@stabrise/scaledp 0.1.0

package/dist/display/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","names":[],"sources":["../../src/display/index.ts"],"sourcesContent":["/**\n * Result rendering, mirroring ScaleDP's notebook display helpers.\n *\n * Python monkey-patches `show_image`, `show_text`, `show_ner` and\n * `visualize_ner` onto the Spark DataFrame and renders Jinja templates into\n * IPython. The browser equivalent is to build DOM elements, so these return an\n * `HTMLElement` the caller places wherever it likes -- and `renderInto` is the\n * one-liner for the common case.\n *\n * Everything is built with text nodes rather than interpolated markup:\n * recognized text and entity words come straight from the document, and a page\n * containing `<` would otherwise corrupt the DOM.\n */\n\nimport type { Box } from '../schemas/box.js'\nimport type { DetectorOutput } from '../schemas/detector-output.js'\nimport type { Document } from '../schemas/document.js'\nimport type { Entity, NerOutput } from '../schemas/entity.js'\nimport type { ScaleDpImage } from '../schemas/image.js'\n\n/**\n * A stable colour per entity group.\n *\n * Python picks a random colour each run, so two renders of the same document\n * never match. Hashing the group name keeps 'PERSON' one colour everywhere.\n */\nexport function colorForGroup(name: string): string {\n    let hash = 0\n    for (let i = 0; i < name.length; i++) hash = (hash * 31 + name.charCodeAt(i)) | 0\n    return `hsl(${Math.abs(hash) % 360}, 70%, 45%)`\n}\n\nfunction element<K extends keyof HTMLElementTagNameMap>(\n    tag: K,\n    props: { text?: string; className?: string; style?: Partial<CSSStyleDeclaration> } = {}\n): HTMLElementTagNameMap[K] {\n    const node = document.createElement(tag)\n    if (props.text !== undefined) node.textContent = props.text\n    if (props.className) node.className = props.className\n    if (props.style) Object.assign(node.style, props.style)\n    return node\n}\n\n/** Replace a container's contents with `node`. Accepts a selector or element. */\nexport function renderInto(target: string | HTMLElement, node: Node): HTMLElement {\n    const host = typeof target === 'string' ? document.querySelector<HTMLElement>(target) : target\n    if (!host) throw new Error(`No element matches ${String(target)}`)\n    host.replaceChildren(node)\n    return host\n}\n\nexport interface ShowImageOptions {\n    /** CSS width, e.g. '600px' or '100%'. */\n    width?: string\n    alt?: string\n}\n\n/**\n * An `<img>` for a ScaleDP image. Mirrors `show_image`.\n *\n * The object URL is revoked once the image has decoded -- holding one per page\n * leaks the whole blob for the lifetime of the document.\n */\nexport function showImage(image: ScaleDpImage, options: ShowImageOptions = {}): HTMLElement {\n    if (image.exception) return errorBlock(image.exception)\n\n    const bytes = new Uint8Array(image.data.byteLength)\n    bytes.set(image.data)\n    const url = URL.createObjectURL(new Blob([bytes.buffer], { type: `image/${image.imageType}` }))\n\n    const img = element('img', { style: { maxWidth: options.width ?? '100%', height: 'auto' } })\n    img.alt = options.alt ?? image.path\n    img.addEventListener('load', () => URL.revokeObjectURL(url), { once: true })\n    img.addEventListener('error', () => URL.revokeObjectURL(url), { once: true })\n    img.src = url\n    return img\n}\n\nexport interface ShowTextOptions {\n    /**\n     * Preserve the document's own layout. Correct when the OCR stage ran with\n     * `keepFormatting`, which encodes the layout in spaces and blank lines.