glost-processor 0.7.0 → 1.0.2

package/src/__benchmarks__/stream-processor.bench.js

@@ -0,0 +1,202 @@
+/**
+ * GLOSTStreamProcessor Performance Benchmarks
+ *
+ * Compares batch (eager) processing against streaming for documents
+ * of various sizes. Includes 10K and 100K word documents.
+ */
+import { bench, describe } from "vitest";
+import { createGLOSTWordNode, createSimpleDocument, createSentenceFromWords, createParagraphFromSentences, createGLOSTRootNode, } from "glost-core";
+import { GLOSTProcessor } from "../processor.js";
+import { GLOSTStreamProcessor } from "../stream-processor.js";
+// ============================================================================
+// Document factories
+// ============================================================================
+/** Create a document with `wordCount` words in a single sentence */
+function makeDocumentByWords(wordCount) {
+    const words = Array.from({ length: wordCount }, (_, i) => createGLOSTWordNode({ value: `word${i}` }));
+    return createSimpleDocument(words, "en", "latin");
+}
+/**
+ * Create a document with `sentenceCount` sentences, each sentence
+ * containing `wordsPerSentence` words, spread across `paragraphCount`
+ * paragraphs.
+ */
+function makeDocumentBySentences(sentenceCount, wordsPerSentence = 10, paragraphCount = 1) {
+    const sentencesPerParagraph = Math.ceil(sentenceCount / paragraphCount);
+    const paragraphs = [];
+    let remaining = sentenceCount;
+    for (let p = 0; p < paragraphCount; p++) {
+        const count = Math.min(remaining, sentencesPerParagraph);
+        const sentences = Array.from({ length: count }, (_, s) => {
+            const words = Array.from({ length: wordsPerSentence }, (_, w) => createGLOSTWordNode({ value: `p${p}s${s}w${w}` }));
+            return createSentenceFromWords(words, "en", "latin");
+        });
+        paragraphs.push(createParagraphFromSentences(sentences));
+        remaining -= count;
+        if (remaining <= 0)
+            break;
+    }
+    return createGLOSTRootNode({
+        lang: "en",
+        script: "latin",
+        children: paragraphs,
+    });
+}
+// ============================================================================
+// Extensions
+// ============================================================================
+function createVisitExtension(id) {
+    return {
+        id,
+        name: `Visit Extension ${id}`,
+        visit: {
+            word: (node) => {
+                return {
+                    ...node,
+                    extras: { ...node.extras, [id]: true },
+                };
+            },
+        },
+    };
+}
+function createChunkVisitExtension(id) {
+    return {
+        id,
+        name: `Chunk Visit Extension ${id}`,
+        streamingSupport: "chunk",
+        visit: {
+            word: (node) => {
+                return {
+                    ...node,
+                    extras: { ...node.extras, [id]: true },
+                };
+            },
+        },
+    };
+}
+// ============================================================================
+// Benchmark: medium documents (for baseline comparison)
+// ============================================================================
+describe("Streaming vs Batch: Medium Documents", () => {
+    const doc1k = makeDocumentBySentences(100, 10); // 1K words
+    const ext = createVisitExtension("test");
+    const chunkExt = createChunkVisitExtension("chunk-test");
