ -73,10 +93,16 @@ impl NeedsParentheses for ExprSubscript {
     fn needs_parentheses(
         &self,
         _parent: AnyNodeRef,
-        _context: &PyFormatContext,
+        context: &PyFormatContext,
     ) -> OptionalParentheses {
         {
-            OptionalParentheses::Never
+            if CallChainLayout::from_expression(self.into(), context.source())
+                == CallChainLayout::Fluent
+            {
+                OptionalParentheses::Multiline
+            } else {
+                self.value.needs_parentheses(self.into(), context)
+            }
         }
     }
 }

crates/ruff_python_formatter/src/expression/mod.rs

@@ -1,13 +1,12 @@
 use std::cmp::Ordering;
 
-use ruff_python_ast as ast;
-use ruff_python_ast::{Expr, Operator};
-
 use ruff_formatter::{
     write, FormatOwnedWithRule, FormatRefWithRule, FormatRule, FormatRuleWithOptions,
 };
+use ruff_python_ast as ast;
 use ruff_python_ast::node::AnyNodeRef;
 use ruff_python_ast::visitor::preorder::{walk_expr, PreorderVisitor};
+use ruff_python_ast::{Expr, Operator};
 
 use crate::builders::parenthesize_if_expands;
 use crate::context::{NodeLevel, WithNodeLevel};
@@ -70,7 +69,7 @@ impl FormatRule<Expr, PyFormatContext<'_>> for FormatExpr {
         let format_expr = format_with(|f| match expression {
             Expr::BoolOp(expr) => expr.format().with_options(Some(parentheses)).fmt(f),
             Expr::NamedExpr(expr) => expr.format().fmt(f),
-            Expr::BinOp(expr) => expr.format().with_options(Some(parentheses)).fmt(f),
+            Expr::BinOp(expr) => expr.format().fmt(f),
             Expr::UnaryOp(expr) => expr.format().fmt(f),
             Expr::Lambda(expr) => expr.format().fmt(f),
             Expr::IfExp(expr) => expr.format().fmt(f),
@@ -100,9 +99,10 @@ impl FormatRule<Expr, PyFormatContext<'_>> for FormatExpr {
 
         let parenthesize = match parentheses {
             Parentheses::Preserve => {
-                is_expression_parenthesized(AnyNodeRef::from(expression), f.context().source())
+                is_expression_parenthesized(expression.into(), f.context().source())
             }
             Parentheses::Always => true,
+            // Fluent style means we already have parentheses
             Parentheses::Never => false,
         };
 
@@ -157,7 +157,7 @@ impl Format<PyFormatContext<'_>> for MaybeParenthesizeExpression<'_> {
 
         let comments = f.context().comments();
         let preserve_parentheses = parenthesize.is_optional()
-            && is_expression_parenthesized(AnyNodeRef::from(*expression), f.context().source());
+            && is_expression_parenthesized((*expression).into(), f.context().source());
 
         let has_comments = comments.has_leading_comments(*expression)
             || comments.has_trailing_own_line_comments(*expression);
@@ -178,9 +178,10 @@ impl Format<PyFormatContext<'_>> for MaybeParenthesizeExpression<'_> {
             Parenthesize::Optional | Parenthesize::IfBreaks => needs_parentheses,
         };
 
+        let can_omit_optional_parentheses = can_omit_optional_parentheses(expression, f.context());
         match needs_parentheses {
             OptionalParentheses::Multiline if *parenthesize != Parenthesize::IfRequired => {
-                if can_omit_optional_parentheses(expression, f.context()) {
+                if can_omit_optional_parentheses {
                     optional_parentheses(&expression.format().with_options(Parentheses::Never))
                         .fmt(f)
                 } else {
@@ -265,7 +266,24 @@ impl<'ast> IntoFormat<PyFormatContext<'ast>> for Expr {
 fn can_omit_optional_parentheses(expr: &Expr, context: &PyFormatContext) -> bool {
     let mut visitor = CanOmitOptionalParenthesesVisitor::new(context.source());
     visitor.visit_subexpression(expr);
-    visitor.can_omit()
+
+    if visitor.max_priority_count > 1 {
+        false
+    } else if visitor.max_priority == OperatorPriority::Attribute {
+        true
+    } else if !visitor.any_parenthesized_expressions {
+        // Only use the more complex IR when there is any expression that we can possibly split by
+        false
+    } else {
+        // Only use the layout if the first or last expression has parentheses of some sort.
+        let first_parenthesized = visitor
+            .first
+            .is_some_and(|first| has_parentheses(first, visitor.source));
+        let last_parenthesized = visitor
+            .last
+            .is_some_and(|last| has_parentheses(last, visitor.source));
+        first_parenthesized || last_parenthesized
+    }
 }
 
