atrep-cli 0.3.1__tar.gz → 0.3.2__tar.gz

{atrep_cli-0.3.1 → atrep_cli-0.3.2}/Cargo.lock

@@ -60,7 +60,7 @@ dependencies = [
 
 [[package]]
 name = "atrep"
-version = "0.3.1"
+version = "0.3.2"
 dependencies = [
  "flate2",
  "serde",
@@ -73,17 +73,10 @@ dependencies = [
 
 [[package]]
 name = "atrep-cli"
-version = "0.3.1"
+version = "0.3.2"
 dependencies = [
  "atrep",
  "clap",
-]
-
-[[package]]
-name = "atrep-lsp"
-version = "0.3.1"
-dependencies = [
- "atrep",
  "lsp-server",
  "lsp-types",
  "serde",

{atrep_cli-0.3.1 → atrep_cli-0.3.2}/Cargo.toml

@@ -3,14 +3,14 @@ resolver = "2"
 members = ["crates/atrep", "crates/atrep-cli"]
 
 [workspace.package]
-version = "0.3.1"
+version = "0.3.2"
 edition = "2024"
 license = "MIT OR Apache-2.0"
 authors = ["Bojan Durickovic <dev@deyanovich.org>"]
 repository = "https://gitlab.com/atrep/atrep-rs"
 
 [workspace.dependencies]
-atrep = { path = "crates/atrep", version = "0.3.1" }
+atrep = { path = "crates/atrep", version = "0.3.2" }
 unicode-normalization = "0.1"
 sha2 = "0.10"
 ureq = "3"
@@ -19,3 +19,5 @@ flate2 = "1"
 serde = { version = "1", features = ["derive"] }
 serde_json = "1"
 clap = { version = "4", features = ["derive"] }
+lsp-server = "0.9"
+lsp-types = "0.97"

{atrep_cli-0.3.1 → atrep_cli-0.3.2}/PKG-INFO

@@ -1,13 +1,13 @@
 Metadata-Version: 2.4
 Name: atrep-cli
-Version: 0.3.1
+Version: 0.3.2
 Classifier: Development Status :: 4 - Beta
 Classifier: Intended Audience :: Developers
 Classifier: Programming Language :: Rust
 Classifier: Topic :: Text Processing :: Markup
 License-File: LICENSE-MIT
 License-File: LICENSE-APACHE
-Summary: The atrep command-line tool: check, respellings, outline, kanonizo, endo importers, morphisms, exo rendering, collation
+Summary: The atrep command-line tool (check, respellings, outline, kanonizo, endo importers, morphisms, exo rendering, collation) and the atrep-lsp language server
 Keywords: atrep,markup,semantic,document,dialect,cli
 Author: Bojan Đuričković
 License-Expression: MIT OR Apache-2.0
@@ -43,11 +43,15 @@ spec revision.
   classes), `dialektos` (`.lektos`/`.dia` definitions, local
   inheritance and import), `parser` (document to Dendron),
   `dendron` (AST + canonical serializer), `kanonizo`, `litosis`.
-- `crates/atrep-cli` — the `atrep` binary.
-- `crates/atrep-lsp` — the `atrep-lsp` language server:
-  diagnostics + semantic-token highlighting for `.atd`/`.atk`
-  documents and `.dia`/`.lektos` definitions (see its README
-  for editor configuration).
+- `crates/atrep-cli` — two binaries, one crate: the `atrep`
+  CLI, and the `atrep-lsp` language server (diagnostics +
+  semantic-token highlighting for `.atd`/`.atk` documents and
+  `.dia`/`.lektos` definitions, and — over the outline —
+  document symbols, folding, onym definition and references,
+  selection ranges, and sim hover for documents; see
+  `crates/atrep-cli/LSP.md` for editor configuration). Both
+  ship with `cargo install atrep-cli` and the PyPI `atrep-cli`
+  wheel.
 