+    bench("batch — 1K words (GLOSTProcessor)", async () => {
+        const proc = new GLOSTProcessor().use(ext);
+        await proc.process(doc1k);
+    });
+    bench("stream — 1K words, batchSize=50", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc1k, { batchSize: 50 })) {
+            // consume
+        }
+    });
+    bench("stream — 1K words, batchSize=10", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc1k, { batchSize: 10 })) {
+            // consume
+        }
+    });
+    bench("stream — 1K words, batchSize=100", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc1k, { batchSize: 100 })) {
+            // consume
+        }
+    });
+});
+// ============================================================================
+// Benchmark: 10K word documents
+// ============================================================================
+describe("Streaming: 10K Word Documents", () => {
+    // 1000 sentences * 10 words = 10K words
+    const doc10k = makeDocumentBySentences(1000, 10);
+    const ext = createVisitExtension("test");
+    const chunkExt = createChunkVisitExtension("chunk-test");
+    bench("batch — 10K words (GLOSTProcessor)", async () => {
+        const proc = new GLOSTProcessor().use(ext);
+        await proc.process(doc10k);
+    });
+    bench("stream — 10K words, batchSize=50", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc10k, { batchSize: 50 })) {
+            // consume
+        }
+    });
+    bench("stream — 10K words, batchSize=100", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc10k, {
+            batchSize: 100,
+        })) {
+            // consume
+        }
+    });
+    bench("stream — 10K words, batchSize=500", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc10k, {
+            batchSize: 500,
+        })) {
+            // consume
+        }
+    });
+});
+// ============================================================================
+// Benchmark: 100K word documents
+// ============================================================================
+describe("Streaming: 100K Word Documents", () => {
+    // 10000 sentences * 10 words = 100K words
+    const doc100k = makeDocumentBySentences(10_000, 10, 100);
+    const ext = createVisitExtension("test");
+    const chunkExt = createChunkVisitExtension("chunk-test");
+    bench("batch — 100K words (GLOSTProcessor)", async () => {
+        const proc = new GLOSTProcessor().use(ext);
+        await proc.process(doc100k);
+    });
+    bench("stream — 100K words, batchSize=50", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc100k, {
+            batchSize: 50,
+        })) {
+            // consume
+        }
+    });
+    bench("stream — 100K words, batchSize=500", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc100k, {
+            batchSize: 500,
+        })) {
+            // consume
+        }
+    });
+    bench("stream — 100K words, batchSize=1000", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const _chunk of proc.stream(doc100k, {
+            batchSize: 1000,
+        })) {
+            // consume
+        }
+    });
+});
+// ============================================================================
+// Benchmark: early termination via break
+// ============================================================================
+describe("Streaming: Early Termination", () => {
+    const doc10k = makeDocumentBySentences(1000, 10);
+    const chunkExt = createChunkVisitExtension("chunk-test");
