| 1 | #!/usr/bin/env python3 |
| 2 | """Generate src/index/languages_table.rs from GitHub Linguist's data. |
| 3 | |
| 4 | Dev-only; the output is committed. Needs PyYAML. |
| 5 | |
| 6 | tools/gen-languages/gen.py [VERSION] # default: the pinned version |
| 7 | |
| 8 | Downloads languages.yml, vendor.yml and documentation.yml for VERSION and |
| 9 | writes the table. Re-run when bumping Linguist. |
| 10 | """ |
| 11 | |
| 12 | import pathlib |
| 13 | import sys |
| 14 | import urllib.request |
| 15 | |
| 16 | import yaml |
| 17 | |
| 18 | VERSION = sys.argv[1] if len(sys.argv) > 1 else "v9.7.0" |
| 19 | BASE = f"https://raw.githubusercontent.com/github-linguist/linguist/{VERSION}/lib/linguist/" |
| 20 | OUT = pathlib.Path(__file__).resolve().parents[2] / "src/index/languages_table.rs" |
| 21 | |
| 22 | # Extensions claimed by several languages, settled by hand (Linguist uses |
| 23 | # content heuristics we don't run). Everything else: see `pick`. |
| 24 | OVERRIDES = { |
| 25 | ".h": "C", |
| 26 | ".m": "Objective-C", |
| 27 | ".pl": "Perl", |
| 28 | ".pm": "Perl", |
| 29 | ".t": "Perl", |
| 30 | ".md": "Markdown", |
| 31 | ".ts": "TypeScript", |
| 32 | ".sql": "SQL", |
| 33 | ".inc": "PHP", |
| 34 | ".v": "Verilog", |
| 35 | ".fs": "F#", |
| 36 | ".r": "R", |
| 37 | ".d": "D", |
| 38 | ".cls": "TeX", |
| 39 | ".pro": "Prolog", |
| 40 | ".s": "Unix Assembly", |
| 41 | ".asm": "Assembly", |
| 42 | ".l": "Lex", |
| 43 | ".ms": "Roff", |
| 44 | ".bf": "Brainfuck", |
| 45 | ".cs": "C#", |
| 46 | ".rs": "Rust", |
| 47 | ".html": "HTML", |
| 48 | ".txt": "Text", |
| 49 | ".json": "JSON", |
| 50 | ".yml": "YAML", |
| 51 | ".xml": "XML", |
| 52 | ".sh": "Shell", |
| 53 | ".mod": "Modula-2", |
| 54 | ".gs": "JavaScript", |
| 55 | ".ex": "Elixir", |
| 56 | ".nl": "NewLisp", |
| 57 | ".lisp": "Common Lisp", |
| 58 | ".cl": "Common Lisp", |
| 59 | ".tst": "Scilab", |
| 60 | ".x": "Logos", |
| 61 | ".b": "Brainfuck", |
| 62 | ".n": "Roff", |
| 63 | ".i": "SWIG", |
| 64 | ".tpl": "Smarty", |
| 65 | ".sc": "Scala", |
| 66 | ".st": "Smalltalk", |
| 67 | ".frm": "VBA", |
| 68 | ".bas": "BASIC", |
| 69 | ".fx": "HLSL", |
| 70 | ".mask": "Unity3D Asset", |
| 71 | ".plt": "Gnuplot", |
| 72 | ".ncl": "NCL", |
| 73 | ".lsp": "Common Lisp", |
| 74 | ".brd": "Eagle", |
| 75 | ".mdx": "MDX", |
| 76 | ".cake": "C#", |
| 77 | ".builds": "XML", |
| 78 | ".ice": "JSON", |
| 79 | ".rpy": "Ren'Py", |
| 80 | ".sol": "Solidity", |
| 81 | ".yaml": "YAML", |
| 82 | ".cfg": "INI", |
| 83 | ".f": "Fortran", |
| 84 | ".es": "JavaScript", |
| 85 | } |
| 86 | |
| 87 | KIND_RANK = {"programming": 0, "markup": 1, "data": 2, "prose": 3} |
| 88 | |
| 89 | |
| 90 | def fetch(name): |
| 91 | with urllib.request.urlopen(BASE + name) as res: |
| 92 | return yaml.safe_load(res.read()) |
| 93 | |
| 94 | |
| 95 | def rust_str(s): |
| 96 | return '"' + s.replace("\\", "\\\\").replace('"', '\\"') + '"' |
| 97 | |
| 98 | |
| 99 | def main(): |
| 100 | languages = fetch("languages.yml") |
| 101 | vendor = fetch("vendor.yml") |
| 102 | documentation = fetch("documentation.yml") |