 ## Usage
 

{atrep_cli-0.3.1/crates/atrep → atrep_cli-0.3.2}/README.md

@@ -26,11 +26,15 @@ spec revision.
   classes), `dialektos` (`.lektos`/`.dia` definitions, local
   inheritance and import), `parser` (document to Dendron),
   `dendron` (AST + canonical serializer), `kanonizo`, `litosis`.
-- `crates/atrep-cli` — the `atrep` binary.
-- `crates/atrep-lsp` — the `atrep-lsp` language server:
-  diagnostics + semantic-token highlighting for `.atd`/`.atk`
-  documents and `.dia`/`.lektos` definitions (see its README
-  for editor configuration).
+- `crates/atrep-cli` — two binaries, one crate: the `atrep`
+  CLI, and the `atrep-lsp` language server (diagnostics +
+  semantic-token highlighting for `.atd`/`.atk` documents and
+  `.dia`/`.lektos` definitions, and — over the outline —
+  document symbols, folding, onym definition and references,
+  selection ranges, and sim hover for documents; see
+  `crates/atrep-cli/LSP.md` for editor configuration). Both
+  ship with `cargo install atrep-cli` and the PyPI `atrep-cli`
+  wheel.
 
 ## Usage
 

{atrep_cli-0.3.1/crates/atrep-cli → atrep_cli-0.3.2/crates/atrep}/README.md

@@ -26,11 +26,15 @@ spec revision.
   classes), `dialektos` (`.lektos`/`.dia` definitions, local
   inheritance and import), `parser` (document to Dendron),
   `dendron` (AST + canonical serializer), `kanonizo`, `litosis`.
-- `crates/atrep-cli` — the `atrep` binary.
-- `crates/atrep-lsp` — the `atrep-lsp` language server:
-  diagnostics + semantic-token highlighting for `.atd`/`.atk`
-  documents and `.dia`/`.lektos` definitions (see its README
-  for editor configuration).
+- `crates/atrep-cli` — two binaries, one crate: the `atrep`
+  CLI, and the `atrep-lsp` language server (diagnostics +
+  semantic-token highlighting for `.atd`/`.atk` documents and
+  `.dia`/`.lektos` definitions, and — over the outline —
+  document symbols, folding, onym definition and references,
+  selection ranges, and sim hover for documents; see
+  `crates/atrep-cli/LSP.md` for editor configuration). Both
+  ship with `cargo install atrep-cli` and the PyPI `atrep-cli`
+  wheel.
 
 ## Usage
 

atrep_cli-0.3.2/crates/atrep-cli/Cargo.toml

@@ -0,0 +1,30 @@
+[package]
+name = "atrep-cli"
+description = "atrep command-line tool and language server (pilot implementation)"
+version.workspace = true
+edition.workspace = true
+license.workspace = true
+repository.workspace = true
+readme = "README.md"
+keywords = ["atrep", "markup", "semantic", "document", "cli"]
+categories = ["command-line-utilities", "text-processing"]
+
+# Two binaries, one crate, one release: the `atrep` CLI for
+# pipelines and the `atrep-lsp` language server for editors
+# (LSP.md). Both ride the crates.io publish and the PyPI wheel.
+[[bin]]
+name = "atrep"
+path = "src/main.rs"
+
+[[bin]]
+name = "atrep-lsp"
+path = "src/lsp/main.rs"
+
+[dependencies]
+atrep.workspace = true
+clap.workspace = true
+lsp-server.workspace = true
+lsp-types.workspace = true
+serde.workspace = true
+serde_json.workspace = true
+unicode-normalization.workspace = true

