chore: Update readme, add more languages (#7)

Author: Pringled

README.md

@@ -1,4 +1,16 @@
-# semble-grammars
+<h1 align="center">
+  semble-grammars<br/>
+  <sub>Prebuilt tree-sitter grammars for Semble</sub>
+</h1>
+
+<div align="center">
+  <h2>
+    <a href="https://pypi.org/project/semble-grammars/"><img src="https://img.shields.io/pypi/v/semble-grammars?color=%23007ec6&label=pypi%20package" alt="Package version"></a>
+    <a href="https://github.com/MinishLab/semble-grammars/blob/main/LICENSE">
+      <img src="https://img.shields.io/badge/license-MIT-green" alt="License - MIT">
+    </a>
+  </h2>
+</div>
 
 Compiled [tree-sitter](https://tree-sitter.github.io/tree-sitter/) grammars for
 [Semble](https://github.com/MinishLab/semble).
@@ -7,6 +19,14 @@ The package installs a platform-specific grammar archive and extracts individual
 libraries to a local cache on first use. It does not access the network at
 runtime.
 
+## Installation
+
+```bash
+pip install semble-grammars
+```
+
+## Usage
+
 ```python
 from semble_grammars import get_parser
 
@@ -18,24 +38,33 @@ tree = parser.parse(b"def hello(): pass")
 macOS arm64 and x86-64, Linux arm64 and x86-64, and Windows arm64 and x86-64.
 
 <details>
-<summary>Included grammars (75)</summary>
+<summary>Included grammars (77)</summary>
 
 The canonical names accepted by `get_parser()` are:
 
-`asciidoc`, `astro`, `bash`, `c`, `clojure`, `cmake`, `cpp`, `csharp`, `css`,
-`dart`, `dockerfile`, `dtd`, `elixir`, `embedded_template`, `erlang`, `fortran`,
-`gitignore`, `go`, `gotmpl`, `graphql`, `groovy`, `haskell`, `hcl`, `html`,
-`ini`, `java`, `javascript`, `jinja2`, `json`, `json5`, `jsonc`, `jsonnet`,
-`julia`, `just`, `kdl`, `kotlin`, `latex`, `lua`, `make`, `markdown`,
-`markdown_inline`, `nix`, `objc`, `ocaml`, `org`, `perl`, `php`, `php_only`,
-`powershell`, `properties`, `proto`, `python`, `r`, `racket`, `rst`, `ruby`,
-`rust`, `scala`, `scheme`, `scss`, `solidity`, `sql`, `starlark`, `svelte`,
-`swift`, `toml`, `tsx`, `typescript`, `typst`, `vim`, `vue`, `wat`, `xml`,
-`yaml`, `zig`.
-
-`py` is an alias for `python`, and `terraform` is an alias for `hcl`.
+`asciidoc`, `astro`, `bash`, `batch`, `c`, `clojure`, `cmake`, `cpp`,
+`csharp`, `css`, `dart`, `dockerfile`, `dtd`, `elixir`, `embedded_template`,
+`erlang`, `fortran`, `gitignore`, `go`, `gotmpl`, `graphql`, `groovy`,
+`haskell`, `hcl`, `heex`, `html`, `ini`, `java`, `javascript`, `jinja2`,
+`json`, `json5`, `jsonc`, `jsonnet`, `julia`, `just`, `kdl`, `kotlin`,
+`latex`, `lua`, `make`, `markdown`, `markdown_inline`, `nix`, `objc`,
+`ocaml`, `org`, `perl`, `php`, `php_only`, `powershell`, `properties`,
+`proto`, `python`, `r`, `racket`, `rst`, `ruby`, `rust`, `scala`, `scheme`,
+`scss`, `solidity`, `sql`, `starlark`, `svelte`, `swift`, `toml`, `tsx`,
+`typescript`, `typst`, `vim`, `vue`, `wat`, `xml`, `yaml`, `zig`.
+
+`py` is an alias for `python`, `terraform` is an alias for `hcl`,
+`embeddedtemplate` is an alias for `embedded_template`, and `zsh` is an alias
+for `bash`.
 </details>
 
+The MIT license above covers `semble-grammars`' own code. Each bundled grammar is a
+separate upstream work with its own license; see
+[`THIRD_PARTY_NOTICES.md`](src/semble_grammars/grammars/THIRD_PARTY_NOTICES.md) for the
+full list.
+
+## Building from source
+
 Build grammars from the pinned sources in
 [`sources.json`](src/semble_grammars/grammars/sources.json):
 

pyproject.toml

@@ -4,8 +4,8 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "semble-grammars"
-description = "Locally-cached tree-sitter grammar bundle for Semble, with no runtime downloads."
-authors = [{name = "MinishLab", email = "info@minishlab.com"}]
+description = "Prebuilt tree-sitter grammars for Semble"
+authors = [{name = "Thomas van Dongen", email = "thomasvdongen@proton.me"}, { name = "Stéphan Tulkens", email = "stephantul@gmail.com"}]
 readme = { file = "README.md", content-type = "text/markdown" }
