From 7ae7e42eb1eb981483cc4183368bec4932b8f1c2 Mon Sep 17 00:00:00 2001
From: Akshay <nerdy@peppe.rs>
Date: Sun, 14 Jul 2024 10:16:15 +0100
Subject: add trawk cli

---
 Cargo.toml    |  16 ++++-
 src/ast.rs    |  17 +----
 src/eval.rs   |  55 +++++++++++++--
 src/lib.rs    |   3 +-
 src/main.rs   | 219 +++++++++++++++++++++++++++++++++++++++++++++++-----------
 src/parser.rs |  58 +++++++++++++++-
 6 files changed, 297 insertions(+), 71 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 147096b..570ec22 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -11,6 +11,20 @@ regex = "1.3"
 thiserror = "1.0.61"
 serde = "1.0.204"
 nom = "7.1.3"
-tree-sitter-python = "0.20"
+
+tree-sitter-md = {version = "0.1", optional = true}
+tree-sitter-typescript = {version = "0.20.1", optional = true}
+tree-sitter-javascript = {version = "0.20.0", optional = true}
+tree-sitter-python = {version = "0.20.4", optional = true}
+tree-sitter-rust = {version = "0.20.3", optional = true}
+argh = "0.1.12"
+
+[features]
+default = ["md", "typescript", "javascript", "rust", "python"]
+md = ["tree-sitter-md"]
+typescript = ["tree-sitter-typescript"]
+javascript = ["tree-sitter-javascript"]
+python = ["tree-sitter-python"]
+rust = ["tree-sitter-rust"]
 
 [dev-dependencies]
diff --git a/src/ast.rs b/src/ast.rs
index 07b5c39..96fe8ab 100644
--- a/src/ast.rs
+++ b/src/ast.rs
@@ -59,6 +59,7 @@ pub enum Statement {
 #[derive(Debug, Eq, PartialEq, Clone)]
 pub enum Expr {
     Node,
+    FieldAccess(Vec<Identifier>),
     Unit,
     Lit(Literal),
     Ident(Identifier),
@@ -71,22 +72,6 @@ pub enum Expr {
 }
 
 impl Expr {
-    pub fn int(int: i128) -> Expr {
-        Self::Lit(Literal::Int(int))
-    }
-
-    pub fn str(s: &str) -> Expr {
-        Self::Lit(Literal::Str(s.to_owned()))
-    }
-
-    pub const fn false_() -> Expr {
-        Self::Lit(Literal::Bool(false))
-    }
-
-    pub const fn true_() -> Expr {
-        Self::Lit(Literal::Bool(true))
-    }
-
     pub fn boxed(self) -> Box<Expr> {
         Box::new(self)
     }
diff --git a/src/eval.rs b/src/eval.rs
index 859979d..e13cec3 100644
--- a/src/eval.rs
+++ b/src/eval.rs
@@ -39,6 +39,7 @@ pub enum Value {
     String(String),
     Boolean(bool),
     Node,
+    FieldAccess(Vec<String>),
 }
 
 impl Value {
@@ -49,6 +50,7 @@ impl Value {
             Self::String(_) => ast::Type::String,
             Self::Boolean(_) => ast::Type::Boolean,
             Self::Node => ast::Type::Node,
+            Self::FieldAccess(_) => ast::Type::Node,
         }
     }
 
@@ -241,6 +243,7 @@ impl fmt::Display for Value {
             Self::String(s) => write!(f, "{s}"),
             Self::Boolean(b) => write!(f, "{b}"),
             Self::Node => write!(f, "<node>"),
+            Self::FieldAccess(items) => write!(f, "<node>.{}", items.join(".")),
         }
     }
 }
@@ -317,7 +320,7 @@ pub enum Error {
     CurrentNodeNotPresent,
 }
 