+    bench("stream — cancel after first chunk (lazy win)", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        for await (const chunk of proc.stream(doc10k, { batchSize: 50 })) {
+            break; // Only consume first chunk
+        }
+    });
+    bench("stream — cancel after 10 chunks vs full batch", async () => {
+        const proc = new GLOSTStreamProcessor().use(chunkExt);
+        let count = 0;
+        for await (const _chunk of proc.stream(doc10k, {
+            batchSize: 50,
+        })) {
+            if (++count >= 10)
+                break;
+        }
+    });
+});
+//# sourceMappingURL=stream-processor.bench.js.map

package/src/__benchmarks__/stream-processor.bench.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"stream-processor.bench.js","sourceRoot":"","sources":["stream-processor.bench.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,QAAQ,CAAC;AACzC,OAAO,EACL,mBAAmB,EACnB,oBAAoB,EACpB,uBAAuB,EACvB,4BAA4B,EAC5B,mBAAmB,GACpB,MAAM,YAAY,CAAC;AAEpB,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AACjD,OAAO,EAAE,oBAAoB,EAAE,MAAM,wBAAwB,CAAC;AAE9D,+EAA+E;AAC/E,qBAAqB;AACrB,+EAA+E;AAE/E,oEAAoE;AACpE,SAAS,mBAAmB,CAAC,SAAiB;IAC5C,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,SAAS,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACvD,mBAAmB,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC,EAAE,EAAE,CAAC,CAC3C,CAAC;IACF,OAAO,oBAAoB,CAAC,KAAK,EAAE,IAAI,EAAE,OAAO,CAAC,CAAC;AACpD,CAAC;AAED;;;;GAIG;AACH,SAAS,uBAAuB,CAC9B,aAAqB,EACrB,gBAAgB,GAAG,EAAE,EACrB,cAAc,GAAG,CAAC;IAElB,MAAM,qBAAqB,GAAG,IAAI,CAAC,IAAI,CACrC,aAAa,GAAG,cAAc,CAC/B,CAAC;IACF,MAAM,UAAU,GAAG,EAAE,CAAC;IACtB,IAAI,SAAS,GAAG,aAAa,CAAC;IAE9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,cAAc,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,EAAE,qBAAqB,CAAC,CAAC;QACzD,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YACvD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,gBAAgB,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAC9D,mBAAmB,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,CAAC,CAClD,CAAC;YACF,OAAO,uBAAuB,CAAC,KAAK,EAAE,IAAI,EAAE,OAAO,CAAC,CAAC;QACvD,CAAC,CAAC,CAAC;QACH,UAAU,CAAC,IAAI,CAAC,4BAA4B,CAAC,SAAS,CAAC,CAAC,CAAC;QACzD,SAAS,IAAI,KAAK,CAAC;QACnB,IAAI,SAAS,IAAI,CAAC;YAAE,MAAM;IAC5B,CAAC;IAED,OAAO,mBAAmB,CAAC;QACzB,IAAI,EAAE,IAAI;QACV,MAAM,EAAE,OAAO;QACf,QAAQ,EAAE,UAAU;KACrB,CAAC,CAAC;AACL,CAAC;AAED,+EAA+E;AAC/E,aAAa;AACb,+EAA+E;AAE/E,SAAS,oBAAoB,CAAC,EAAU;IACtC,OAAO;QACL,EAAE;QACF,IAAI,EAAE,mBAAmB,EAAE,EAAE;QAC7B,KAAK,EAAE;YACL,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE;gBACb,OAAO;oBACL,GAAG,IAAI;oBACP,MAAM,EAAE,EAAE,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,CAAC,EAAE,IAAI,EAAE;iBACvC,CAAC;YACJ,CAAC;SACF;KACF,CAAC;AACJ,CAAC;AAED,SAAS,yBAAyB,CAAC,EAAU;IAC3C,OAAO;QACL,EAAE;QACF,IAAI,EAAE,yBAAyB,EAAE,EAAE;QACnC,gBAAgB,EAAE,OAAO;QACzB,KAAK,EAAE;YACL,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE;gBACb,OAAO;oBACL,GAAG,IAAI;oBACP,MAAM,EAAE,EAAE,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,CAAC,EAAE,IAAI,EAAE;iBACvC,CAAC;YACJ,CAAC;SACF;KACF,CAAC;AACJ,CAAC;AAED,+EAA+E;AAC/E,wDAAwD;AACxD,+EAA+E;AAE/E,QAAQ,CAAC,sCAAsC,EAAE,GAAG,EAAE;IACpD,MAAM,KAAK,GAAG,uBAAuB,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW;IAC3D,MAAM,GAAG,GAAG,oBAAoB,CAAC,MAAM,CAAC,CAAC;IACzC,MAAM,QAAQ,GAAG,yBAAyB,CAAC,YAAY,CAAC,CAAC;IAEzD,KAAK,CAAC,mCAAmC,EAAE,KAAK,IAAI,EAAE;QACpD,MAAM,IAAI,GAAG,IAAI,cAAc,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAC3C,MAAM,