\n     */\n    preserveLayout?: boolean\n    maxHeight?: string\n}\n\n/** A `<pre>` of the recognized text. Mirrors `show_text`. */\nexport function showText(document_: Document, options: ShowTextOptions = {}): HTMLElement {\n    if (document_.exception) return errorBlock(document_.exception)\n\n    return element('pre', {\n        text: document_.text,\n        className: 'scaledp-text',\n        style: {\n            fontFamily: 'ui-monospace, SFMono-Regular, Menlo, monospace',\n            fontSize: '12px',\n            whiteSpace: options.preserveLayout === false ? 'pre-wrap' : 'pre',\n            overflowX: 'auto',\n            maxHeight: options.maxHeight ?? '30rem',\n            margin: '0',\n        },\n    })\n}\n\n/** Pretty-printed JSON. Mirrors `show_json`. */\nexport function showJson(value: unknown, indent = 2): HTMLElement {\n    const text = typeof value === 'string' ? value : JSON.stringify(value, null, indent)\n    return element('pre', {\n        text,\n        className: 'scaledp-json',\n        style: { fontFamily: 'ui-monospace, monospace', fontSize: '12px', overflowX: 'auto' },\n    })\n}\n\nexport interface ShowNerOptions {\n    /** Maximum rows; 0 shows all. Python defaults to 20. */\n    limit?: number\n    /** Only these groups. */\n    whiteList?: readonly string[]\n}\n\n/** A table of entities. Mirrors `show_ner`. */\nexport function showNer(ner: NerOutput, options: ShowNerOptions = {}): HTMLElement {\n    if (ner.exception) return errorBlock(ner.exception)\n\n    const allowed = new Set(options.whiteList ?? [])\n    let entities = allowed.size > 0 ? ner.entities.filter((e) => allowed.has(e.entity_group)) : ner.entities\n\n    const limit = options.limit ?? 20\n    const total = entities.length\n    if (limit > 0) entities = entities.slice(0, limit)\n\n    if (total === 0) return element('p', { text: 'No entities found.' })\n\n    const table = element('table', { className: 'scaledp-ner' })\n    table.style.borderCollapse = 'collapse'\n\n    const header = table.insertRow()\n    for (const label of ['Type', 'Text', 'Score', 'Start', 'End', 'Boxes']) {\n        const th = document.createElement('th')\n        th.textContent = label\n        th.style.cssText = 'border:1px solid #ddd;padding:4px 8px;text-align:left'\n        header.append(th)\n    }\n\n    for (const entity of entities) {\n        const tr = table.insertRow()\n        const swatch = element('span', {\n            style: {\n                display: 'inline-block',\n                width: '8px',\n                height: '8px',\n                borderRadius: '50%',\n                marginRight: '6px',\n                background: colorForGroup(entity.entity_group),\n            },\n        })\n        const cells: (string | Node)[] = [\n            entity.entity_group,\n            entity.word,\n            entity.score.toFixed(3),\n            String(entity.start),\n            String(entity.end),\n            String(entity.boxes.length),\n        ]\n        cells.forEach((value, index) => {\n            const cell = tr.insertCell()\n            cell.style.cssText = 'border:1px solid #ddd;padding:4px 8px'\n            if (index === 0) cell.append(swatch)\n            cell.append(typeof value === 'string' ? document.createTextNode(value) : value)\n        })\n    }\n\n    const wrapper = element('div')\n    wrapper.append(table)\n    if (limit > 0 && total > limit) {\n        wrapper.append(element('p', { text: `Showing ${limit} of ${total} entities.