atrep_cli-0.3.2/crates/atrep-cli/LSP.md

@@ -0,0 +1,95 @@
+# atrep-lsp
+
+Language server for **atrep** documents (`.atd`/`.atk`) and
+dialektos definitions (`.dia`/`.lektos`), over the `atrep`
+engine. stdio transport. It is the second binary of the
+`atrep-cli` crate — a thin transport over the engine, released
+and distributed together with the `atrep` CLI (crates.io and
+the PyPI `atrep-cli` wheel both carry it).
+
+Surface:
+
+- **Diagnostics** — the buffer is parsed on open and on every
+  change (`atrep` `check_source`: documents and definitions
+  are routed on the declaration line); the first parse error is
+  published with its source location.
+- **Semantic tokens** (full-document) — lexical highlighting of
+  the declaration (`@@@!id@version`), sim and episim symbols,
+  monosim parameters (numeric taxis vs. onym), comments
+  (single- and multi-line, at any sigil-run depth), and escaped
+  sigils. When the document's dialektos resolves (locally beside
+  the file, or from the embedded std set — litogramma included),
+  symbols are matched against it with the parser's own
+  longest-match rule, and undefined sims render distinctly.
+  Alias-sigil (`\`) documents are handled.
+
+Structural features, for documents whose outline assembles (the
+engine's `atrep outline` data: block spans recorded by the
+parser, milestones, onyms, deixes — dialektos-generic, so they
+work for litogramma, lexigramma, at-docbook, or any dialektos
+that resolves):
+
+- **Document symbols** — the nested block tree (sim name plus
+  lemma; symbol, genoses, and onym in the detail), with each
+  block's milestones and standalone onym anchors as leaf
+  children, so a coordinate such as `steph:17a` is one symbol
+  search away.
+- **Folding ranges** — one per multi-line block; the opening
+  line stays visible.
+- **Go to definition** — on a `(onym)` group: the onym's
+  declaration (the block's episim line or the standalone
+  anchor). **References** — every deixis pointing at the onym,
+  plus the declaration when the client asks for it.
+- **Selection ranges** — the enclosing-block chain at the
+  cursor, innermost first: the structural text objects.
+- **Hover** on a sim symbol or episim — the sim's name, form
+  (endo-simmere, para-simmere, monosim), symbol pair, and the
+  definition's descriptions from the resolved dialektos.
+
+Not yet: atramento (the loose syntax compiles via `atrep endo`;
+its own LSP comes later), incremental sync, completions, rename.
+
+## Install
+
+Any of:
+
+```
+cargo install atrep-cli        # crates.io: atrep + atrep-lsp
+pip install atrep-cli          # PyPI wheel, same two binaries
+cargo build -p atrep-cli --release   # from a checkout
+```
+
+The binary is `atrep-lsp`; it speaks LSP on stdin/stdout.
+Editor package managers that install from PyPI or crates.io
+(mason.nvim, for one) can fetch it through either registry.
+
+## Editor configuration
+
+**Helix** (`languages.toml`):
+
+```toml
+[language-server.atrep]
+command = "atrep-lsp"
+
+[[language]]
+name = "atrep"
+scope = "source.atrep"
+file-types = ["atd", "atk", "dia", "lektos"]
+language-servers = ["atrep"]
+```
+
+**Neovim** (0.11+):
+
+```lua
+vim.filetype.add({ extension = { atd = "atrep", atk = "atrep",
+                                 dia = "atrep", lektos = "atrep" } })
+vim.lsp.config("atrep", {
+  cmd = { "atrep-lsp" },
+  filetypes = { "atrep" },
+})
+vim.lsp.enable("atrep")
+```
+
+**VS Code**: needs a thin client extension (planned); any generic
+LSP client extension pointed at `atrep-lsp` with the file types
+above works meanwhile.

{atrep_cli-0.3.1 → atrep_cli-0.3.2/crates/atrep-cli}/README.md

@@ -26,11 +26,15 @@ spec revision.
   classes), `dialektos` (`.lektos`/`.dia` definitions, local
   inheritance and import), `parser` (document to Dendron),
   `dendron` (AST + canonical serializer), `kanonizo`, `litosis`.
-- `crates/atrep-cli` — the `atrep` binary.
-- `crates/atrep-lsp` — the `atrep-lsp` language server:
-  diagnostics + semantic-token highlighting for `.atd`/`.atk`
-  documents and `.dia`/`.lektos` definitions (see its README
-  for editor configuration).
+- `crates/atrep-cli` — two binaries, one crate: the `atrep`
+  CLI, and the `atrep-lsp` language server (diagnostics +
+  semantic-token highlighting for `.atd`/`.atk` documents and
+  `.dia`/`.lektos` definitions, and — over the outline —
+  document symbols, folding, onym definition and references,
+  selection ranges, and sim hover for documents; see
+  `crates/atrep-cli/LSP.md` for editor configuration). Both
+  ship with `cargo install atrep-cli` and the PyPI `atrep-cli`
+  wheel.
 