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;IAC5B,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,iCAAiC,EAAE,KAAK,IAAI,EAAE;QAClD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC;YACjE,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,iCAAiC,EAAE,KAAK,IAAI,EAAE;QAClD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC;YACjE,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,kCAAkC,EAAE,KAAK,IAAI,EAAE;QACnD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,GAAG,EAAE,CAAC,EAAE,CAAC;YAClE,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEH,+EAA+E;AAC/E,gCAAgC;AAChC,+EAA+E;AAE/E,QAAQ,CAAC,+BAA+B,EAAE,GAAG,EAAE;IAC7C,wCAAwC;IACxC,MAAM,MAAM,GAAG,uBAAuB,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACjD,MAAM,GAAG,GAAG,oBAAoB,CAAC,MAAM,CAAC,CAAC;IACzC,MAAM,QAAQ,GAAG,yBAAyB,CAAC,YAAY,CAAC,CAAC;IAEzD,KAAK,CAAC,oCAAoC,EAAE,KAAK,IAAI,EAAE;QACrD,MAAM,IAAI,GAAG,IAAI,cAAc,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAC3C,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;IAC7B,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,kCAAkC,EAAE,KAAK,IAAI,EAAE;QACnD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC;YAClE,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,mCAAmC,EAAE,KAAK,IAAI,EAAE;QACpD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE;YAC7C,SAAS,EAAE,GAAG;SACf,CAAC,EAAE,CAAC;YACH,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,mCAAmC,EAAE,KAAK,IAAI,EAAE;QACpD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE;YAC7C,SAAS,EAAE,GAAG;SACf,CAAC,EAAE,CAAC;YACH,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEH,+EAA+E;AAC/E,iCAAiC;AACjC,+EAA+E;AAE/E,QAAQ,CAAC,gCAAgC,EAAE,GAAG,EAAE;IAC9C,0CAA0C;IAC1C,MAAM,OAAO,GAAG,uBAAuB,CAAC,MAAM,EAAE,EAAE,EAAE,GAAG,CAAC,CAAC;IACzD,MAAM,GAAG,GAAG,oBAAoB,CAAC,MAAM,CAAC,CAAC;IACzC,MAAM,QAAQ,GAAG,yBAAyB,CAAC,YAAY,CAAC,CAAC;IAEzD,KAAK,CAAC,qCAAqC,EAAE,KAAK,IAAI,EAAE;QACtD,MAAM,IAAI,GAAG,IAAI,cAAc,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAC3C,MAAM,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IAC9B,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,mCAAmC,EAAE,KAAK,IAAI,EAAE;QACpD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE;YAC9C,SAAS,EAAE,EAAE;SACd,CAAC,EAAE,CAAC;YACH,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,oCAAoC,EAAE,KAAK,IAAI,EAAE;QACrD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE;YAC9C,SAAS,EAAE,GAAG;SACf,CAAC,EAAE,CAAC;YACH,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,KAAK,CAAC,qCAAqC,EAAE,KAAK,IAAI,EAAE;QACtD,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE;YAC9C,SAAS,EAAE,IAAI;SAChB,CAAC,EAAE,CAAC;YACH,UAAU;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEH,+EAA+E;AAC/E,yCAAyC;AACzC,+EAA+E;AAE/E,QAAQ,CAAC,8BAA8B,EAAE,GAAG,EAAE;IAC5C,MAAM,MAAM,GAAG,uBAAuB,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACjD,MAAM,QAAQ,GAAG,yBAAyB,CAAC,YAAY,CAAC,CAAC;IAEzD,KAAK,CAAC,8CAA8C,EAAE,KAAK,IAAI,EAAE;QAC/D,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC;YACjE,MAAM,CAAC,2BAA2B;QACpC,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,KAAK,CACH,+CAA+C,EAC/C,KAAK,IAAI,EAAE;QACT,MAAM,IAAI,GAAG,IAAI,oBAAoB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACtD,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE;YAC7C,SAAS,EAAE,EAAE;SACd,CAAC,EAAE,CAAC;YACH,IAAI,EAAE,KAAK,IAAI,EAAE;gBAAE,MAAM;QAC3B,CAAC;IACH,CAAC,CACF,CAAC;AACJ,CAAC,CAAC,CAAC"}

package/src/index.d.ts

@@ -0,0 +1,59 @@
+/**
+ * GLOST Processor
+ *
+ * Unified-style processor for GLOST documents.