| 103 | |
| 104 | names = sorted(languages) |
| 105 | index = {name: i for i, name in enumerate(names)} |
| 106 | |
| 107 | def pick(ext, candidates): |
| 108 | if ext in OVERRIDES and OVERRIDES[ext] in candidates: |
| 109 | return OVERRIDES[ext] |
| 110 | primary = [c for c in candidates if languages[c].get("extensions", [None])[0] == ext] |
| 111 | pool = primary or candidates |
| 112 | return min(pool, key=lambda c: (KIND_RANK.get(languages[c]["type"], 9), c)) |
| 113 | |
| 114 | by_ext = {} |
| 115 | by_name = {} |
| 116 | for name in names: |
| 117 | lang = languages[name] |
| 118 | for ext in lang.get("extensions", []): |
| 119 | by_ext.setdefault(ext.lower(), []).append(name) |
| 120 | for filename in lang.get("filenames", []): |
| 121 | by_name.setdefault(filename, []).append(name) |
| 122 | |
| 123 | extensions = sorted((ext, index[pick(ext, c)]) for ext, c in by_ext.items()) |
| 124 | filenames = sorted( |
| 125 | (f, index[min(c, key=lambda n: (KIND_RANK.get(languages[n]["type"], 9), n))]) |
| 126 | for f, c in by_name.items() |
| 127 | ) |
| 128 | |
| 129 | out = [] |
| 130 | w = out.append |
| 131 | w(f"// @generated by tools/gen-languages from GitHub Linguist {VERSION}. Do not edit.\n") |
| 132 | w("// Linguist is MIT licensed: https://github.com/github-linguist/linguist\n\n") |
| 133 | w("use super::languages::{Kind, Language};\n\n") |
| 134 | w(f"pub const LINGUIST_VERSION: &str = {rust_str(VERSION)};\n\n") |
| 135 | w("/// Every language; `group` folds a language into its parent for statistics.\n") |
| 136 | w("pub static LANGUAGES: &[Language] = &[\n") |
| 137 | for name in names: |
| 138 | lang = languages[name] |
| 139 | kind = lang["type"].capitalize() |
| 140 | group = lang.get("group") |
| 141 | group = f"Some({index[group]})" if group in index else "None" |
| 142 | w(f" Language {{ name: {rust_str(name)}, kind: Kind::{kind}, group: {group} }},\n") |
| 143 | w("];\n\n") |
| 144 | w("/// Lowercase extensions (with the dot), sorted, to an index into `LANGUAGES`.\n") |
| 145 | w("pub static EXTENSIONS: &[(&str, u16)] = &[\n") |
| 146 | for ext, i in extensions: |
| 147 | w(f" ({rust_str(ext)}, {i}),\n") |
| 148 | w("];\n\n") |
| 149 | w("/// Exact file names, sorted, to an index into `LANGUAGES`.\n") |
| 150 | w("pub static FILENAMES: &[(&str, u16)] = &[\n") |
| 151 | for f, i in filenames: |
| 152 | w(f" ({rust_str(f)}, {i}),\n") |
| 153 | w("];\n\n") |
| 154 | for title, const, patterns in [ |
| 155 | ("Paths of vendored code (regular expressions).", "VENDOR", vendor), |
| 156 | ("Paths of documentation (regular expressions).", "DOCUMENTATION", documentation), |
| 157 | ]: |
| 158 | w(f"/// {title}\n") |
| 159 | w(f"pub static {const}: &[&str] = &[\n") |
| 160 | for p in patterns: |
| 161 | w(f" {rust_str(p)},\n") |
| 162 | w("];\n\n") |
| 163 | OUT.write_text("".join(out).rstrip() + "\n") |
| 164 | print(f"wrote {OUT} ({len(names)} languages, {len(extensions)} extensions, " |
| 165 | f"{len(filenames)} file names)") |
| 166 | |
| 167 | |
| 168 | if __name__ == "__main__": |
| 169 | main() |