 #[derive(Clone, Debug)]
@@ -365,8 +383,10 @@ impl<'input> CanOmitOptionalParenthesesVisitor<'input> {
                 self.last = Some(expr);
                 return;
             }
-            Expr::Subscript(_) => {
+            Expr::Subscript(ast::ExprSubscript { value, .. }) => {
-                // Don't walk the value. Splitting before the value looks weird.
+                self.any_parenthesized_expressions = true;
+                // Only walk the function, the subscript is always parenthesized
+                self.visit_expr(value);
                 // Don't walk the slice, because the slice is always parenthesized.
                 return;
             }
@@ -409,26 +429,6 @@ impl<'input> CanOmitOptionalParenthesesVisitor<'input> {
 
         walk_expr(self, expr);
     }
-
-    fn can_omit(self) -> bool {
-        if self.max_priority_count > 1 {
-            false
-        } else if self.max_priority == OperatorPriority::Attribute {
-            true
-        } else if !self.any_parenthesized_expressions {
-            // Only use the more complex IR when there is any expression that we can possibly split by
-            false
-        } else {
-            // Only use the layout if the first or last expression has parentheses of some sort.
-            let first_parenthesized = self
-                .first
-                .is_some_and(|first| has_parentheses(first, self.source));
-            let last_parenthesized = self
-                .last
-                .is_some_and(|last| has_parentheses(last, self.source));
-            first_parenthesized || last_parenthesized
-        }
-    }
 }
 
 impl<'input> PreorderVisitor<'input> for CanOmitOptionalParenthesesVisitor<'input> {
@@ -448,6 +448,124 @@ impl<'input> PreorderVisitor<'input> for CanOmitOptionalParenthesesVisitor<'inpu
     }
 }
 
+/// A call chain consists only of attribute access (`.` operator), function/method calls and
+/// subscripts. We use fluent style for the call chain if there are at least two attribute dots
+/// after call parentheses or subscript brackets. In case of fluent style the parentheses/bracket
+/// will close on the previous line and the dot gets its own line, otherwise the line will start
+/// with the closing parentheses/bracket and the dot follows immediately after.
+///
+/// Below, the left hand side of the addition has only a single attribute access after a call, the
+/// second `.filter`. The first `.filter` is a call, but it doesn't follow a call. The right hand
+/// side has two, the `.limit_results` after the call and the `.filter` after the subscript, so it
+/// gets formatted in fluent style. The outer expression we assign to `blogs` has zero since the
+/// `.all` follows attribute parentheses and not call parentheses.
+///
+/// ```python
+/// blogs = (
+///     Blog.objects.filter(
+///         entry__headline__contains="Lennon",
+///     ).filter(
+///         entry__pub_date__year=2008,
+///     )
+///     + Blog.objects.filter(
+///         entry__headline__contains="McCartney",
+///     )
+///     .limit_results[:10]
+///     .filter(
+///         entry__pub_date__year=2010,
+///     )
+/// ).all()
+/// ```
+#[derive(Copy, Clone, Debug, Default, PartialEq, Eq)]
+pub enum CallChainLayout {
+    /// The root of a call chain
+    #[default]
+    Default,
+
+    /// A nested call chain element that uses fluent style.
+    Fluent,
+
+    /// A nested call chain element not using fluent style.
+    NonFluent,
+}
+
+impl CallChainLayout {
+    pub(crate) fn from_expression(mut expr: AnyNodeRef, source: &str) -> Self {
+        let mut attributes_after_parentheses = 0;
+        loop {
+            match expr {
+                AnyNodeRef::ExprAttribute(ast::ExprAttribute { value, .. }) => {
+                    expr = AnyNodeRef::from(value.as_ref());
+                    // ```
+                    // f().g
+                    // ^^^ value
+                    // data[:100].T`
+                    // ^^^^^^^^^^ value
+                    // ```
+                    if matches!(value.as_ref(), Expr::Call(_) | Expr::Subscript(_)) {
+                        attributes_after_parentheses += 1;
+                    } else if is_expression_parenthesized(expr, source) {
+                        // `(a).b`. We preserve these parentheses so don't recurse
+                        attributes_after_parentheses += 1;
+                        break;
+                    }
+                }
+                // ```
+                // f()
+                // ^^^ expr
+                // ^ func
+                // data[:100]
+                // ^^^^^^^^^^ expr
+                // ^^^^ value
+                // ```
+                AnyNodeRef::ExprCall(ast::ExprCall { func: inner, .. })
+                | AnyNodeRef::ExprSubscript(ast::ExprSubscript { value: inner, .. }) => {
+                    expr = AnyNodeRef::from(inner.as_ref());
+                }
+                _ => {
+                    // We to format the following in fluent style:
+                    // ```
+                    // f2 = (a).w().t(1,)
+                    //       ^ expr
+                    // ```
+                    if is_expression_parenthesized(expr, source) {
+                        attributes_after_parentheses += 1;
+                    }
+                    break;
+                }
+            }
+            // We preserve these parentheses so don't recurse
+            if is_expression_parenthesized(expr, source) {
+                break;
+            }
+        }
+        if attributes_after_parentheses < 2 {
+            CallChainLayout::NonFluent
+        } else {
+            CallChainLayout::Fluent
+        }
+    }
+
+    /// Determine whether to actually apply fluent layout in attribute, call and subscript
+    /// formatting
+    pub(crate) fn apply_in_node<'a>(
+        self,
+        item: impl Into<AnyNodeRef<'a>>,
+        f: &mut PyFormatter,
+    ) -> CallChainLayout {
+        match self {
+            CallChainLayout::Default => {
+                if f.context().node_level().is_parenthesized() {
+                    CallChainLayout::from_expression(item.into(), f.context().source())
+                } else {
+                    CallChainLayout::NonFluent
+                }
+            }
+            layout @ (CallChainLayout::Fluent | CallChainLayout::NonFluent) => layout,
+        }
+    }
+}
+
 fn has_parentheses(expr: &Expr, source: &str) -> bool {
     has_own_parentheses(expr) || is_expression_parenthesized(AnyNodeRef::from(expr), source)
 }