` }))\n    }\n    return wrapper\n}\n\nexport interface VisualizeNerOptions {\n    /** Only highlight these groups. */\n    labelsList?: readonly string[]\n    /** Render the group name beside each highlight. */\n    showLabels?: boolean\n}\n\n/**\n * The document text with entities highlighted inline. Mirrors `visualize_ner`.\n *\n * Splices spans by character offset, which is exactly what `Entity.start`/`end`\n * index. Overlapping entities are dropped rather than nested: the highest-\n * scoring one wins, because two spans cannot occupy the same characters in a\n * flat text run.\n */\nexport function visualizeNer(\n    document_: Document,\n    ner: NerOutput,\n    options: VisualizeNerOptions = {}\n): HTMLElement {\n    if (document_.exception) return errorBlock(document_.exception)\n    if (ner.exception) return errorBlock(ner.exception)\n\n    const allowed = new Set(options.labelsList ?? [])\n    const entities = (\n        allowed.size > 0 ? ner.entities.filter((e) => allowed.has(e.entity_group)) : ner.entities\n    )\n        .filter((e) => e.start >= 0 && e.end > e.start && e.end <= document_.text.length)\n        .sort((a, b) => a.start - b.start || b.score - a.score)\n\n    const container = element('div', {\n        className: 'scaledp-ner-text',\n        style: {\n            fontFamily: 'ui-monospace, SFMono-Regular, Menlo, monospace',\n            fontSize: '12px',\n            whiteSpace: 'pre-wrap',\n            lineHeight: '1.9',\n        },\n    })\n\n    let cursor = 0\n    for (const entity of entities) {\n        if (entity.start < cursor) continue // overlaps an already-rendered span\n\n        if (entity.start > cursor) {\n            container.append(document.createTextNode(document_.text.slice(cursor, entity.start)))\n        }\n\n        const color = colorForGroup(entity.entity_group)\n        const mark = element('span', {\n            text: document_.text.slice(entity.start, entity.end),\n            style: {\n                background: color,\n                color: '#fff',\n                borderRadius: '3px',\n                padding: '1px 3px',\n            },\n        })\n        mark.title = `${entity.entity_group} (${entity.score.toFixed(3)})`\n        container.append(mark)\n\n        if (options.showLabels) {\n            container.append(\n                element('sup', {\n                    text: entity.entity_group,\n                    style: { color, fontSize: '9px', marginLeft: '2px' },\n                })\n            )\n        }\n        cursor = entity.end\n    }\n    container.append(document.createTextNode(document_.text.slice(cursor)))\n    return container\n}\n\n/** A summary table of detected boxes. */\nexport function showBoxes(output: DetectorOutput | Document, limit = 20): HTMLElement {\n    if (output.exception) return errorBlock(output.exception)\n\n    const boxes: Box[] = output.bboxes\n    const table = element('table')\n    table.style.borderCollapse = 'collapse'\n\n    const header = table.insertRow()\n    for (const label of ['Text', 'Score', 'x', 'y', 'w', 'h', 'angle']) {\n        const th = document.createElement('th')\n        