-type Result = std::result::Result<Value, Error>;
+pub type Result = std::result::Result<Value, Error>;
 
 pub struct Context<'a> {
     variables: HashMap<ast::Identifier, Variable>,
@@ -385,6 +388,7 @@ impl<'a> Context<'a> {
             ast::Expr::IfExpr(if_expr) => self.eval_if(if_expr),
             ast::Expr::Block(block) => self.eval_block(block),
             ast::Expr::Node => Ok(Value::Node),
+            ast::Expr::FieldAccess(items) => Ok(Value::FieldAccess(items.to_owned())),
         }
     }
 
@@ -544,12 +548,33 @@ impl<'a> Context<'a> {
                 }
                 Ok(Value::Unit)
             }
-            ("text", [arg]) if self.eval_expr(arg)? == Value::Node => {
-                let node = self
-                    .cursor
-                    .as_ref()
-                    .ok_or(Error::CurrentNodeNotPresent)?
-                    .node();
+            ("text", [arg]) => {
+                let node = match self.eval_expr(arg)? {
+                    Value::Node => self
+                        .cursor
+                        .as_ref()
+                        .ok_or(Error::CurrentNodeNotPresent)?
+                        .node(),
+                    Value::FieldAccess(fields) => {
+                        let mut node = self
+                            .cursor
+                            .as_ref()
+                            .ok_or(Error::CurrentNodeNotPresent)?
+                            .node();
+                        for field in &fields {
+                            node = node
+                                .child_by_field_name(field.as_bytes())
+                                .ok_or_else(|| Error::FailedLookup(field.to_owned()))?;
+                        }
+                        node
+                    }
+                    v => {
+                        return Err(Error::TypeMismatch {
+                            expected: ast::Type::Node,
+                            got: v.ty(),
+                        })
+                    }
+                };
                 let text = node
                     .utf8_text(self.input_src.as_ref().unwrap().as_bytes())
                     .unwrap();
@@ -629,6 +654,22 @@ impl<'a> Context<'a> {
     }
 }
 
+pub fn evaluate(file: &str, program: &str, language: tree_sitter::Language) -> Result {
+    let mut parser = tree_sitter::Parser::new();
+    let _ = parser.set_language(language);
+
+    let tree = parser.parse(file, None).unwrap();
+    let cursor = tree.walk();
+
+    let program = ast::Program::new().from_str(program).unwrap();
+    let mut ctx = Context::new(tree_sitter_md::language())
+        .with_input(file.to_owned())
+        .with_cursor(cursor)
+        .with_program(program)?;
+
+    ctx.eval()
+}
+
 #[cfg(test)]
 mod test {
     use super::*;
diff --git a/src/lib.rs b/src/lib.rs
index 8780b74..fca5dd5 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -3,5 +3,4 @@ mod eval;
 mod parser;
 mod string;
 
-pub use ast::Program;
-pub use eval::Context;
+pub use eval::evaluate;
diff --git a/src/main.rs b/src/main.rs
index 09a15ef..6196a32 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -1,47 +1,182 @@
-use trawk::{Context, Program};
+/// TBSP: tree-based source processor
+#[derive(argh::FromArgs)]
+struct Cli {
+    /// read the TBSP program source from a file
+    #[argh(option, short = 'f')]
+    program_file: std::path::PathBuf,
+
+    /// set the language that the file is written in
+    #[argh(option, short = 'l')]
+    language: String,
+
+    /// input file to process
+    #[argh(positional)]
+    file: Option<std::path::PathBuf>,
+}
 