+ *
+ * @packageDocumentation
+ *
+ * @example
+ * ```typescript
+ * import { glost } from "glost-processor";
+ * import { transcription } from "glost-transcription";
+ * import { translation } from "glost-translation";
+ *
+ * const processor = glost()
+ *   .use(transcription, { scheme: "ipa" })
+ *   .use(translation, { target: "en" })
+ *   .freeze();
+ *
+ * const result = await processor.process(document);
+ * ```
+ */
+export { GLOSTProcessor } from "./processor.js";
+export type { FrozenProcessor } from "./processor.js";
+export type { Plugin, PluginSpec, Preset, ProcessorOptions, ProcessingResult, ProcessingError, ProcessingWarning, ProcessingStats, BeforeHook, AfterHook, ErrorHook, SkipHook, ProgressHook, ProgressStats, } from "./types.js";
+export { GLOSTStreamProcessor } from "./stream-processor.js";
+export type { FrozenStreamProcessor, StreamOptions, ProcessedChunk, } from "./stream-processor.js";
+import { GLOSTProcessor } from "./processor.js";
+import type { ProcessorOptions } from "./types.js";
+/**
+ * Create a new GLOST processor
+ *
+ * Factory function for creating a new processor instance.
+ * Similar to `unified()` from the unified ecosystem.
+ *
+ * @param options - Initial processor options
+ * @returns A new processor instance
+ *
+ * @example
+ * ```typescript
+ * import { glost } from "glost-processor";
+ *
+ * const processor = glost()
+ *   .use(plugin1)
+ *   .use(plugin2)
+ *   .use(plugin3);
+ *
+ * const result = await processor.process(document);
+ * ```
+ *
+ * @example
+ * ```typescript
+ * // With options
+ * const processor = glost({ lenient: true, debug: true })
+ *   .use(plugin1)
+ *   .use(plugin2);
+ * ```
+ */
+export declare function glost(options?: ProcessorOptions): GLOSTProcessor;
+//# sourceMappingURL=index.d.ts.map

package/src/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAChD,YAAY,EAAE,eAAe,EAAE,MAAM,gBAAgB,CAAC;AACtD,YAAY,EACV,MAAM,EACN,UAAU,EACV,MAAM,EACN,gBAAgB,EAChB,gBAAgB,EAChB,eAAe,EACf,iBAAiB,EACjB,eAAe,EACf,UAAU,EACV,SAAS,EACT,SAAS,EACT,QAAQ,EACR,YAAY,EACZ,aAAa,GACd,MAAM,YAAY,CAAC;AACpB,OAAO,EAAE,oBAAoB,EAAE,MAAM,uBAAuB,CAAC;AAC7D,YAAY,EACV,qBAAqB,EACrB,aAAa,EACb,cAAc,GACf,MAAM,uBAAuB,CAAC;AAE/B,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAChD,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAEnD;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AACH,wBAAgB,KAAK,CAAC,OAAO,CAAC,EAAE,gBAAgB,GAAG,cAAc,CAEhE"}

package/src/index.js

@@ -0,0 +1,57 @@
+/**
+ * GLOST Processor
+ *
+ * Unified-style processor for GLOST documents.
+ *
+ * @packageDocumentation
+ *
+ * @example
+ * ```typescript
+ * import { glost } from "glost-processor";
+ * import { transcription } from "glost-transcription";
+ * import { translation } from "glost-translation";
+ *
+ * const processor = glost()
+ *   .use(transcription, { scheme: "ipa" })
+ *   .use(translation, { target: "en" })
+ *   .freeze();
+ *
+ * const result = await processor.process(document);
+ * ```
+ */
+export { GLOSTProcessor } from "./processor.js";
+export { GLOSTStreamProcessor } from "./stream-processor.js";
+import { GLOSTProcessor } from "./processor.js";
+/**
+ * Create a new GLOST processor
+ *
+ * Factory function for creating a new processor instance.
+ * Similar to `unified()` from the unified ecosystem.