crates/ruff_python_formatter/src/expression/parentheses.rs

@@ -316,3 +316,20 @@ impl<'ast> Format<PyFormatContext<'ast>> for FormatInParenthesesOnlyGroup<'_, 'a
         }
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use crate::expression::parentheses::is_expression_parenthesized;
+    use ruff_python_ast::node::AnyNodeRef;
+    use ruff_python_parser::parse_expression;
+
+    #[test]
+    fn test_has_parentheses() {
+        let expression = r#"(b().c("")).d()"#;
+        let expr = parse_expression(expression, "<filename>").unwrap();
+        assert!(!is_expression_parenthesized(
+            AnyNodeRef::from(&expr),
+            expression
+        ));
+    }
+}

crates/ruff_python_formatter/tests/snapshots/black_compatibility@simple_cases__comments4.py.snap

@@ -1,348 +0,0 @@
----
-source: crates/ruff_python_formatter/tests/fixtures.rs
-input_file: crates/ruff_python_formatter/resources/test/fixtures/black/simple_cases/comments4.py
----
-## Input
-
-```py
-from com.my_lovely_company.my_lovely_team.my_lovely_project.my_lovely_component import (
-    MyLovelyCompanyTeamProjectComponent,  # NOT DRY
-)
-from com.my_lovely_company.my_lovely_team.my_lovely_project.my_lovely_component import (
-    MyLovelyCompanyTeamProjectComponent as component,  # DRY
-)
-
-
-class C:
-    @pytest.mark.parametrize(
-        ("post_data", "message"),
-        [
-            # metadata_version errors.
-            (
-                {},
-                "None is an invalid value for Metadata-Version. Error: This field is"
-                " required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "-1"},
-                "'-1' is an invalid value for Metadata-Version. Error: Unknown Metadata"
-                " Version see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            # name errors.
-            (
-                {"metadata_version": "1.2"},
-                "'' is an invalid value for Name. Error: This field is required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "1.2", "name": "foo-"},
-                "'foo-' is an invalid value for Name. Error: Must start and end with a"
-                " letter or numeral and contain only ascii numeric and '.', '_' and"
-                " '-'. see https://packaging.python.org/specifications/core-metadata",
-            ),
-            # version errors.
-            (
-                {"metadata_version": "1.2", "name": "example"},
-                "'' is an invalid value for Version. Error: This field is required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "1.2", "name": "example", "version": "dog"},
-                "'dog' is an invalid value for Version. Error: Must start and end with"
-                " a letter or numeral and contain only ascii numeric and '.', '_' and"
-                " '-'. see https://packaging.python.org/specifications/core-metadata",
-            ),
-        ],
-    )
-    def test_fails_invalid_post_data(
-        self, pyramid_config, db_request, post_data, message
-    ):
-        pyramid_config.testing_securitypolicy(userid=1)
-        db_request.POST = MultiDict(post_data)
-
-
-def foo(list_a, list_b):
-    results = (
-        User.query.filter(User.foo == "bar")
-        .filter(  # Because foo.
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        )
-        .filter(User.xyz.is_(None))
-        # Another comment about the filtering on is_quux goes here.
-        .filter(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True)))
-        .order_by(User.created_at.desc())
-        .with_for_update(key_share=True)
-        .all()
-    )
-    return results
-
-
-def foo2(list_a, list_b):
-    # Standalone comment reasonably placed.
-    return (
-        User.query.filter(User.foo == "bar")
-        .filter(
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        )
-        .filter(User.xyz.is_(None))
-    )
-
-
-def foo3(list_a, list_b):
-    return (
-        # Standalone comment but weirdly placed.
-        User.query.filter(User.foo == "bar")
-        .filter(
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        )
-        .filter(User.xyz.is_(None))
-    )
-```
-
-## Black Differences
-
-```diff
---- Black
-+++ Ruff
-@@ -58,37 +58,28 @@
- 
- def foo(list_a, list_b):
-     results = (
--        User.query.filter(User.foo == "bar")
--        .filter(  # Because foo.
-+        User.query.filter(User.foo == "bar").filter(  # Because foo.
-             db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
--        )
--        .filter(User.xyz.is_(None))
-+        ).filter(User.xyz.is_(None)).
-         # Another comment about the filtering on is_quux goes here.
--        .filter(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True)))
--        .order_by(User.created_at.desc())
--        .with_for_update(key_share=True)
--        .all()
-+        filter(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True))).order_by(
-+            User.created_at.desc()
-+        ).with_for_update(key_share=True).all()
-     )
-     return results
- 
- 
- def foo2(list_a, list_b):
-     # Standalone comment reasonably placed.
--    return (
--        User.query.filter(User.foo == "bar")
--        .filter(
--            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
--        )
--        .filter(User.xyz.is_(None))
--    )
-+    return User.query.filter(User.foo == "bar").filter(
-+        db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-+    ).filter(User.xyz.is_(None))
- 
- 
- def foo3(list_a, list_b):
-     return (
-         # Standalone comment but weirdly placed.
--        User.query.filter(User.foo == "bar")
--        .filter(
-+        User.query.filter(User.foo == "bar").filter(
-             db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
--        )
--        .filter(User.xyz.is_(None))
-+        ).filter(User.xyz.is_(None))
-     )
-```
-
-## Ruff Output
-
-```py
-from com.my_lovely_company.my_lovely_team.my_lovely_project.my_lovely_component import (
-    MyLovelyCompanyTeamProjectComponent,  # NOT DRY
-)
-from com.my_lovely_company.my_lovely_team.my_lovely_project.my_lovely_component import (