 fn main() {
-    let src = r#"
-bar = 0
-def foo():
-    baz = 5
-        "#
-    .to_owned();
-
-    let program = Program::new()
-        .from_str(
-            r#"
-    BEGIN {
-        bool in_def = false;
-    }
-    pre function_definition {
-        in_def = true;
+    let cli: Cli = argh::from_env();
+
+    let program = std::fs::read_to_string(&cli.program_file).unwrap_or_else(|e| {
+        eprintln!(
+            "failed to read program-file from `{}`: {e}",
+            cli.program_file.display()
+        );
+        std::process::exit(-1);
+    });
+
+    let language = match cli.language.as_str() {
+        "md" => tree_sitter_md::language(),
+        "typescript" => tree_sitter_typescript::language_typescript(),
+        "javascript" => tree_sitter_javascript::language(),
+        "python" => tree_sitter_python::language(),
+        "rust" => tree_sitter_rust::language(),
+        lang => {
+            eprintln!("unknown language `{lang}`");
+            std::process::exit(-1);
+        }
+    };
+
+    let file = cli
+        .file
+        .map(std::fs::read_to_string)
+        .unwrap_or_else(try_consume_stdin)
+        .unwrap_or_else(|e| {
+            eprintln!("{e}");
+            std::process::exit(-1)
+        });
+
+    trawk::evaluate(&file, &program, language).unwrap_or_else(|e| {
+        eprintln!("{e:?}");
+        std::process::exit(-1);
+    });
+}
+
+fn try_consume_stdin() -> std::io::Result<String> {
+    let mut buffer = String::new();
+    let mut lock = std::io::stdin().lock();
+
+    while let Ok(n) = std::io::Read::read_to_string(&mut lock, &mut buffer) {
+        if n == 0 {
+            break;
+        }
     }
-    post function_definition {
-        in_def = false;
+
+    if buffer.is_empty() {
+        Err(std::io::Error::other("empty stdin"))
+    } else {
+        Ok(buffer)
     }
-    pre identifier {
-        if (in_def) {
-            print(text(node));
-            print(" ");
-            print("in def\n");
-        } else {
-        };
-    }"#,
-        )
-        .unwrap();
-
-    let mut parser = tree_sitter::Parser::new();
-    let _ = parser.set_language(tree_sitter_python::language());
-
-    let tree = parser.parse(&src, None).unwrap();
-    let cursor = tree.walk();
-
-    let mut ctx = Context::new(tree_sitter_python::language())
-        .with_input(src)
-        .with_cursor(cursor)
-        .with_program(program)
-        .unwrap();
-
-    let _ = ctx.eval();
 }
+
+// fn main() {
+//     let src = r#"
+// # foo1
+//
+// bar
+//
+// ## foo1.1
+//
+// bar baz
+//
+// # foo2
+//
+// bar baz
+//
+// ```
+// fn main() {
+// }
+// ```
+//
+// - foo
+// - bar
+// - baz
+//
+//         "#
+//     .to_owned();
+//
+//     let program = Program::new()
+//         .from_str(
+//             r#"
+//     BEGIN {
+//         int depth = 0;
+//
+//         print("<html>\n");
+//         print("<body>\n");
+//     }
+//
+//     enter section {
+//         depth += 1;
+//     }
+//     leave section {
+//         depth -= 1;
+//     }
+//
+//     enter atx_heading {
+//         print("<h");
+//         print(depth);
+//         print(">");
+//     }
+//     leave atx_heading {
+//         print("</h");
+//         print(depth);
+//         print(">\n");
+//     }
+//
+//     enter paragraph {
+//         print("<p>");
+//     }
+//     leave paragraph {
+//         print("</p>\n");
+//     }
+//
+//     enter list {
+//         print("<ol>");
+//     }
+//     leave list {
+//         print("</ol>\n");
+//     }
+//
+//     enter list_item {
+//         print("<li>");
+//     }
+//     leave list_item {
+//         print("</li>\n");
+//     }
+//
+//     enter fenced_code_block {
+//         print("<pre>");
+//     }
+//     leave fenced_code_block {
+//         print("</pre>\n");
+//     }
+//
+//     enter inline {
+//         print(text(node));
+//     }
+//     enter code_fence_content {
+//         print(text(node));
+//     }
+//
+//     END {
+//         print("</body>\n");
+//         print("</html>\n");
+//     }
+//     "#,
+//         )
+//         .unwrap();
+//
+//     let mut parser = tree_sitter::Parser::new();
+//     let _ = parser.set_language(&tree_sitter_md::language());
+//
+//     let tree = parser.parse(&src, None).unwrap();
+//     let cursor = tree.walk();
+//
+//     let mut ctx = Context::new(tree_sitter_md::language())
+//         .with_input(src)
+//         .with_cursor(cursor)
+//         .with_program(program)
+//         .unwrap();
+//
+//     let _ = ctx.eval();
+// }
diff --git a/src/parser.rs b/src/parser.rs
index 3a020dc..d705a11 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -1,10 +1,10 @@
 use nom::{
     branch::alt,
-    bytes::complete::tag,
+    bytes::complete::{is_not, tag},
     character::complete::{alpha1, alphanumeric1, char, multispace0, multispace1, one_of},
     combinator::{map, opt, recognize, value},
     error::ParseError,
-    multi::{many0, many0_count, many1, separated_list0},
+    multi::{many0, many0_count, many1, separated_list0, separated_list1},
     sequence::{delimited, pair, preceded, terminated, tuple},
     IResult, Parser,
 };
@@ -21,6 +21,11 @@ where
     delimited(multispace0, inner, multispace0)
 }
 
