seendiff 0.0.2

package/src/highlight.js

@@ -0,0 +1,220 @@
+import hljs from "highlight.js";
+
+export const MAX_LINES = 10_000;
+export const MAX_BYTES = 1_000_000;
+export const MAX_MEAN_LINE_LEN = 500;
+
+const EXT_LANG = {
+    ".js": "javascript", ".mjs": "javascript", ".cjs": "javascript",
+    ".jsx": "javascript", ".ts": "typescript", ".tsx": "typescript",
+    ".py": "python", ".pyw": "python",
+    ".rb": "ruby", ".go": "go", ".rs": "rust",
+    ".java": "java", ".kt": "kotlin", ".kts": "kotlin",
+    ".c": "c", ".h": "c", ".cc": "cpp", ".cpp": "cpp", ".cxx": "cpp",
+    ".hpp": "cpp", ".hh": "cpp", ".cs": "csharp",
+    ".php": "php", ".swift": "swift", ".m": "objectivec", ".mm": "objectivec",
+    ".scala": "scala", ".sh": "bash", ".bash": "bash", ".zsh": "bash",
+    ".ps1": "powershell", ".pl": "perl", ".pm": "perl", ".lua": "lua",
+    ".r": "r", ".jl": "julia", ".ex": "elixir", ".exs": "elixir",
+    ".erl": "erlang", ".hs": "haskell", ".clj": "clojure", ".cljs": "clojure",
+    ".ml": "ocaml", ".fs": "fsharp", ".fsx": "fsharp", ".dart": "dart",
+    ".sql": "sql", ".html": "xml", ".htm": "xml", ".xml": "xml",
+    ".svg": "xml", ".css": "css", ".scss": "scss", ".less": "less",
+    ".json": "json", ".jsonc": "json", ".yaml": "yaml", ".yml": "yaml",
+    ".toml": "ini", ".ini": "ini", ".cfg": "ini", ".md": "markdown",
+    ".markdown": "markdown", ".dockerfile": "dockerfile", ".vue": "xml",
+    ".graphql": "graphql", ".gql": "graphql", ".proto": "protobuf",
+    ".diff": "diff", ".patch": "diff", ".makefile": "makefile",
+    ".cmake": "cmake", ".vim": "vim", ".zig": "zig", ".nim": "nim",
+    ".sol": "solidity",
+};
+
+function extOf(filename) {
+    const base = filename.split("/").pop() || filename;
+    if (base === "Dockerfile") return ".dockerfile";
+    if (base === "Makefile" || base === "makefile") return ".makefile";
+    if (base === "CMakeLists.txt") return ".cmake";
+    const i = base.lastIndexOf(".");
+    return i >= 0 ? base.slice(i).toLowerCase() : "";
+}
+
+const SCOPE_MAP = [
+    [/^keyword$/, "k"],
+    [/^built_in$/, "nb"],
+    [/^type$/, "kt"],
+    [/^literal$/, "kc"],
+    [/^number$/, "m"],
+    [/^operator$/, "o"],
+    [/^punctuation$/, "p"],
+    [/^comment$/, "c1"],
+    [/^doctag$/, "c1"],
+    [/^meta/, "cp"],
+    [/^string$/, "s1"],
+    [/^regexp$/, "sr"],
+    [/^subst$/, "si"],
+    [/^symbol$/, "ss"],
+    [/^class$/, "nc"],
+    [/^title\.class/, "nc"],
+    [/^title\.function/, "nf"],
+    [/^title\.function_/, "nf"],
+    [/^function$/, "nf"],
+    [/^title$/, "nf"],
+    [/^params$/, "n"],
+    [/^variable$/, "nv"],
+    [/^variable\.language/, "nb"],
+    [/^variable\.constant/, "no"],
+    [/^property$/, "n"],
+    [/^attr$/, "na"],
+    [/^attribute$/, "na"],
+    [/^tag$/, "nt"],
+    [/^name$/, "nt"],
+    [/^selector-tag$/, "nt"],
+    [/^selector-id$/, "nd"],
+    [/^selector-class$/, "nc"],
+    [/^selector-attr$/, "na"],
+    [/^selector-pseudo$/, "nd"],
+    [/^template-tag$/, "k"],
+    [/^template-variable$/, "nv"],
+    [/^addition$/, "gi"],
+    [/^deletion$/, "gd"],
+    [/^emphasis$/, "ge"],
+    [/^strong$/, "gs"],
+    [/^formula$/, "n"],
+    [/^section$/, "gh"],
+    [/^bullet$/, "nt"],
+    [/^link$/, "nt"],
+    [/^quote$/, "c1"],
+    [/^code$/, "s1"],
+];
+
+function shortClass(scopeAttr) {
+    const parts = scopeAttr.split(/\s+/).filter(Boolean);
+    for (const part of parts) {
+        for (const [re, cls] of SCOPE_MAP) {
+            if (re.test(part)) return cls;
+        }
+    }
+    return "n";
+}
+
+export function langFor(filename) {
+    const ext = extOf(filename);
+    const lang = EXT_LANG[ext];
+    if (!lang) return null;
+    return hljs.getLanguage(lang) ? lang : null;