th.textContent = label\n        th.style.cssText = 'border:1px solid #ddd;padding:4px 8px;text-align:left'\n        header.append(th)\n    }\n\n    for (const box of limit > 0 ? boxes.slice(0, limit) : boxes) {\n        const tr = table.insertRow()\n        for (const value of [\n            box.text,\n            box.score.toFixed(3),\n            String(box.x),\n            String(box.y),\n            String(box.width),\n            String(box.height),\n            box.angle.toFixed(1),\n        ]) {\n            const cell = tr.insertCell()\n            cell.style.cssText = 'border:1px solid #ddd;padding:4px 8px'\n            cell.textContent = value\n        }\n    }\n\n    const wrapper = element('div')\n    wrapper.append(table)\n    if (limit > 0 && boxes.length > limit) {\n        wrapper.append(element('p', { text: `Showing ${limit} of ${boxes.length} boxes.` }))\n    }\n    return wrapper\n}\n\nfunction errorBlock(message: string): HTMLElement {\n    return element('pre', {\n        text: message,\n        style: {\n            color: '#b00020',\n            whiteSpace: 'pre-wrap',\n            fontFamily: 'ui-monospace, monospace',\n            fontSize: '12px',\n        },\n    })\n}\n\nexport type { DetectorOutput, Document, Entity, NerOutput, ScaleDpImage }\n"],"mappings":";;;;;;;AA0BA,SAAgB,cAAc,MAAsB;CAChD,IAAI,OAAO;CACX,KAAK,IAAI,IAAI,GAAG,IAAI,KAAK,QAAQ,KAAK,OAAQ,OAAO,KAAK,KAAK,WAAW,CAAC,IAAK;CAChF,OAAO,OAAO,KAAK,IAAI,IAAI,IAAI,IAAI;AACvC;AAEA,SAAS,QACL,KACA,QAAqF,CAAC,GAC9D;CACxB,MAAM,OAAO,SAAS,cAAc,GAAG;CACvC,IAAI,MAAM,SAAS,KAAA,GAAW,KAAK,cAAc,MAAM;CACvD,IAAI,MAAM,WAAW,KAAK,YAAY,MAAM;CAC5C,IAAI,MAAM,OAAO,OAAO,OAAO,KAAK,OAAO,MAAM,KAAK;CACtD,OAAO;AACX;;AAGA,SAAgB,WAAW,QAA8B,MAAyB;CAC9E,MAAM,OAAO,OAAO,WAAW,WAAW,SAAS,cAA2B,MAAM,IAAI;CACxF,IAAI,CAAC,MAAM,MAAM,IAAI,MAAM,sBAAsB,OAAO,MAAM,GAAG;CACjE,KAAK,gBAAgB,IAAI;CACzB,OAAO;AACX;;;;;;;AAcA,SAAgB,UAAU,OAAqB,UAA4B,CAAC,GAAgB;CACxF,IAAI,MAAM,WAAW,OAAO,WAAW,MAAM,SAAS;CAEtD,MAAM,QAAQ,IAAI,WAAW,MAAM,KAAK,UAAU;CAClD,MAAM,IAAI,MAAM,IAAI;CACpB,MAAM,MAAM,IAAI,gBAAgB,IAAI,KAAK,CAAC,MAAM,MAAM,GAAG,EAAE,MAAM,SAAS,MAAM,YAAY,CAAC,CAAC;CAE9F,MAAM,MAAM,QAAQ,OAAO,EAAE,OAAO;EAAE,UAAU,QAAQ,SAAS;EAAQ,QAAQ;CAAO,EAAE,CAAC;CAC3F,IAAI,MAAM,QAAQ,OAAO,MAAM;CAC/B,IAAI,iBAAiB,cAAc,IAAI,gBAAgB,GAAG,GAAG,EAAE,MAAM,KAAK,CAAC;CAC3E,IAAI,iBAAiB,eAAe,IAAI,gBAAgB,GAAG,GAAG,EAAE,MAAM,KAAK,CAAC;CAC5E,IAAI,MAAM;CACV,OAAO;AACX;;AAYA,SAAgB,SAAS,WAAqB,UAA2B,CAAC,GAAgB;CACtF,IAAI,UAAU,WAAW,OAAO,WAAW,UAAU,SAAS;CAE9D,OAAO,QAAQ,OAAO;EAClB,MAAM,UAAU;EAChB,WAAW;EACX,OAAO;GACH,YAAY;GACZ,UAAU;GACV,YAAY,QAAQ,mBAAmB,QAAQ,aAAa;GAC5D,WAAW;GACX,WAAW,QAAQ,aAAa;GAChC,QAAQ;EACZ;CACJ,CAAC;AACL;;AAGA,SAAgB,SAAS,OAAgB,SAAS,GAAgB;CAE9D,OAAO,QAAQ,OAAO;EAClB,MAFS,OAAO,UAAU,WAAW,QAAQ,KAAK,UAAU,OAAO,MAAM,MAAM;EAG/E,WAAW;EACX,OAAO;GAAE,YAAY;GAA2B,UAAU;GAAQ,WAAW;EAAO;CACxF,CAAC;AACL;;AAUA,SAAgB,QAAQ,KAAgB,UAA0B,CAAC,GAAgB;CAC/E,IAAI,IAAI,WAAW,OAAO,WAAW,IAAI,SAAS;CAElD,MAAM,UAAU,IAAI,IAAI,QAAQ,aAAa,CAAC,CAAC;CAC/C,IAAI,WAAW,QAAQ,OAAO,IAAI,IAAI,SAAS,QAAQ,MAAM,QAAQ,IAAI,EAAE,YAAY,CAAC,IAAI,IAAI;CAEhG,MAAM,QAAQ,QAAQ,SAAS;CAC/B,MAAM,QAAQ,SAAS;CACvB,IAAI,QAAQ,GAAG,WAAW,SAAS,MAAM,GAAG,KAAK;CAEjD,IAAI,UAAU,GAAG,OAAO,QAAQ,KAAK,EAAE,MAAM,qBAAqB,CAAC;CAEnE,MAAM,QAAQ,QAAQ,SAAS,EAAE,WAAW,cAAc,CAAC;CAC3D,MAAM,MAAM,iBAAiB;CAE7B,MAAM,SAAS,MAAM,UAAU;CAC/B,KAAK,MAAM,SAAS;EAAC;EAAQ;EAAQ;EAAS;EAAS;EAAO;CAAO,GAAG;EACpE,MAAM,KAAK,SAAS,cAAc,IAAI;EACtC,GAAG,cAAc;EACjB,GAAG,MAAM,UAAU;EACnB,OAAO,OAAO,EAAE;CACpB;CAEA,KAAK,MAAM,UAAU,UAAU;EAC3B,MAAM,KAAK,MAAM,UAAU;EAC3B,MAAM,SAAS,QAAQ,QAAQ,EAC3B,OAAO;GACH,SAAS;GACT,OAAO;GACP,QAAQ;GACR,cAAc;GACd,aAAa;GACb,YAAY,cAAc,OAAO,YAAY;EACjD,EACJ,CAAC;EASD;GAPI,OAAO;GACP,OAAO;GACP,OAAO,MAAM,QAAQ,CAAC;GACtB,OAAO,OAAO,KAAK;GACnB,OAAO,OAAO,GAAG;GACjB,OAAO,OAAO,MAAM,MAAM;EAE1B,CAAC,CAAC,SAAS,OAAO,UAAU;GAC5B,MAAM,OAAO,GAAG,WAAW;GAC3B,KAAK,MAAM,UAAU;GACrB,IAAI,UAAU,GAAG,KAAK,OAAO,MAAM;GACnC,KAAK,OAAO,OAAO,UAAU,WAAW,SAAS,eAAe,KAAK,IAAI,KAAK;EAClF,CAAC;CACL;CAEA,MAAM,UAAU,QAAQ,KAAK;CAC7B,QAAQ,OAAO,KAAK;CACpB,IAAI,QAAQ,KAAK,QAAQ,OACrB,QAAQ,OAAO,QAAQ,KAAK,EAAE,MAAM,WAAW,MAAM,MAAM,MAAM,YAAY,CAAC,CAAC;CAEnF,OAAO;AACX;;;;;;;;;AAiBA,SAAgB,aACZ,WACA,KACA,UAA+B,CAAC,GACrB;CACX,IAAI,UAAU,WAAW,OAAO,WAAW,UAAU,SAAS;CAC9D,IAAI,IAAI,WAAW,OAAO,WAAW,IAAI,SAAS;CAElD,MAAM,UAAU,IAAI,IAAI,QAAQ,cAAc,CAAC,CAAC;CAChD,MAAM,YACF,QAAQ,OAAO,IAAI,IAAI,SAAS,QAAQ,MAAM,QAAQ,IAAI,EAAE,YAAY,CAAC,IAAI,IAAI,SAAA,CAEhF,QAAQ,MAAM,EAAE,SAAS,KAAK,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,UAAU,KAAK,MAAM,CAAC,CAChF,MAAM,GAAG,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,EAAE,KAAK;CAE1D,MAAM,YAAY,QAAQ,OAAO;EAC7B,WAAW;EACX,OAAO;GACH,YAAY;GACZ,UAAU;GACV,YAAY;GACZ,YAAY;EAChB;CACJ,CAAC;CAED,IAAI,SAAS;CACb,KAAK,MAAM,UAAU,UAAU;EAC3B,IAAI,OAAO,QAAQ,QAAQ;EAE3B,IAAI,OAAO,QAAQ,QACf,UAAU,OAAO,SAAS,eAAe,UAAU,KAAK,MAAM,QAAQ,OAAO,KAAK,CAAC,CAAC;EAGxF,MAAM,QAAQ,cAAc,OAAO,YAAY;EAC/C,MAAM,OAAO,QAAQ,QAAQ;GACzB,MAAM,UAAU,KAAK,MAAM,OAAO,OAAO,OAAO,GAAG;GACnD,OAAO;IACH,YAAY;IACZ,OAAO;IACP,cAAc;IACd,SAAS;GACb;EACJ,CAAC;EACD,KAAK,QAAQ,GAAG,OAAO,aAAa,IAAI,OAAO,MAAM,QAAQ,CAAC,EAAE;EAChE,UAAU,OAAO,IAAI;EAErB,IAAI,QAAQ,YACR,UAAU,OACN,QAAQ,OAAO;GACX,MAAM,OAAO;GACb,OAAO;IAAE;IAAO,UAAU;IAAO,YAAY;GAAM;EACvD,CAAC,CACL;EAEJ,SAAS,OAAO;CACpB;CACA,UAAU,OAAO,SAAS,eAAe,UAAU,KAAK,MAAM,MAAM,CAAC,CAAC;CACtE,OAAO;AACX;;AAGA,SAAgB,UAAU,QAAmC,QAAQ,IAAiB;CAClF,IAAI,OAAO,WAAW,OAAO,WAAW,OAAO,SAAS;CAExD,MAAM,QAAe,OAAO;CAC5B,MAAM,QAAQ,QAAQ,OAAO;CAC7B,MAAM,MAAM,iBAAiB;CAE7B,MAAM,SAAS,MAAM,UAAU;CAC/B,KAAK,MAAM,SAAS;EAAC;EAAQ;EAAS;EAAK;EAAK;EAAK;EAAK;CAAO,GAAG;EAChE,MAAM,KAAK,SAAS,cAAc,IAAI;EACtC,GAAG,cAAc;EACjB,GAAG,MAAM,UAAU;EACnB,OAAO,OAAO,EAAE;CACpB;CAEA,KAAK,MAAM,OAAO,QAAQ,IAAI,MAAM,MAAM,GAAG,KAAK,IAAI,OAAO;EACzD,MAAM,KAAK,MAAM,UAAU;EAC3B,KAAK,MAAM,SAAS;GAChB,IAAI;GACJ,IAAI,MAAM,QAAQ,CAAC;GACnB,OAAO,IAAI,CAAC;GACZ,OAAO,IAAI,CAAC;GACZ,OAAO,IAAI,KAAK;GAChB,OAAO,IAAI,MAAM;GACjB,IAAI,MAAM,QAAQ,CAAC;EACvB,GAAG;GACC,MAAM,OAAO,GAAG,WAAW;GAC3B,KAAK,MAAM,UAAU;GACrB,KAAK,cAAc;EACvB;CACJ;CAEA,MAAM,UAAU,QAAQ,KAAK;CAC7B,QAAQ,OAAO,KAAK;CACpB,IAAI,QAAQ,KAAK,MAAM,SAAS,OAC5B,QAAQ,OAAO,QAAQ,KAAK,EAAE,MAAM,WAAW,MAAM,MAAM,MAAM,OAAO,SAAS,CAAC,CAAC;CAEvF,OAAO;AACX;AAEA,SAAS,WAAW,SAA8B;CAC9C,OAAO,QAAQ,OAAO;EAClB,MAAM;EACN,OAAO;GACH,OAAO;GACP,YAAY;GACZ,YAAY;GACZ,UAAU;EACd;CACJ,CAAC;AACL"}

package/dist/document-B8I61TiY.d.ts

@@ -0,0 +1,16 @@
+import { n as Box } from "./box-DAfzwfhA.js";
+//#region src/schemas/document.d.ts
+interface Document {
+  path: string;
+  text: string;
+  /** Producer of this document: 'text' | 'ocr' | 'pdf' | an engine name. */
+  type: string;
+  bboxes: Box[];
+  exception: string;
+}
+declare function createDocument(init?: Partial<Document>): Document;
+/** Python's `Document.merge` puts the *argument* first and joins with a newline. */
+declare function mergeDocuments(self: Document, other: Document): Document;
+//#endregion