 ## Usage
 

atrep_cli-0.3.2/crates/atrep-cli/src/lsp/analysis.rs

@@ -0,0 +1,102 @@
+//! Per-document analysis: diagnostics via atrep, dialektos
+//! resolution for the scanner.
+
+use std::path::{Path, PathBuf};
+
+use atrep::dialektos::{self, Dialektos};
+use lsp_types::{Diagnostic, DiagnosticSeverity, Position, Range};
+use unicode_normalization::UnicodeNormalization;
+
+/// Parse the buffer (document or definition, routed on the
+/// declaration) and map the error, if any, to LSP diagnostics.
+pub fn diagnostics(source: &str, path: &Path) -> Vec<Diagnostic> {
+    match atrep::check_source(source, path) {
+        Ok(_) => Vec::new(),
+        Err(e) => {
+            let range = e
+                .location
+                .as_ref()
+                .map(|loc| location_range(source, loc.line, loc.col))
+                .unwrap_or_else(|| Range::new(Position::new(0, 0), Position::new(0, 0)));
+            vec![Diagnostic {
+                range,
+                severity: Some(DiagnosticSeverity::ERROR),
+                source: Some("atrep".into()),
+                message: e.to_string(),
+                ..Default::default()
+            }]
+        }
+    }
+}
+
+/// Range for a 1-based (line, col) error location: from the column
+/// to the end of that line (the core reports positions, not spans).
+fn location_range(source: &str, line: usize, col: usize) -> Range {
+    let l = line.saturating_sub(1);
+    let text = source.lines().nth(l).unwrap_or("");
+    let mut start16 = 0u32;
+    let mut end16 = 0u32;
+    for (i, c) in text.chars().enumerate() {
+        if i < col.saturating_sub(1) {
+            start16 += c.len_utf16() as u32;
+        }
+        end16 += c.len_utf16() as u32;
+    }
+    if end16 <= start16 {
+        start16 = 0;
+    }
+    Range::new(
+        Position::new(l as u32, start16),
+        Position::new(l as u32, end16),
+    )
+}
+
+/// Resolve the buffer's declared dialektos (documents only;
+/// definitions highlight structurally). NFC-normalized like the
+/// parser input; resolution is local-to-the-file's-directory plus
+/// the embedded std set, exactly as atrep resolves it.
+pub fn resolve_dialektos(source: &str, path: &Path) -> Option<Dialektos> {
+    if atrep::is_definition_source(source) {
+        return None;
+    }
+    let id = atrep::scan::declared_id(source)?;
+    let normalized: String = id.nfc().collect();
+    let dir: PathBuf = path.parent().unwrap_or(Path::new(".")).to_path_buf();
+    dialektos::resolve(&dir, &normalized).ok()
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn clean_litogramma_document() {
+        let src = "@@@!litogramma\n\n@#(1) Title\ntext\n#@\n";
+        assert!(diagnostics(src, Path::new("t.atd")).is_empty());
+    }
+
+    #[test]
+    fn missing_declaration_is_flagged() {
+        let d = diagnostics("just prose\n", Path::new("t.atd"));
+        assert_eq!(d.len(), 1);
+        assert_eq!(d[0].range.start.line, 0);
+    }
+
+    #[test]
+    fn unresolvable_dialektos_is_flagged() {
+        let d = diagnostics("@@@!no-such-dialekt\n", Path::new("t.atd"));
+        assert_eq!(d.len(), 1);
+    }
+
+    #[test]
+    fn std_dialektos_resolves() {
+        let src = "@@@!litogramma\n\ntext\n";
+        assert!(resolve_dialektos(src, Path::new("t.atd")).is_some());
+    }
+
+    #[test]
+    fn definition_source_gets_no_dialektos() {
+        let src = "@@@!atrep\n\n@=== note\n";
+        assert!(resolve_dialektos(src, Path::new("x.dia")).is_none());
+    }
+}