+}
+
+const NAMED_ENTITIES = { amp: "&", lt: "<", gt: ">", quot: '"', apos: "'" };
+const ENTITY_RE = /&(#x[0-9a-fA-F]+|#\d+|[a-zA-Z]+);/g;
+function decodeEntities(s) {
+    return s.replace(ENTITY_RE, (m, body) => {
+        if (body[0] === "#") {
+            const code = body[1] === "x" || body[1] === "X"
+                ? parseInt(body.slice(2), 16)
+                : parseInt(body.slice(1), 10);
+            return Number.isFinite(code) ? String.fromCodePoint(code) : m;
+        }
+        return NAMED_ENTITIES[body] ?? m;
+    });
+}
+
+function parseHljsHtml(html) {
+    const tokens = [];
+    const classStack = [];
+    let i = 0;
+    const n = html.length;
+    let buf = "";
+
+    function flush() {
+        if (buf) {
+            const cls = classStack.length ? shortClass(classStack[classStack.length - 1]) : "";
+            tokens.push([cls, decodeEntities(buf)]);
+            buf = "";
+        }
+    }
+
+    while (i < n) {
+        if (html.startsWith("<span class=\"", i)) {
+            flush();
+            const start = i + '<span class="'.length;
+            const end = html.indexOf('"', start);
+            classStack.push(html.slice(start, end).replace(/^hljs-/, ""));
+            i = html.indexOf(">", end) + 1;
+            continue;
+        }
+        if (html.startsWith("</span>", i)) {
+            flush();
+            classStack.pop();
+            i += "</span>".length;
+            continue;
+        }
+        buf += html[i];
+        i++;
+    }
+    flush();
+    return tokens;
+}
+
+export function tokenizeLines(text, lang, nLines) {
+    let html;
+    try {
+        html = hljs.highlight(text, { language: lang, ignoreIllegals: true }).value;
+    } catch {
+        html = text.replace(/&/g, "&amp;").replace(/</g, "&lt;").replace(/>/g, "&gt;");
+    }
+    const flat = parseHljsHtml(html);
+    const lines = [];
+    let cur = [];
+    for (const [cls, value] of flat) {
+        const parts = value.split("\n");
+        parts.forEach((piece, i) => {
+            if (i) {
+                lines.push(cur);
+                cur = [];
+            }
+            if (piece) cur.push([cls, piece]);
+        });
+    }
+    lines.push(cur);
+    while (lines.length < nLines) lines.push([]);
+    return lines.slice(0, nLines);
+}
+
+export function guard(rows) {
+    if (rows.length > MAX_LINES) return `file too large (${rows.length} lines)`;
+    const total = rows.reduce((n, r) => n + r.text.length, 0);
+    if (total > MAX_BYTES) return `file too large (${Math.floor(total / 1024)} kB)`;
+    if (rows.length && total / rows.length > MAX_MEAN_LINE_LEN) return "minified content (very long lines)";
+    return null;
+}
+
+export function highlightRows(rows, filename) {
+    if (guard(rows) !== null) return null;
+    const lang = langFor(filename);
+    if (!lang) return null;
+    try {
+        const newRows = rows.filter((r) => r.type !== "del");
+        const oldRows = rows.filter((r) => r.type !== "add");
+        const newToks = tokenizeLines(newRows.map((r) => r.text).join("\n"), lang, newRows.length);
+        const oldToks = tokenizeLines(oldRows.map((r) => r.text).join("\n"), lang, oldRows.length);
+        const out = [];
+        let ni = 0;
+        let oi = 0;
+        for (const r of rows) {
+            if (r.type === "del") {
+                out.push(oldToks[oi]);
+                oi++;
+            } else if (r.type === "add") {
+                out.push(newToks[ni]);
+                ni++;
+            } else {
+                out.push(newToks[ni]);
+                ni++;
+                oi++;
+            }
+        }
+        return out;
+    } catch {
+        return null;
+    }
+}