+export { createDocument as n, mergeDocuments as r, Document as t };
+//# sourceMappingURL=document-B8I61TiY.d.ts.map

package/dist/entity-CedtRhU1.d.ts

@@ -0,0 +1,22 @@
+import { n as Box } from "./box-DAfzwfhA.js";
+//#region src/schemas/entity.d.ts
+interface Entity {
+  entity_group: string;
+  score: number;
+  word: string;
+  /** Character offset into the source document text. */
+  start: number;
+  end: number;
+  /** Boxes the entity's characters fall inside; empty when there is no OCR layer. */
+  boxes: Box[];
+}
+interface NerOutput {
+  path: string;
+  entities: Entity[];
+  exception: string;
+  json: string;
+}
+declare function createNerOutput(init?: Partial<NerOutput>): NerOutput;
+//#endregion
+export { NerOutput as n, createNerOutput as r, Entity as t };
+//# sourceMappingURL=entity-CedtRhU1.d.ts.map

package/dist/entity-D6Hxaugj.js

@@ -0,0 +1,13 @@
+//#region src/schemas/entity.ts
+function createNerOutput(init = {}) {
+	return {
+		path: init.path ?? "memory",
+		entities: init.entities ?? [],
+		exception: init.exception ?? "",
+		json: init.json ?? ""
+	};
+}
+//#endregion
+export { createNerOutput as t };
+
+//# sourceMappingURL=entity-D6Hxaugj.js.map

package/dist/entity-D6Hxaugj.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"entity-D6Hxaugj.js","names":[],"sources":["../src/schemas/entity.ts"],"sourcesContent":["/** Port of `scaledp/schemas/Entity.py` and `NerOutput.py`. */\n\nimport type { Box } from './box.js'\n\nexport interface Entity {\n    entity_group: string\n    score: number\n    word: string\n    /** Character offset into the source document text. */\n    start: number\n    end: number\n    /** Boxes the entity's characters fall inside; empty when there is no OCR layer. */\n    boxes: Box[]\n}\n\nexport interface NerOutput {\n    path: string\n    entities: Entity[]\n    exception: string\n    json: string\n}\n\nexport function createNerOutput(init: Partial<NerOutput> = {}): NerOutput {\n    return {\n        path: init.path ?? 'memory',\n        entities: init.entities ?? [],\n        exception: init.exception ?? '',\n        json: init.json ?? '',\n    }\n}\n"],"mappings":";AAsBA,SAAgB,gBAAgB,OAA2B,CAAC,GAAc;CACtE,OAAO;EACH,MAAM,KAAK,QAAQ;EACnB,UAAU,KAAK,YAAY,CAAC;EAC5B,WAAW,KAAK,aAAa;EAC7B,MAAM,KAAK,QAAQ;CACvB;AACJ"}

package/dist/image-CAH2rLv9.js

@@ -0,0 +1,511 @@
+//#region src/core/geometry.ts
+const EPSILON = 1e-9;
+/** Cross product of (o->a) and (o->b). > 0 means counter-clockwise. */
+function cross(o, a, b) {
+	return (a[0] - o[0]) * (b[1] - o[1]) - (a[1] - o[1]) * (b[0] - o[0]);
+}
+/**
+* Monotone-chain convex hull. Returns hull points counter-clockwise in a
+* y-down image coordinate system, without the duplicated closing point.
+*/
+function convexHull(points) {
+	const pts = [...points].sort((p, q) => p[0] === q[0] ? p[1] - q[1] : p[0] - q[0]);
+	const uniq = [];
+	for (const p of pts) {
+		const last = uniq[uniq.length - 1];
+		if (!last || last[0] !== p[0] || last[1] !== p[1]) uniq.push(p);
+	}
+	if (uniq.length < 3) return uniq;
+	const build = (source) => {
+		const chain = [];
+		for (const p of source) {
+			while (chain.length >= 2 && cross(chain[chain.length - 2], chain[chain.length - 1], p) <= 0) chain.pop();
+			chain.push(p);
+		}
+		chain.pop();
+		return chain;
+	};
+	return [...build(uniq), ...build([...uniq].reverse())];
+}
+/**
+* Minimum-area enclosing rectangle via rotating calipers.