+ *
+ * @param options - Initial processor options
+ * @returns A new processor instance
+ *
+ * @example
+ * ```typescript
+ * import { glost } from "glost-processor";
+ *
+ * const processor = glost()
+ *   .use(plugin1)
+ *   .use(plugin2)
+ *   .use(plugin3);
+ *
+ * const result = await processor.process(document);
+ * ```
+ *
+ * @example
+ * ```typescript
+ * // With options
+ * const processor = glost({ lenient: true, debug: true })
+ *   .use(plugin1)
+ *   .use(plugin2);
+ * ```
+ */
+export function glost(options) {
+    return new GLOSTProcessor(options);
+}
+//# sourceMappingURL=index.js.map

package/src/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAkBhD,OAAO,EAAE,oBAAoB,EAAE,MAAM,uBAAuB,CAAC;AAO7D,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAGhD;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AACH,MAAM,UAAU,KAAK,CAAC,OAA0B;IAC9C,OAAO,IAAI,cAAc,CAAC,OAAO,CAAC,CAAC;AACrC,CAAC"}

package/src/processor.d.ts

@@ -0,0 +1,256 @@
+/**
+ * GLOST Processor
+ *
+ * Unified-style processor for GLOST documents with fluent API.
+ * Similar to unified/remark processors.
+ *
+ * @packageDocumentation
+ */
+import type { GLOSTRoot } from "glost-core";
+import type { PluginSpec, Preset, ProcessorOptions, ProcessingResult, BeforeHook, AfterHook, ErrorHook, SkipHook, ProgressHook } from "./types.js";
+/**
+ * GLOST Processor
+ *
+ * Fluent API for processing GLOST documents through plugin pipelines.
+ *
+ * @example
+ * ```typescript
+ * import { GLOSTProcessor } from "glost-processor";
+ *
+ * const processor = new GLOSTProcessor()
+ *   .use(transcription, { scheme: "ipa" })
+ *   .use(translation, { target: "en" })
+ *   .use(frequency);
+ *
+ * const result = await processor.process(document);
+ * ```
+ */
+export declare class GLOSTProcessor {
+    private plugins;
+    private hooks;
+    private dataStore;
+    private options;
+    private frozen;
+    /**
+     * Create a new processor instance
+     *
+     * @param options - Initial processor options
+     */
+    constructor(options?: ProcessorOptions);
+    /**
+     * Use a plugin, preset, or extension
+     *
+     * @param spec - Plugin function, extension object, preset, or plugin ID
+     * @param options - Plugin options
+     * @returns This processor for chaining
+     *
+     * @example
+     * ```typescript
+     * processor
+     *   .use(transcription, { scheme: "ipa" })
+     *   .use(translation)
+     *   .use("frequency");
+     * ```
+     */
+    use(spec: PluginSpec | Preset, options?: any): this;
+    /**
+     * Apply a preset (collection of plugins)
+     *
+     * @param preset - Preset to apply
+     * @returns This processor for chaining
+     */
+    private usePreset;
+    /**
+     * Register a hook to run before a plugin
+     *
+     * @param pluginId - Plugin ID to hook into
+     * @param hook - Hook function to run
+     * @returns This processor for chaining
+     *
+     * @example
+     * ```typescript
+     * processor.before("translation", (doc) => {
+     *   console.log("About to translate");
+     * });
+     * ```
+     */
+    before(pluginId: string, hook: BeforeHook): this;
+    /**
+     * Register a hook to run after a plugin
+     *
+     * @param pluginId - Plugin ID to hook into
+     * @param hook - Hook function to run
+     * @returns This processor for chaining
+     *
+     * @example
+     * ```typescript
+     * processor.after("translation", (doc) => {
+     *   console.log("Translation complete");
+     * });
+     * ```
+     */
+    after(pluginId: string, hook: AfterHook): this;
+    /**
+     * Register an error handler
+     *
+     * @param hook - Error handler function
+     * @returns This processor for chaining
+     *
+     * @example
+     * ```typescript
+     * processor.onError((error, plugin) => {
+     *   console.error(`Plugin ${plugin} failed:`, error);
+     * });
+     * ```
+     */
+    onError(hook: ErrorHook): this;
+    /**