 dynamic = ["version"]
 license = "MIT"

src/semble_grammars/grammars/THIRD_PARTY_NOTICES.md

@@ -10,6 +10,7 @@ full license and notice text of each.
 | asciidoc | https://github.com/cathaysia/tree-sitter-asciidoc | `e0710115e7b060ce6681b5182a49a792813006f1` | Apache-2.0 |
 | astro | https://github.com/virchau13/tree-sitter-astro | `213f6e6973d9b456c6e50e86f19f66877e7ef0ee` | MIT |
 | bash | https://github.com/tree-sitter/tree-sitter-bash | `a06c2e4415e9bc0346c6b86d401879ffb44058f7` (v0.25.1) | MIT |
+| batch | https://github.com/wharflab/tree-sitter-batch | `f477847c89b7db79b0d5e3527ba39318bf054e2f` (v0.9.0) | MIT |
 | c | https://github.com/tree-sitter/tree-sitter-c | `b780e47fc780ddc8da13afa35a3f4ed5c157823d` (v0.24.2) | MIT |
 | clojure | https://github.com/sogaiu/tree-sitter-clojure | `3a1ace906c151dd631cf6f149b5083f2b60e6a9e` (v0.0.13) | CC0-1.0 |
 | cmake | https://github.com/uyha/tree-sitter-cmake | `ca627bb5828616b6246aafdc3c3222789e728e37` (v0.7.4) | MIT |
@@ -30,6 +31,7 @@ full license and notice text of each.
 | groovy | https://github.com/murtaza64/tree-sitter-groovy | `deb0dcf8c4544f07564060f6e9b9f6e4b0bfc27d` | MIT |
 | haskell | https://github.com/tree-sitter/tree-sitter-haskell | `0975ef72fc3c47b530309ca93937d7d143523628` | MIT |
 | hcl | https://github.com/tree-sitter-grammars/tree-sitter-hcl | `fad991865fee927dd1de5e172fb3f08ac674d914` (v1.2.0) | Apache-2.0 |
+| heex | https://github.com/phoenixframework/tree-sitter-heex | `5842537f734d7c12685bf27d6005313e3e5a47a0` (v0.9.0) | MIT |
 | html | https://github.com/tree-sitter/tree-sitter-html | `5a5ca8551a179998360b4a4ca2c0f366a35acc03` (v0.23.2) | MIT |
 | ini | https://github.com/justinmk/tree-sitter-ini | `f0285fe577ad298ad79f8633e643ad60646e3027` (v1.4.0) | Apache-2.0 |
 | java | https://github.com/tree-sitter/tree-sitter-java | `94703d5a6bed02b98e438d7cad1136c01a60ba2c` (v0.23.5) | MIT |

src/semble_grammars/grammars/sources.json

@@ -679,5 +679,23 @@
       "src/parser.c",
       "src/scanner.c"
     ]
+  },
+  "batch": {
+    "repository": "https://github.com/wharflab/tree-sitter-batch",
+    "commit": "f477847c89b7db79b0d5e3527ba39318bf054e2f",
+    "ref": "v0.9.0",
+    "spdx_license": "MIT",
+    "sources": [
+      "src/parser.c"
+    ]
+  },
+  "heex": {
+    "repository": "https://github.com/phoenixframework/tree-sitter-heex",
+    "commit": "5842537f734d7c12685bf27d6005313e3e5a47a0",
+    "ref": "v0.9.0",
+    "spdx_license": "MIT",
+    "sources": [
+      "src/parser.c"
+    ]
   }
 }

src/semble_grammars/loader.py

@@ -18,6 +18,9 @@
     "py": "python",
     # Terraform files are valid HCL; there is no separate compiled grammar.
     "terraform": "hcl",
+    "embeddedtemplate": "embedded_template",
+    # Zsh is largely a superset of POSIX shell; there is no separate compiled grammar.
+    "zsh": "bash",
 }
 
 _OS_NAMES = {"darwin": "macos", "linux": "linux", "windows": "windows"}

src/semble_grammars/version.py

@@ -1 +1 @@
-__version__ = "0.1.1"
+__version__ = "0.1.2"

tests/test_loader.py

@@ -20,6 +20,8 @@
         ("python", b"def foo(:\n", True),
         ("json", b'{"a": [1, 2, 3]}', False),
         ("sql", b"SELECT * FROM foo WHERE x = 1;", False),
+        ("batch", b"@echo off\nset FOO=bar\n", False),
+        ("heex", b"<div><%= @name %></div>\n", False),
     ],
 )
 def test_get_parser_parses_bundled_languages(name, source, expect_error):
@@ -32,6 +34,14 @@ def test_terraform_is_an_alias_for_hcl():
     assert get_language("terraform") == get_language("hcl")
 
 
+def test_embeddedtemplate_is_an_alias_for_embedded_template():
+    assert get_language("embeddedtemplate") == get_language("embedded_template")
+
+
+def test_zsh_is_an_alias_for_bash():
+    assert get_language("zsh") == get_language("bash")
+
+
 def test_unknown_language_raises_with_available_languages_listed():
     with pytest.raises(LanguageNotFoundError, match="nonexistent"):
         get_parser("nonexistent")