add treehouse fix command that generates ulids for you

Cargo.lock

@@ -312,6 +312,12 @@ dependencies = [
  "termcolor",
 ]
 
+[[package]]
+name = "equivalent"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5443807d6dff69373d433ab9ef5378ad8df50ca6298caf15de6e52e24aaf54d5"
+
 [[package]]
 name = "errno"
 version = "0.3.2"
@@ -397,6 +403,17 @@ dependencies = [
  "version_check",
 ]
 
+[[package]]
+name = "getrandom"
+version = "0.2.10"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "be4136b2a15dd319360be1c07d9933517ccf0be8f16bf62a3bee4f0d618df427"
+dependencies = [
+ "cfg-if",
+ "libc",
+ "wasi",
+]
+
 [[package]]
 name = "gimli"
 version = "0.27.3"
@@ -417,6 +434,12 @@ dependencies = [
  "thiserror",
 ]
 
+[[package]]
+name = "hashbrown"
+version = "0.14.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2c6201b9ff9fd90a5a3bac2e56a830d0caa509576f0e503818ee82c181b3437a"
+
 [[package]]
 name = "heck"
 version = "0.4.1"
@@ -498,6 +521,16 @@ dependencies = [
  "want",
 ]
 
+[[package]]
+name = "indexmap"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d5477fe2230a79769d8dc68e0eabf5437907c0457a5614a9e8dddb67f65eb65d"
+dependencies = [
+ "equivalent",
+ "hashbrown",
+]
+
 [[package]]
 name = "is-terminal"
 version = "0.4.9"
@@ -721,6 +754,12 @@ version = "0.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184"
 
+[[package]]
+name = "ppv-lite86"
+version = "0.2.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5b40af805b3121feab8a3c29f04d8ad262fa8e0561883e7653e024ae4479e6de"
+
 [[package]]
 name = "proc-macro2"
 version = "1.0.66"
@@ -750,6 +789,36 @@ dependencies = [
  "proc-macro2",
 ]
 
+[[package]]
+name = "rand"
+version = "0.8.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "34af8d1a0e25924bc5b7c43c079c942339d8f0a8b57c39049bef581b46327404"
+dependencies = [
+ "libc",
+ "rand_chacha",
+ "rand_core",
+]
+
+[[package]]
+name = "rand_chacha"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e6c10a63a0fa32252be49d21e7709d4d4baf8d231c2dbce1eaa8141b9b127d88"
+dependencies = [
+ "ppv-lite86",
+ "rand_core",
+]
+
+[[package]]
+name = "rand_core"
+version = "0.6.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ec0be4795e2f6a28069bec0b5ff3e2ac9bafc99e6a9a7dc3547996c5c816922c"
+dependencies = [
+ "getrandom",
+]
+
 [[package]]
 name = "redox_syscall"
 version = "0.3.5"
@@ -1028,6 +1097,23 @@ dependencies = [
  "tokio",
 ]
 
+[[package]]
+name = "toml_datetime"
+version = "0.6.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7cda73e2f1397b1262d6dfdcef8aafae14d1de7748d66822d3bfeeb6d03e5e4b"
+
+[[package]]
+name = "toml_edit"
+version = "0.19.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f8123f27e969974a3dfba720fdb560be359f57b44302d280ba72e76a74480e8a"
+dependencies = [
+ "indexmap",
+ "toml_datetime",
+ "winnow",
+]
+
 [[package]]
 name = "tower"
 version = "0.4.13"
@@ -1129,11 +1215,14 @@ dependencies = [
  "handlebars",
  "log",
  "pulldown-cmark",
+ "rand",
  "serde",
  "tokio",
+ "toml_edit",
  "tower-http",
  "tower-livereload",
  "treehouse-format",
+ "ulid",
  "walkdir",
 ]
 
@@ -1162,6 +1251,15 @@ version = "0.1.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ed646292ffc8188ef8ea4d1e0e0150fb15a5c2e12ad9b8fc191ae7a8a7f3c4b9"
 
+[[package]]
+name = "ulid"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "13a3aaa69b04e5b66cc27309710a569ea23593612387d67daaf102e73aa974fd"
+dependencies = [
+ "rand",
+]
+
 [[package]]
 name = "unicase"
 version = "2.6.0"
@@ -1316,3 +1414,12 @@ name = "windows_x86_64_msvc"
 version = "0.48.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b6ade45bc8bf02ae2aa34a9d54ba660a1a58204da34ba793c00d83ca3730b5f1"
+
+[[package]]
+name = "winnow"
+version = "0.5.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d09770118a7eb1ccaf4a594a221334119a44a814fcb0d31c5b85e83e97227a97"
+dependencies = [
+ "memchr",
+]

content/index.tree

@@ -2,6 +2,7 @@
   id = "treehouse"
 - # treehouse
 
+    % # example
     - welcome to the treehouse!
 