+     * Register a skip handler
+     *
+     * @param hook - Skip handler function
+     * @returns This processor for chaining
+     *
+     * @example
+     * ```typescript
+     * processor.onSkip((plugin, reason) => {
+     *   console.log(`Plugin ${plugin} skipped: ${reason}`);
+     * });
+     * ```
+     */
+    onSkip(hook: SkipHook): this;
+    /**
+     * Register a progress handler
+     *
+     * @param hook - Progress handler function
+     * @returns This processor for chaining
+     *
+     * @example
+     * ```typescript
+     * processor.onProgress((stats) => {
+     *   console.log(`Progress: ${stats.completed}/${stats.total}`);
+     * });
+     * ```
+     */
+    onProgress(hook: ProgressHook): this;
+    /**
+     * Set or get data in the processor data store
+     *
+     * @param key - Data key
+     * @param value - Data value (omit to get)
+     * @returns Value if getting, this processor if setting
+     *
+     * @example
+     * ```typescript
+     * processor.data("config", { theme: "dark" });
+     * const config = processor.data("config");
+     * ```
+     */
+    data(key: string): any;
+    data(key: string, value: any): this;
+    /**
+     * Freeze the processor
+     *
+     * Returns a frozen processor that cannot be modified.
+     * Useful for reusing the same configuration across multiple documents.
+     *
+     * @returns A frozen copy of this processor
+     *
+     * @example
+     * ```typescript
+     * const frozen = processor
+     *   .use(transcription)
+     *   .use(translation)
+     *   .freeze();
+     *
+     * // Can process multiple documents with the same pipeline
+     * const result1 = await frozen.process(doc1);
+     * const result2 = await frozen.process(doc2);
+     * ```
+     */
+    freeze(): FrozenProcessor;
+    /**
+     * Process a document through the pipeline
+     *
+     * @param document - GLOST document to process
+     * @returns Promise resolving to the processed document
+     *
+     * @example
+     * ```typescript
+     * const result = await processor.process(document);
+     * console.log(result);
+     * ```
+     */
+    process(document: GLOSTRoot): Promise<GLOSTRoot>;
+    /**
+     * Process a document and return detailed metadata
+     *
+     * @param document - GLOST document to process
+     * @returns Promise resolving to processing result with metadata
+     *
+     * @example
+     * ```typescript
+     * const result = await processor.processWithMeta(document);
+     * console.log(result.metadata.appliedPlugins);
+     * console.log(result.metadata.stats.totalTime);
+     * ```
+     */
+    processWithMeta(document: GLOSTRoot): Promise<ProcessingResult>;
+    /**
+     * Process a document synchronously (only if all plugins are sync)
+     *
+     * @param document - GLOST document to process
+     * @returns The processed document
+     * @throws {Error} If any plugin is async
+     *
+     * @example
+     * ```typescript
+     * const result = processor.processSync(document);
+     * ```
+     */
+    processSync(document: GLOSTRoot): GLOSTRoot;
+    /**
+     * Resolve all plugins to extensions
+     */
+    private resolveExtensions;
+    /**
+     * Resolve a single plugin to an extension
+     */
+    private resolvePlugin;
+    /**
+     * Check if a spec is a preset
+     */
+    private isPreset;
+    /**
+     * Run before hooks for a plugin
+     */
+    private runBeforeHooks;
+    /**
+     * Run after hooks for a plugin
+     */
+    private runAfterHooks;
+    /**
+     * Emit error to error handlers
+     */
+    private emitError;
+    /**
+     * Emit skip to skip handlers
+     */
+    private emitSkip;
+    /**
+     * Emit progress to progress handlers
+     */
+    private emitProgress;
+    /**
+     * Assert that the processor is not frozen
+     */
+    private assertNotFrozen;
+}
+/**
+ * Frozen processor type
+ *
+ * A frozen processor cannot be modified, only used for processing.