+*
+* The minimum-area rectangle always has a side flush with a convex-hull edge,
+* so testing one orientation per hull edge is exhaustive.
+*/
+function minAreaRect(points) {
+	if (points.length === 0) return {
+		center: [0, 0],
+		size: [0, 0],
+		angle: 0
+	};
+	const hull = convexHull(points);
+	if (hull.length < 2) {
+		const p = hull[0] ?? points[0];
+		return {
+			center: [p[0], p[1]],
+			size: [0, 0],
+			angle: 0
+		};
+	}
+	let best = null;
+	for (let i = 0; i < hull.length; i++) {
+		const a = hull[i];
+		const b = hull[(i + 1) % hull.length];
+		const len = Math.hypot(b[0] - a[0], b[1] - a[1]);
+		if (len < EPSILON) continue;
+		const ux = (b[0] - a[0]) / len;
+		const uy = (b[1] - a[1]) / len;
+		const vx = -uy;
+		const vy = ux;
+		let minU = Infinity;
+		let maxU = -Infinity;
+		let minV = Infinity;
+		let maxV = -Infinity;
+		for (const p of hull) {
+			const pu = p[0] * ux + p[1] * uy;
+			const pv = p[0] * vx + p[1] * vy;
+			if (pu < minU) minU = pu;
+			if (pu > maxU) maxU = pu;
+			if (pv < minV) minV = pv;
+			if (pv > maxV) maxV = pv;
+		}
+		const w = maxU - minU;
+		const h = maxV - minV;
+		const area = w * h;
+		if (best === null || area < best.area) {
+			const cu = (minU + maxU) / 2;
+			const cv = (minV + maxV) / 2;
+			best = {
+				area,
+				center: [cu * ux + cv * vx, cu * uy + cv * vy],
+				w,
+				h,
+				angle: Math.atan2(uy, ux) * 180 / Math.PI
+			};
+		}
+	}
+	if (best === null) return {
+		center: [0, 0],
+		size: [0, 0],
+		angle: 0
+	};
+	let { angle, w, h } = best;
+	while (angle <= 0) {
+		angle += 90;
+		[w, h] = [h, w];
+	}
+	while (angle > 90) {
+		angle -= 90;
+		[w, h] = [h, w];
+	}
+	return {
+		center: best.center,
+		size: [w, h],
+		angle
+	};
+}
+/**
+* The 4 corners of a rotated rect, in cv2.boxPoints order.
+* For an upright rect in image coordinates that is: BL, TL, TR, BR.
+*/
+function boxPoints(rect) {
+	const [cx, cy] = rect.center;
+	const [w, h] = rect.size;
+	const rad = rect.angle * Math.PI / 180;
+	const b = Math.cos(rad) * .5;
+	const a = Math.sin(rad) * .5;
+	const p0 = [cx - a * h - b * w, cy + b * h - a * w];
+	const p1 = [cx + a * h - b * w, cy - b * h - a * w];
+	return [
+		p0,
+		p1,
+		[2 * cx - p0[0], 2 * cy - p0[1]],
+		[2 * cx - p1[0], 2 * cy - p1[1]]
+	];
+}
+/** Shoelace area of a simple polygon; always non-negative. */
+function polygonArea(points) {
+	let sum = 0;
+	for (let i = 0; i < points.length; i++) {
+		const a = points[i];
+		const b = points[(i + 1) % points.length];
+		sum += a[0] * b[1] - b[0] * a[1];
+	}
+	return Math.abs(sum) / 2;
+}
+/** Perimeter of a closed polygon. */
+function polygonPerimeter(points) {
+	let sum = 0;
+	for (let i = 0; i < points.length; i++) {
+		const a = points[i];
+		const b = points[(i + 1) % points.length];
+		sum += Math.hypot(b[0] - a[0], b[1] - a[1]);
+	}
+	return sum;
+}
+//#endregion
+//#region src/schemas/box.ts
+/**
+* Port of `scaledp/schemas/Box.py`.
+*
+* A Box is NOT xyxy and NOT a polygon. `x`/`y` is the top-left of the
+* *axis-aligned* box of the same size centred on the rotated rect's centre,
+* and `angle` is degrees about that same centre. `width` is always the longer
+* side. Getting this wrong silently shifts every downstream consumer, so the
+* conversions below mirror the Python implementation exactly.