         - treehouse is a brand new static website generator, inspired by the likes of Jekyll and Hugo, but offering a writing experience more close to Logseq

content/secret.tree

@@ -1 +1,2 @@
+% id = "01H87RB4MB7RM6V2K1DR8XBV8N"
 - He is behind the tree.

content/ulid-test.tree

@@ -0,0 +1,35 @@
+% always_expanded = true
+  id = "treehouse"  # this one has a custom id
+- root
+
+    % id = "01H87R133DNH9JRGS4ZWN586M4"
+      # this one is missing an ULID
+    - child
+
+        % id = "01H87R133D75W92HRESGAXBTC5"
+
+            # this one is formatted a bit weirdly
+        - another child
+
+            % id = "custom-id"
+            - intertwined with one that has a custom ID
+
+            % id = "01H87R133D4MXM01VGTEXHPVAR"
+
+              # this is on a separate line and lacks indentation
+            - and another child
+
+    % always_expanded = true
+      id = "01H87R133D9GDXRD7WA14ECDES"
+    - this one has a custom attribute but not an ID
+
+        % id = "01H87R7PVRJ7FTWKC1SXHGQXDC"
+        - hellow
+
+    % id = "01H87R133DHQPZWJVTMWPKVGP0"
+    - <!-- and that one is plain empty, including the content. or conten't, for that matter. -->
+
+% id = "01H87R133DRJ7VKEMBSJKHYRJN"
+
+    # comment
+-

crates/treehouse-format/src/pull.rs

@@ -108,6 +108,7 @@ impl<'a> Parser<'a> {
             self.advance();
             let after_percent = self.position;
             self.eat_indented_lines_until(indent_level, |c| c == '-' || c == '+')?;
+            self.eat_as_long_as(' ');
             let end = self.position;
             Some(Attributes {
                 percent: start..after_percent,

crates/treehouse/Cargo.toml

@@ -18,8 +18,11 @@ handlebars = "4.3.7"
 pulldown-cmark = { version = "0.9.3", default-features = false }
 serde = { version = "1.0.183", features = ["derive"] }
 tokio = { version = "1.32.0", features = ["full"] }
+toml_edit = "0.19.14"
 tower-http = { version = "0.4.3", features = ["fs"] }
 tower-livereload = "0.8.0"
 walkdir = "2.3.3"
+ulid = "1.0.0"
+rand = "0.8.5"
 
 

crates/treehouse/src/cli/diagnostics.rs

@@ -0,0 +1,42 @@
+use anyhow::Context;
+use codespan_reporting::{
+    diagnostic::Diagnostic,
+    files::SimpleFiles,
+    term::termcolor::{ColorChoice, StandardStream},
+};
+
+pub type Files = SimpleFiles<String, String>;
+pub type FileId = <Files as codespan_reporting::files::Files<'static>>::FileId;
+
+pub struct Diagnosis {
+    pub files: Files,
+    pub diagnostics: Vec<Diagnostic<FileId>>,
+}
+
+impl Diagnosis {
+    pub fn new() -> Self {
+        Self {
+            files: Files::new(),
+            diagnostics: vec![],
+        }
+    }
+
+    /// Get the source code of a file, assuming it was previously registered.
+    pub fn get_source(&self, file_id: FileId) -> &str {
+        self.files
+            .get(file_id)
+            .expect("file should have been registered previously")
+            .source()
+    }
+
+    pub fn report(&self) -> anyhow::Result<()> {
+        let writer = StandardStream::stderr(ColorChoice::Auto);
+        let config = codespan_reporting::term::Config::default();
+        for diagnostic in &self.diagnostics {
+            codespan_reporting::term::emit(&mut writer.lock(), &config, &self.files, diagnostic)
+                .context("could not emit diagnostic")?;
+        }
+
+        Ok(())
+    }
+}