+// TODO use this
+fn _parse_comment<'a>(i: &'a str) -> IResult<&'a str, ()> {
+    value((), pair(tag("//"), is_not("\n\r")))(i)
+}
+
 fn parse_unit<'a>(i: &'a str) -> IResult<&'a str, ()> {
     let open = char('(');
     let close = char(')');
@@ -169,8 +174,16 @@ fn parse_mul<'a>(i: &'a str) -> IResult<&'a str, Expr> {
     alt((recursive, base)).parse(i)
 }
 
+fn parse_field_access<'a>(i: &'a str) -> IResult<&'a str, Vec<Identifier>> {
+    let node = tag("node");
+    let dot = ws(char('.'));
+    let fields = separated_list1(ws(char('.')), map(parse_name, str::to_owned));
+    map(tuple((node, dot, fields)), |(_, _, fields)| fields)(i)
+}
+
 fn parse_atom<'a>(i: &'a str) -> IResult<&'a str, Expr> {
     let inner = alt((
+        map(parse_field_access, Expr::FieldAccess),
         map(tag("node"), |_| Expr::Node),
         map(parse_block, Expr::Block),
         map(parse_if, Expr::IfExpr),
@@ -337,7 +350,7 @@ fn parse_pattern<'a>(i: &str) -> IResult<&str, Pattern> {
         tuple((parse_modifier, multispace0, parse_ident)),
         |(modifier, _, kind)| Pattern::Node(NodePattern { modifier, kind }),
     );
-    alt((begin, end, node)).parse(i)
+    ws(alt((begin, end, node))).parse(i)
 }
 
 pub fn parse_stanza<'a>(i: &str) -> IResult<&str, Stanza> {
@@ -358,6 +371,25 @@ pub fn parse_file(i: &str) -> IResult<&str, Vec<Stanza>> {
 mod test {
     use super::*;
 
+    // test helpers
+    impl Expr {
+        pub fn int(int: i128) -> Expr {
+            Self::Lit(Literal::Int(int))
+        }
+
+        pub fn str(s: &str) -> Expr {
+            Self::Lit(Literal::Str(s.to_owned()))
+        }
+
+        pub const fn false_() -> Expr {
+            Self::Lit(Literal::Bool(false))
+        }
+
+        pub const fn true_() -> Expr {
+            Self::Lit(Literal::Bool(true))
+        }
+    }
+
     #[test]
     fn test_parse_unit() {
         assert_eq!(parse_unit("()"), Ok(("", ())))
@@ -534,6 +566,26 @@ mod test {
         );
     }
 
+    #[test]
+    fn test_parse_node() {
+        assert_eq!(parse_expr(r#" node "#), Ok(("", Expr::Node)));
+        assert_eq!(
+            parse_expr(r#" node.foo "#),
+            Ok(("", Expr::FieldAccess(vec!["foo".to_owned()])))
+        );
+        assert_eq!(
+            parse_expr(
+                r#" node
+                .foo
+                .bar"#
+            ),
+            Ok((
+                "",
+                Expr::FieldAccess(vec!["foo".to_owned(), "bar".to_owned()])
+            ))
+        );
+    }
+
     #[test]
     fn test_parse_if() {
         assert_eq!(
-- 
cgit v1.2.3