+ */
+export type FrozenProcessor = Omit<GLOSTProcessor, "use" | "before" | "after" | "onError" | "onSkip" | "onProgress" | "data" | "freeze"> & {
+    readonly frozen: true;
+};
+//# sourceMappingURL=processor.d.ts.map

package/src/processor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"processor.d.ts","sourceRoot":"","sources":["processor.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,YAAY,CAAC;AAG5C,OAAO,KAAK,EACV,UAAU,EACV,MAAM,EACN,gBAAgB,EAChB,gBAAgB,EAEhB,UAAU,EACV,SAAS,EACT,SAAS,EACT,QAAQ,EACR,YAAY,EAIb,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;GAgBG;AACH,qBAAa,cAAc;IACzB,OAAO,CAAC,OAAO,CAAkD;IACjE,OAAO,CAAC,KAAK,CAMX;IACF,OAAO,CAAC,SAAS,CAA+B;IAChD,OAAO,CAAC,OAAO,CAAwB;IACvC,OAAO,CAAC,MAAM,CAAS;IAEvB;;;;OAIG;gBACS,OAAO,GAAE,gBAAqB;IAO1C;;;;;;;;;;;;;;OAcG;IACH,GAAG,CAAC,IAAI,EAAE,UAAU,GAAG,MAAM,EAAE,OAAO,CAAC,EAAE,GAAG,GAAG,IAAI;IAanD;;;;;OAKG;IACH,OAAO,CAAC,SAAS;IAYjB;;;;;;;;;;;;;OAaG;IACH,MAAM,CAAC,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,UAAU,GAAG,IAAI;IAShD;;;;;;;;;;;;;OAaG;IACH,KAAK,CAAC,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,GAAG,IAAI;IAS9C;;;;;;;;;;;;OAYG;IACH,OAAO,CAAC,IAAI,EAAE,SAAS,GAAG,IAAI;IAM9B;;;;;;;;;;;;OAYG;IACH,MAAM,CAAC,IAAI,EAAE,QAAQ,GAAG,IAAI;IAM5B;;;;;;;;;;;;OAYG;IACH,UAAU,CAAC,IAAI,EAAE,YAAY,GAAG,IAAI;IAMpC;;;;;;;;;;;;OAYG;IACH,IAAI,CAAC,GAAG,EAAE,MAAM,GAAG,GAAG;IACtB,IAAI,CAAC,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,GAAG,GAAG,IAAI;IAUnC;;;;;;;;;;;;;;;;;;;OAmBG;IACH,MAAM,IAAI,eAAe;IAezB;;;;;;;;;;;OAWG;IACG,OAAO,CAAC,QAAQ,EAAE,SAAS,GAAG,OAAO,CAAC,SAAS,CAAC;IAKtD;;;;;;;;;;;;OAYG;IACG,eAAe,CAAC,QAAQ,EAAE,SAAS,GAAG,OAAO,CAAC,gBAAgB,CAAC;IAkHrE;;;;;;;;;;;OAWG;IACH,WAAW,CAAC,QAAQ,EAAE,SAAS,GAAG,SAAS;IAM3C;;OAEG;YACW,iBAAiB;IAa/B;;OAEG;YACW,aAAa;IAuB3B;;OAEG;IACH,OAAO,CAAC,QAAQ;IAShB;;OAEG;YACW,cAAc;IAO5B;;OAEG;YACW,aAAa;IAO3B;;OAEG;IACH,OAAO,CAAC,SAAS;IAUjB;;OAEG;IACH,OAAO,CAAC,QAAQ;IAUhB;;OAEG;IACH,OAAO,CAAC,YAAY;IAUpB;;OAEG;IACH,OAAO,CAAC,eAAe;CAKxB;AAED;;;;GAIG;AACH,MAAM,MAAM,eAAe,GAAG,IAAI,CAChC,cAAc,EACd,KAAK,GAAG,QAAQ,GAAG,OAAO,GAAG,SAAS,GAAG,QAAQ,GAAG,YAAY,GAAG,MAAM,GAAG,QAAQ,CACrF,GAAG;IACF,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC;CACvB,CAAC"}