crates/treehouse/src/cli/fix.rs

@@ -0,0 +1,150 @@
+use std::ops::Range;
+
+use anyhow::Context;
+use treehouse_format::ast::Branch;
+
+use super::{
+    diagnostics::{Diagnosis, FileId},
+    parse::{self, parse_toml_with_diagnostics, parse_tree_with_diagnostics},
+    FixArgs,
+};
+
+struct Fix {
+    range: Range<usize>,
+    replacement: String,
+}
+
+#[derive(Default)]
+struct State {
+    fixes: Vec<Fix>,
+}
+
+fn dfs_fix_branch(diagnosis: &mut Diagnosis, file_id: FileId, state: &mut State, branch: &Branch) {
+    let mut rng = rand::thread_rng();
+    let ulid = ulid::Generator::new()
+        .generate_with_source(&mut rng)
+        .expect("failed to generate ulid for block"); // (wtf moment. you know how big the 80-bit combination space is?)
+
+    let indent = " ".repeat(branch.indent_level);
+    if let Some(attributes) = branch.attributes.clone() {
+        // Scenario: Attributes need to be parsed as TOML and the id attribute has to be added into
+        // the top-level table. Then we also need to pretty-print everything to match the right
+        // indentation level.
+        if let Ok(mut toml) =
+            parse_toml_with_diagnostics(diagnosis, file_id, attributes.data.clone())
+        {
+            if !toml.contains_key("id") {
+                toml["id"] = toml_edit::value(ulid.to_string());
+                toml.key_decor_mut("id")
+                    .unwrap()
+                    .set_prefix(" ".repeat(branch.indent_level + 2));
+            }
+            let mut toml_string = toml.to_string();
+
+            // This is incredibly janky and barely works.
+            let leading_spaces: usize = toml_string.chars().take_while(|&c| c == ' ').count();
+            match leading_spaces {
+                0 => toml_string.insert(0, ' '),
+                1 => (),
+                _ => toml_string.replace_range(0..leading_spaces - 1, ""),
+            }
+
+            let toml_string = fix_indent_in_generated_toml(&toml_string, branch.indent_level);
+
+            state.fixes.push(Fix {
+                range: attributes.data.clone(),
+                replacement: toml_string,
+            })
+        }
+    } else {
+        // Scenario: No attributes at all.
+        // In this case we can do a fast path where we generate the `% id = "whatever"` string
+        // directly, not going through toml_edit.
+        state.fixes.push(Fix {
+            range: branch.kind_span.start..branch.kind_span.start,
+            replacement: format!("% id = \"{ulid}\"\n{indent}"),
+        });
+    }
+
+    // Then we fix child branches.
+    for child in &branch.children {
+        dfs_fix_branch(diagnosis, file_id, state, child);
+    }
+}
+
+fn fix_indent_in_generated_toml(toml: &str, min_indent_level: usize) -> String {
+    let toml = toml.trim_end();
+
+    let mut result = String::with_capacity(toml.len());
+
+    for (i, line) in toml.lines().enumerate() {
+        if line.is_empty() {
+            result.push('\n');
+        } else {
+            let desired_line_indent_level = if i == 0 { 1 } else { min_indent_level + 2 };
+            let leading_spaces: usize = line.chars().take_while(|&c| c == ' ').count();
+            let needed_indentation = desired_line_indent_level.saturating_sub(leading_spaces);
+            for _ in 0..needed_indentation {
+                result.push(' ');
+            }
+            result.push_str(line);
+            result.push('\n');
+        }
+    }
+
+    for _ in 0..min_indent_level {
+        result.push(' ');
+    }
+
+    result
+}
+
+pub fn fix_file(
+    diagnosis: &mut Diagnosis,
+    file_id: FileId,
+) -> Result<String, parse::ErrorsEmitted> {
+    parse_tree_with_diagnostics(diagnosis, file_id).map(|roots| {
+        let mut source = diagnosis.get_source(file_id).to_owned();
+        let mut state = State::default();
+
+        for branch in &roots.branches {
+            dfs_fix_branch(diagnosis, file_id, &mut state, branch);
+        }
+
+        // Doing a depth-first search of the branches yields fixes from the beginning of the file
+        // to its end. The most efficient way to apply all the fixes then is to reverse their order,
+        // which lets us modify the source string in place because the fix ranges always stay
+        // correct.
+        for fix in state.fixes.iter().rev() {
+            source.replace_range(fix.range.clone(), &fix.replacement);
+        }
+
+        source
+    })
+}
+
+pub fn fix_file_cli(fix_args: FixArgs) -> anyhow::Result<()> {
+    let utf8_filename = fix_args.file.to_string_lossy().into_owned();
+    let file = std::fs::read_to_string(&fix_args.file).context("cannot read file to fix")?;
+
+    let mut diagnosis = Diagnosis::new();
+    let file_id = diagnosis.files.add(utf8_filename, file);
+
+    if let Ok(fixed) = fix_file(&mut diagnosis, file_id) {
+        if fix_args.apply {
+            // Try to write the backup first. If writing that fails, bail out without overwriting
+            // the source file.
+            if let Some(backup_path) = fix_args.backup {
+                std::fs::write(backup_path, diagnosis.get_source(file_id))
+                    .context("cannot write backup; original file will not be overwritten")?;
+            }
+            std::fs::write(&fix_args.file, fixed).context("cannot overwrite original file")?;
+        } else {
+            println!("{fixed}");
+        }
+    } else {
+        diagnosis.report()?;
+    }
+
+    Ok(())
+}