-    MyLovelyCompanyTeamProjectComponent as component,  # DRY
-)
-
-
-class C:
-    @pytest.mark.parametrize(
-        ("post_data", "message"),
-        [
-            # metadata_version errors.
-            (
-                {},
-                "None is an invalid value for Metadata-Version. Error: This field is"
-                " required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "-1"},
-                "'-1' is an invalid value for Metadata-Version. Error: Unknown Metadata"
-                " Version see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            # name errors.
-            (
-                {"metadata_version": "1.2"},
-                "'' is an invalid value for Name. Error: This field is required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "1.2", "name": "foo-"},
-                "'foo-' is an invalid value for Name. Error: Must start and end with a"
-                " letter or numeral and contain only ascii numeric and '.', '_' and"
-                " '-'. see https://packaging.python.org/specifications/core-metadata",
-            ),
-            # version errors.
-            (
-                {"metadata_version": "1.2", "name": "example"},
-                "'' is an invalid value for Version. Error: This field is required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "1.2", "name": "example", "version": "dog"},
-                "'dog' is an invalid value for Version. Error: Must start and end with"
-                " a letter or numeral and contain only ascii numeric and '.', '_' and"
-                " '-'. see https://packaging.python.org/specifications/core-metadata",
-            ),
-        ],
-    )
-    def test_fails_invalid_post_data(
-        self, pyramid_config, db_request, post_data, message
-    ):
-        pyramid_config.testing_securitypolicy(userid=1)
-        db_request.POST = MultiDict(post_data)
-
-
-def foo(list_a, list_b):
-    results = (
-        User.query.filter(User.foo == "bar").filter(  # Because foo.
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        ).filter(User.xyz.is_(None)).
-        # Another comment about the filtering on is_quux goes here.
-        filter(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True))).order_by(
-            User.created_at.desc()
-        ).with_for_update(key_share=True).all()
-    )
-    return results
-
-
-def foo2(list_a, list_b):
-    # Standalone comment reasonably placed.
-    return User.query.filter(User.foo == "bar").filter(
-        db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-    ).filter(User.xyz.is_(None))
-
-
-def foo3(list_a, list_b):
-    return (
-        # Standalone comment but weirdly placed.
-        User.query.filter(User.foo == "bar").filter(
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        ).filter(User.xyz.is_(None))
-    )
-```
-
-## Black Output
-
-```py
-from com.my_lovely_company.my_lovely_team.my_lovely_project.my_lovely_component import (
-    MyLovelyCompanyTeamProjectComponent,  # NOT DRY
-)
-from com.my_lovely_company.my_lovely_team.my_lovely_project.my_lovely_component import (
-    MyLovelyCompanyTeamProjectComponent as component,  # DRY
-)
-
-
-class C:
-    @pytest.mark.parametrize(
-        ("post_data", "message"),
-        [
-            # metadata_version errors.
-            (
-                {},
-                "None is an invalid value for Metadata-Version. Error: This field is"
-                " required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "-1"},
-                "'-1' is an invalid value for Metadata-Version. Error: Unknown Metadata"
-                " Version see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            # name errors.
-            (
-                {"metadata_version": "1.2"},
-                "'' is an invalid value for Name. Error: This field is required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "1.2", "name": "foo-"},
-                "'foo-' is an invalid value for Name. Error: Must start and end with a"
-                " letter or numeral and contain only ascii numeric and '.', '_' and"
-                " '-'. see https://packaging.python.org/specifications/core-metadata",
-            ),
-            # version errors.
-            (
-                {"metadata_version": "1.2", "name": "example"},
-                "'' is an invalid value for Version. Error: This field is required. see"
-                " https://packaging.python.org/specifications/core-metadata",
-            ),
-            (
-                {"metadata_version": "1.2", "name": "example", "version": "dog"},
-                "'dog' is an invalid value for Version. Error: Must start and end with"
-                " a letter or numeral and contain only ascii numeric and '.', '_' and"
-                " '-'. see https://packaging.python.org/specifications/core-metadata",
-            ),
-        ],
-    )
-    def test_fails_invalid_post_data(
-        self, pyramid_config, db_request, post_data, message
-    ):
-        pyramid_config.testing_securitypolicy(userid=1)
-        db_request.POST = MultiDict(post_data)
-
-
-def foo(list_a, list_b):
-    results = (
-        User.query.filter(User.foo == "bar")
-        .filter(  # Because foo.
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        )
-        .filter(User.xyz.is_(None))
-        # Another comment about the filtering on is_quux goes here.
-        .filter(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True)))
-        .order_by(User.created_at.desc())
-        .with_for_update(key_share=True)
-        .all()
-    )
-    return results
-
-
-def foo2(list_a, list_b):
-    # Standalone comment reasonably placed.
-    return (
-        User.query.filter(User.foo == "bar")
-        .filter(
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        )
-        .filter(User.xyz.is_(None))
-    )
-
-
-def foo3(list_a, list_b):
-    return (
-        # Standalone comment but weirdly placed.
-        User.query.filter(User.foo == "bar")
-        .filter(
-            db.or_(User.field_a.astext.in_(list_a), User.field_b.astext.in_(list_b))
-        )
-        .filter(User.xyz.is_(None))
-    )
-```
-
-

crates/ruff_python_formatter/tests/snapshots/black_compatibility@simple_cases__expression.py.snap

@@ -300,40 +300,7 @@ last_call()
  )  # note: no trailing comma pre-3.6
  call(*gidgets[:2])
  call(a, *gidgets[:2])
-@@ -208,24 +209,14 @@
+@@ -328,13 +329,18 @@
- what_is_up_with_those_new_coord_names = (coord_names | set(vars_to_create)) - set(
-     vars_to_remove
- )
--result = (
--    session.query(models.Customer.id)
--    .filter(
--        models.Customer.account_id == account_id, models.Customer.email == email_address
--    )
--    .order_by(models.Customer.id.asc())
--    .all()
--)
--result = (
--    session.query(models.Customer.id)
--    .filter(
--        models.Customer.account_id == account_id, models.Customer.email == email_address
--    )
--    .order_by(
--        models.Customer.id.asc(),
--    )
--    .all()
--)
-+result = session.query(models.Customer.id).filter(
-+    models.Customer.account_id == account_id, models.Customer.email == email_address
-+).order_by(models.Customer.id.asc()).all()
-+result = session.query(models.Customer.id).filter(
-+    models.Customer.account_id == account_id, models.Customer.email == email_address
-+).order_by(
-+    models.Customer.id.asc(),
-+).all()
- Ø = set()
- authors.łukasz.say_thanks()
- mapping = {
-@@ -328,13 +319,18 @@
  ):
      return True
  if (
@@ -355,7 +322,7 @@ last_call()
      ^ aaaaaaaa.i << aaaaaaaa.k >> aaaaaaaa.l**aaaaaaaa.m // aaaaaaaa.n
  ):
      return True
-@@ -342,7 +338,8 @@
+@@ -342,7 +348,8 @@
      ~aaaaaaaaaaaaaaaa.a
      + aaaaaaaaaaaaaaaa.b
      - aaaaaaaaaaaaaaaa.c * aaaaaaaaaaaaaaaa.d @ aaaaaaaaaaaaaaaa.e
@@ -581,14 +548,24 @@ what_is_up_with_those_new_coord_names = (coord_names + set(vars_to_create)) + se
 what_is_up_with_those_new_coord_names = (coord_names | set(vars_to_create)) - set(
     vars_to_remove
 )
-result = session.query(models.Customer.id).filter(
+result = (
-    models.Customer.account_id == account_id, models.Customer.email == email_address
+    session.query(models.Customer.id)
-).order_by(models.Customer.id.asc()).all()
+    .filter(
-result = session.query(models.Customer.id).filter(
+        models.Customer.account_id == account_id, models.Customer.email == email_address
-    models.Customer.account_id == account_id, models.Customer.email == email_address
+    )
-).order_by(
+    .order_by(models.Customer.id.asc())
-    models.Customer.id.asc(),
+    .all()
-).all()
+)
+result = (
+    session.query(models.Customer.id)
+    .filter(
+        models.Customer.account_id == account_id, models.Customer.email == email_address
+    )
+    .order_by(
+        models.Customer.id.asc(),
+    )
+    .all()
+)
 Ø = set()
 authors.łukasz.say_thanks()
 mapping = {

crates/ruff_python_formatter/tests/snapshots/black_compatibility@simple_cases__fmtonoff.py.snap