crates/treehouse/src/cli/mod.rs

@@ -1,5 +1,10 @@
+pub mod diagnostics;
+pub mod fix;
+mod parse;
 pub mod regenerate;
 
+use std::path::PathBuf;
+
 use clap::{Args, Parser, Subcommand};
 
 #[derive(Parser)]
@@ -12,6 +17,9 @@ pub struct ProgramArgs {
 pub enum Command {
     /// Regenerate the website.
     Regenerate(#[clap(flatten)] RegenerateArgs),
+
+    /// Populate missing metadata in blocks.
+    Fix(#[clap(flatten)] FixArgs),
 }
 
 #[derive(Args)]
@@ -20,3 +28,19 @@ pub struct RegenerateArgs {
     #[clap(short, long)]
     pub serve: bool,
 }
+
+#[derive(Args)]
+pub struct FixArgs {
+    /// Which file to fix. The fixed file will be printed into stdout so that you have a chance to
+    /// see the changes.
+    pub file: PathBuf,
+
+    /// If you're happy with the suggested changes, specifying this will apply them to the file
+    /// (overwrite it in place.)
+    #[clap(long)]
+    pub apply: bool,
+
+    /// Write the previous version back to the specified path.
+    #[clap(long)]
+    pub backup: Option<PathBuf>,
+}

crates/treehouse/src/cli/parse.rs

@@ -0,0 +1,57 @@
+use std::{ops::Range, str::FromStr};
+
+use codespan_reporting::diagnostic::{Diagnostic, Label, LabelStyle, Severity};
+use treehouse_format::ast::Roots;
+
+use super::diagnostics::{Diagnosis, FileId};
+
+pub struct ErrorsEmitted;
+
+pub fn parse_tree_with_diagnostics(
+    diagnosis: &mut Diagnosis,
+    file_id: FileId,
+) -> Result<Roots, ErrorsEmitted> {
+    let input = diagnosis.get_source(file_id);
+    Roots::parse(&mut treehouse_format::pull::Parser { input, position: 0 }).map_err(|error| {
+        diagnosis.diagnostics.push(Diagnostic {
+            severity: Severity::Error,
+            code: Some("tree".into()),
+            message: error.kind.to_string(),
+            labels: vec![Label {
+                style: LabelStyle::Primary,
+                file_id,
+                range: error.range,
+                message: String::new(),
+            }],
+            notes: vec![],
+        });
+        ErrorsEmitted
+    })
+}
+
+pub fn parse_toml_with_diagnostics(
+    diagnosis: &mut Diagnosis,
+    file_id: FileId,
+    range: Range<usize>,
+) -> Result<toml_edit::Document, ErrorsEmitted> {
+    let input = &diagnosis.get_source(file_id)[range.clone()];
+    toml_edit::Document::from_str(input).map_err(|error| {
+        diagnosis.diagnostics.push(Diagnostic {
+            severity: Severity::Error,
+            code: Some("toml".into()),
+            message: error.message().to_owned(),
+            labels: error
+                .span()
+                .map(|span| Label {
+                    style: LabelStyle::Primary,
+                    file_id,
+                    range: range.start + span.start..range.start + span.end,
+                    message: String::new(),
+                })
+                .into_iter()
+                .collect(),
+            notes: vec![],
+        });
+        ErrorsEmitted
+    })
+}

crates/treehouse/src/cli/regenerate.rs

@@ -7,8 +7,7 @@ use anyhow::{bail, Context};
 use axum::Router;
 use codespan_reporting::{
     diagnostic::{Diagnostic, Label, LabelStyle, Severity},
-    files::{Files as _, SimpleFiles},
+    files::Files as _,
-    term::termcolor::{ColorChoice, StandardStream},
 };
 use copy_dir::copy_dir;
 use handlebars::Handlebars;
@@ -16,24 +15,17 @@ use log::{debug, info};
 use serde::Serialize;
 use tower_http::services::ServeDir;
 use tower_livereload::LiveReloadLayer;
-use treehouse_format::ast::Roots;
 use walkdir::WalkDir;
 