@@ -287,7 +287,7 @@ d={'a':1,
          # fmt: on
          goes + here,
          andhere,
-@@ -80,38 +94,36 @@
+@@ -80,38 +94,42 @@
  
  def import_as_names():
      # fmt: off
@@ -330,14 +330,19 @@ d={'a':1,
 -        .filter(models.Customer.account_id == account_id,
 -                models.Customer.email == email_address)\
 -        .order_by(models.Customer.id.asc())\
--        .all()
++    result = (
-+    result = session.query(models.Customer.id).filter(
++        session.query(models.Customer.id)
-+        models.Customer.account_id == account_id, models.Customer.email == email_address
++        .filter(
-+    ).order_by(models.Customer.id.asc()).all()
++            models.Customer.account_id == account_id,
++            models.Customer.email == email_address,
++        )
++        .order_by(models.Customer.id.asc())
+         .all()
++    )
      # fmt: on
  
  
-@@ -132,10 +144,10 @@
+@@ -132,10 +150,10 @@
      """Another known limitation."""
      # fmt: on
      # fmt: off
@@ -352,7 +357,7 @@ d={'a':1,
      # fmt: on
      # fmt: off
      # ...but comments still get reformatted even though they should not be
-@@ -153,9 +165,7 @@
+@@ -153,9 +171,7 @@
              )
          )
          # fmt: off
@@ -363,7 +368,7 @@ d={'a':1,
          # fmt: on
      _type_comment_re = re.compile(
          r"""
-@@ -178,7 +188,7 @@
+@@ -178,7 +194,7 @@
          $
          """,
          # fmt: off
@@ -372,7 +377,7 @@ d={'a':1,
          # fmt: on
      )
  
-@@ -216,8 +226,7 @@
+@@ -216,8 +232,7 @@
      xxxxxxxxxx_xxxxxxxxxxx_xxxxxxx_xxxxxxxxx=5,
  )
  # fmt: off
@@ -512,9 +517,15 @@ def yield_expr():
 
 def example(session):
     # fmt: off
-    result = session.query(models.Customer.id).filter(
+    result = (
-        models.Customer.account_id == account_id, models.Customer.email == email_address
+        session.query(models.Customer.id)
-    ).order_by(models.Customer.id.asc()).all()
+        .filter(
+            models.Customer.account_id == account_id,
+            models.Customer.email == email_address,
+        )
+        .order_by(models.Customer.id.asc())
+        .all()
+    )
     # fmt: on
 
 

crates/ruff_python_formatter/tests/snapshots/black_compatibility@simple_cases__function.py.snap

@@ -107,7 +107,7 @@ def __await__(): return (yield)
 ```diff
 --- Black
 +++ Ruff
-@@ -65,18 +65,14 @@
+@@ -65,6 +65,7 @@
  
  def spaces2(result=_core.Value(None)):
      assert fut is self._read_fut, (fut, self._read_fut)
@@ -115,22 +115,6 @@ def __await__(): return (yield)
  
  
  def example(session):
--    result = (
--        session.query(models.Customer.id)
--        .filter(
--            models.Customer.account_id == account_id,
--            models.Customer.email == email_address,
--        )
--        .order_by(models.Customer.id.asc())
--        .all()
--    )
-+    result = session.query(models.Customer.id).filter(
-+        models.Customer.account_id == account_id,
-+        models.Customer.email == email_address,
-+    ).order_by(models.Customer.id.asc()).all()
- 
- 
- def long_lines():
 ```
 