-use crate::html::tree::branches_to_html;
+use crate::{cli::parse::parse_tree_with_diagnostics, html::tree::branches_to_html};
+
+use super::diagnostics::{Diagnosis, FileId};
 
 #[derive(Default)]
 struct Generator {
     tree_files: Vec<PathBuf>,
 }
 
-type Files = SimpleFiles<String, String>;
-type FileId = <Files as codespan_reporting::files::Files<'static>>::FileId;
-
-pub struct Diagnosis {
-    pub files: Files,
-    pub diagnostics: Vec<Diagnostic<FileId>>,
-}
-
 impl Generator {
     fn add_directory_rec(&mut self, directory: &Path) -> anyhow::Result<()> {
         for entry in WalkDir::new(directory) {
@@ -108,10 +100,7 @@ impl Generator {
     }
 
     fn generate_all_files(&self, dirs: &Dirs<'_>) -> anyhow::Result<Diagnosis> {
-        let mut diagnosis = Diagnosis {
+        let mut diagnosis = Diagnosis::new();
-            files: Files::new(),
-            diagnostics: vec![],
-        };
 
         let mut handlebars = Handlebars::new();
         let tree_template = Self::register_template(
@@ -145,20 +134,10 @@ impl Generator {
                 }
             };
             let file_id = diagnosis.files.add(utf8_filename.into_owned(), source);
-            let source = diagnosis
-                .files
-                .get(file_id)
-                .expect("file was just added to the list")
-                .source();
 
-            let parse_result = Roots::parse(&mut treehouse_format::pull::Parser {
+            if let Ok(roots) = parse_tree_with_diagnostics(&mut diagnosis, file_id) {
-                input: source,
-                position: 0,
-            });
-
-            match parse_result {
-                Ok(roots) => {
                 let mut tree = String::new();
+                let source = diagnosis.get_source(file_id);
                 branches_to_html(&mut tree, &roots.branches, source);
 
                 let template_data = TemplateData { tree };
@@ -178,19 +157,6 @@ impl Generator {
 
                 std::fs::write(target_path, templated_html)?;
             }
-                Err(error) => diagnosis.diagnostics.push(Diagnostic {
-                    severity: Severity::Error,
-                    code: Some("tree".into()),
-                    message: error.kind.to_string(),
-                    labels: vec![Label {
-                        style: LabelStyle::Primary,
-                        file_id,
-                        range: error.range,
-                        message: String::new(),
-                    }],
-                    notes: vec![],
-                }),
-            }
         }
 
         Ok(diagnosis)
@@ -223,12 +189,7 @@ pub fn regenerate(dirs: &Dirs<'_>) -> anyhow::Result<()> {
     generator.add_directory_rec(dirs.content_dir)?;
     let diagnosis = generator.generate_all_files(dirs)?;
 
-    let writer = StandardStream::stderr(ColorChoice::Auto);
+    diagnosis.report()?;
-    let config = codespan_reporting::term::Config::default();
-    for diagnostic in &diagnosis.diagnostics {
-        codespan_reporting::term::emit(&mut writer.lock(), &config, &diagnosis.files, diagnostic)
-            .context("could not emit diagnostic")?;
-    }
 
     Ok(())
 }

crates/treehouse/src/main.rs

@@ -2,6 +2,7 @@ use std::path::Path;
 
 use clap::Parser;
 use cli::{
+    fix::fix_file_cli,
     regenerate::{self, regenerate_or_report_error, Dirs},
     Command, ProgramArgs,
 };
@@ -32,6 +33,8 @@ async fn fallible_main() -> anyhow::Result<()> {
                 regenerate::web_server().await?;
             }
         }
+
+        Command::Fix(fix_args) => fix_file_cli(fix_args)?,
     }
 
     Ok(())