 ## Ruff Output
@@ -207,10 +191,15 @@ def spaces2(result=_core.Value(None)):
 
 
 def example(session):
-    result = session.query(models.Customer.id).filter(
+    result = (
-        models.Customer.account_id == account_id,
+        session.query(models.Customer.id)
-        models.Customer.email == email_address,
+        .filter(
-    ).order_by(models.Customer.id.asc()).all()
+            models.Customer.account_id == account_id,
+            models.Customer.email == email_address,
+        )
+        .order_by(models.Customer.id.asc())
+        .all()
+    )
 
 
 def long_lines():

crates/ruff_python_formatter/tests/snapshots/format@call_chains.py.snap

@@ -0,0 +1,340 @@
+---
+source: crates/ruff_python_formatter/tests/fixtures.rs
+input_file: crates/ruff_python_formatter/resources/test/fixtures/ruff/call_chains.py
+---
+## Input
+```py
+# Test cases for call chains and optional parentheses, with and without fluent style
+
+raise OsError("") from a.aaaaa(
+    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
+).a(aaaa)
+
+raise OsError(
+    "sökdjffffsldkfjlhsakfjhalsökafhsöfdahsödfjösaaksjdllllllllllllll"
+) from a.aaaaa(
+    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
+).a(
+    aaaa
+)
+
+a1 = Blog.objects.filter(entry__headline__contains="Lennon").filter(
+    entry__pub_date__year=2008
+)
+
+a2 = Blog.objects.filter(
+    entry__headline__contains="Lennon",
+).filter(
+    entry__pub_date__year=2008,
+)
+
+raise OsError("") from (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+)
+
+raise OsError("sökdjffffsldkfjlhsakfjhalsökafhsöfdahsödfjösaaksjdllllllllllllll") from (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+)
+
+# Break only after calls and indexing
+b1 = (
+    session.query(models.Customer.id)
+    .filter(
+        models.Customer.account_id == account_id, models.Customer.email == email_address
+    )
+    .count()
+)
+
+b2 = (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .limit_results[:10]
+    .filter(
+        entry__pub_date__month=10,
+    )
+)
+
+# Nested call chains
+c1 = (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    ).filter(
+        entry__pub_date__year=2008,
+    )
+    + Blog.objects.filter(
+        entry__headline__contains="McCartney",
+    )
+    .limit_results[:10]
+    .filter(
+        entry__pub_date__year=2010,
+    )
+).all()
+
+# Test different cases with trailing end of line comments:
+# * fluent style, fits: no parentheses -> ignore the expand_parent
+# * fluent style, doesn't fit: break all soft line breaks
+# * default, fits: no parentheses
+# * default, doesn't fit: parentheses but no soft line breaks
+
+# Fits, either style
+d11 = x.e().e().e() #
+d12 = (x.e().e().e()) #
+d13 = (
+    x.e() #
+    .e()
+    .e()
+)
+
+# Doesn't fit, default
+d2 = (
+    x.e().esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkfsdddd()  #
+)
+
+# Doesn't fit, fluent style
+d3 = (
+    x.e()  #
+    .esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk()
+    .esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk()
+)
+
+# Don't drop the bin op parentheses
+e1 = (1 + 2).w().t()
+e2 = (1 + 2)().w().t()
+e3 = (1 + 2)[1].w().t()
+
+# Treat preserved parentheses correctly
+f1 = (b().c()).d(1,)
+f2 = b().c().d(1,)
+f3 = (b).c().d(1,)
+f4 = (a)(b).c(1,)
+f5 = (a.b()).c(1,)
+
+# Indent in the parentheses without breaking
+g1 = (
+    queryset.distinct().order_by(field.name).values_list(field_name_flat_long_long=True)
+)
+
+# Fluent style in subexpressions
+if (
+    not a()
+    .b()
+    .cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc()
+):
+    pass
+h2 = (
+    bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
+    + ccccccccccccccccccccccccc()
+    .dddddddddddddddddddddd()
+    .eeeeeeeeee()
+    .ffffffffffffffffffffff()
+)
+
+# Parentheses aren't allowed on statement level, don't use fluent style here
+if True:
+    (alias).filter(content_typeold_content_type).update(
+        content_typenew_contesadfasfdant_type
+    )
+
+zero(
+    one,
+).two(
+    three,
+).four(
+    five,
+)
+
+
+```
+
+## Output
+```py
+# Test cases for call chains and optional parentheses, with and without fluent style
+
+raise OsError("") from a.aaaaa(
+    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
+).a(aaaa)
+
+raise OsError(
+    "sökdjffffsldkfjlhsakfjhalsökafhsöfdahsödfjösaaksjdllllllllllllll"
+) from a.aaaaa(
+    aksjdhflsakhdflkjsadlfajkslhfdkjsaldajlahflashdfljahlfksajlhfajfjfsaahflakjslhdfkjalhdskjfa
+).a(aaaa)
+
+a1 = Blog.objects.filter(entry__headline__contains="Lennon").filter(
+    entry__pub_date__year=2008
+)
+
+a2 = Blog.objects.filter(
+    entry__headline__contains="Lennon",
+).filter(
+    entry__pub_date__year=2008,
+)
+
+raise OsError("") from (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+)
+
+raise OsError("sökdjffffsldkfjlhsakfjhalsökafhsöfdahsödfjösaaksjdllllllllllllll") from (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+    .filter(
+        entry__pub_date__year=2008,
+    )
+)
+
+# Break only after calls and indexing
+b1 = (
+    session.query(models.Customer.id)
+    .filter(
+        models.Customer.account_id == account_id, models.Customer.email == email_address
+    )
+    .count()
+)
+
+b2 = (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    )
+    .limit_results[:10]
+    .filter(
+        entry__pub_date__month=10,
+    )
+)
+
+# Nested call chains
+c1 = (
+    Blog.objects.filter(
+        entry__headline__contains="Lennon",
+    ).filter(
+        entry__pub_date__year=2008,
+    )
+    + Blog.objects.filter(
+        entry__headline__contains="McCartney",
+    )
+    .limit_results[:10]
+    .filter(
+        entry__pub_date__year=2010,
+    )
+).all()
+
+# Test different cases with trailing end of line comments:
+# * fluent style, fits: no parentheses -> ignore the expand_parent
+# * fluent style, doesn't fit: break all soft line breaks
+# * default, fits: no parentheses
+# * default, doesn't fit: parentheses but no soft line breaks
+
+# Fits, either style
+d11 = x.e().e().e()  #
+d12 = x.e().e().e()  #
+d13 = (
+    x.e()  #
+    .e()
+    .e()
+)
+
+# Doesn't fit, default
+d2 = x.e().esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkfsdddd()  #
+
+# Doesn't fit, fluent style
+d3 = (
+    x.e()  #
+    .esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk()
+    .esadjkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk()
+)
+
+# Don't drop the bin op parentheses
+e1 = (1 + 2).w().t()
+e2 = (1 + 2)().w().t()
+e3 = (1 + 2)[1].w().t()
+
+# Treat preserved parentheses correctly
+f1 = (b().c()).d(
+    1,
+)
+f2 = (
+    b()
+    .c()
+    .d(
+        1,
+    )
+)
+f3 = (
+    (b)
+    .c()
+    .d(
+        1,
+    )
+)
+f4 = (a)(b).c(
+    1,
+)
+f5 = (a.b()).c(
+    1,
+)
+
+# Indent in the parentheses without breaking
+g1 = (
+    queryset.distinct().order_by(field.name).values_list(field_name_flat_long_long=True)
+)
+
+# Fluent style in subexpressions
+if (
+    not a()
+    .b()
+    .cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc()
+):
+    pass
+h2 = (
+    bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
+    + ccccccccccccccccccccccccc()
+    .dddddddddddddddddddddd()
+    .eeeeeeeeee()
+    .ffffffffffffffffffffff()
+)
+
+# Parentheses aren't allowed on statement level, don't use fluent style here
+if True:
+    (alias).filter(content_typeold_content_type).update(
+        content_typenew_contesadfasfdant_type
+    )
+
+zero(
+    one,
+).two(
+    three,
+).four(
+    five,
+)
+```
+
+
+

crates/ruff_python_formatter/tests/snapshots/format@expression__attribute.py.snap

@@ -143,10 +143,10 @@ a = Namespace()
 
 
 (
-    a.
+    a.  # trailing dot comment
     # comment
     # in between
-    b  # trailing dot comment  # trailing identifier comment
+    b  # trailing identifier comment
 )
 
 
@@ -157,10 +157,10 @@ a.aaaaaaaaaaaaaaaaaaaaa.lllllllllllllllllllllllllllloooooooooong.chaaaaaaaaaaaaa
 # chain if and only if we need them, that is if there are own line comments inside
 # the chain.
 x1 = (
-    a.b.
+    a.b.  # comment 2
     # comment 1
     # comment 3
-    c.d  # comment 2
+    c.d
 )
 
 x20 = a.b

crates/ruff_python_formatter/tests/snapshots/format@expression__call.py.snap

@@ -58,7 +58,7 @@ f(
     hey_this_is_a_very_long_call=1, it_has_funny_attributes_asdf_asdf=1, too_long_for_the_line=1, really=True
 )
 
-# TODO(konstin): Call chains/fluent interface (https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#call-chains)
+# Call chains/fluent interface (https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#call-chains)
 result = (
     session.query(models.Customer.id)
     .filter(
@@ -178,11 +178,16 @@ f(
     really=True,
 )
 
-# TODO(konstin): Call chains/fluent interface (https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#call-chains)
+# Call chains/fluent interface (https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#call-chains)
-result = session.query(models.Customer.id).filter(
+result = (
-    models.Customer.account_id == 10000,
+    session.query(models.Customer.id)
-    models.Customer.email == "user@example.org",
+    .filter(
-).order_by(models.Customer.id.asc()).all()
+        models.Customer.account_id == 10000,
+        models.Customer.email == "user@example.org",
+    )
+    .order_by(models.Customer.id.asc())
+    .all()
+)
 # TODO(konstin): Black has this special case for comment placement where everything stays in one line
 f("aaaaaaaa", "aaaaaaaa", "aaaaaaaa", "aaaaaaaa", "aaaaaaaa", "aaaaaaaa", "aaaaaaaa")