+*/
+/** `abs(angle) >= 3` — matches Python's `is_rotated`, which tolerates OCR jitter. */
+const ROTATION_EPSILON_DEGREES = 3;
+function createBox(init = {}) {
+	return {
+		text: init.text ?? "",
+		score: init.score ?? 0,
+		x: init.x ?? 0,
+		y: init.y ?? 0,
+		width: init.width ?? 0,
+		height: init.height ?? 0,
+		angle: init.angle ?? 0
+	};
+}
+function isRotated(box) {
+	return Math.abs(box.angle) >= 3;
+}
+function bbox(box, padding = 0) {
+	return [
+		box.x - padding,
+		box.y - padding,
+		box.x + box.width + padding,
+		box.y + box.height + padding
+	];
+}
+function shape(box, padding = 0) {
+	const [x0, y0, x1, y1] = bbox(box, padding);
+	return [[x0, y0], [x1, y1]];
+}
+/**
+* Scale a box, then apply padding.
+*
+* Note the padding is asymmetric, and deliberately so: Python subtracts it from
+* the origin and adds it to the size, so the box grows by `padding` on its left
+* and top while its right and bottom edges stay put. Growing all four sides
+* would need `padding * 2` on the size.
+*/
+function scaleBox(box, factor, padding = 0) {
+	return {
+		...box,
+		x: Math.round(box.x * factor - padding),
+		y: Math.round(box.y * factor - padding),
+		width: Math.round(box.width * factor + padding),
+		height: Math.round(box.height * factor + padding)
+	};
+}
+function boxFromBBox(box, opts = {}) {
+	const [x0, y0, x1, y1] = box;
+	return {
+		text: opts.text ?? "",
+		score: opts.score ?? 0,
+		x: Math.round(x0),
+		y: Math.round(y0),
+		width: Math.round(x1 - x0),
+		height: Math.round(y1 - y0),
+		angle: opts.angle ?? 0
+	};
+}
+/**
+* Build a Box from exactly 4 polygon points — port of Python `Box.from_polygon`.
+*
+* `width` is forced to be the longer side (subtracting 90 degrees from the angle
+* to compensate), then the angle is normalised to (-90, 270]. `x`/`y` are derived
+* from the centre, NOT from the polygon's bounding box.
+*/
+function boxFromPolygon(points, opts = {}) {
+	if (points.length !== 4) throw new Error(`boxFromPolygon expects exactly 4 points, received ${points.length}`);
+	const padding = opts.padding ?? 0;
+	const rect = minAreaRect(points);
+	const [cx, cy] = rect.center;
+	let [width, height] = rect.size;
+	let angle = rect.angle;
+	if (width < height) {
+		[width, height] = [height, width];
+		angle -= 90;
+	}
+	angle = (angle % 360 + 360) % 360;
+	if (angle > 270) angle -= 360;
+	width = Math.max(1, Math.round(width) + padding * 2);
+	height = Math.max(1, Math.round(height) + padding * 2);
+	return {
+		text: opts.text ?? "",
+		score: opts.score ?? 1,
+		x: Math.round(cx - width / 2),
+		y: Math.round(cy - height / 2),
+		width,
+		height,
+		angle
+	};
+}
+/** Axis-aligned intersection-over-union. Ignores `angle`, exactly as Python does. */
+function boxIou(a, b) {
+	const [ax0, ay0, ax1, ay1] = bbox(a);
+	const [bx0, by0, bx1, by1] = bbox(b);
+	const ix = Math.min(ax1, bx1) - Math.max(ax0, bx0);
+	const iy = Math.min(ay1, by1) - Math.max(ay0, by0);
+	if (ix <= 0 || iy <= 0) return 0;
+	const intersection = ix * iy;
+	const union = a.width * a.height + b.width * b.height - intersection;
+	return union <= 0 ? 0 : intersection / union;
+}
+/** Union of two boxes. Merging discards rotation — Python resets `angle` to 0. */
+function mergeBoxes(a, b) {
+	const [ax0, ay0, ax1, ay1] = bbox(a);
+	const [bx0, by0, bx1, by1] = bbox(b);
+	const x = Math.min(ax0, bx0);
+	const y = Math.min(ay0, by0);
+	return {
+		text: `${a.text} ${b.text}`.trim(),
+		score: Math.min(a.score, b.score),
+		x,
+		y,
+		width: Math.max(ax1, bx1) - x,
+		height: Math.max(ay1, by1) - y,
+		angle: 0
+	};
+}
+/**
+* Whether two boxes sit on the same text line.
+*
+* For near-horizontal boxes this compares vertical centres against the average
+* height. For rotated boxes it projects the centre offset onto the line's normal
+* (dx = -sin, dy = cos), so the test follows the text's own baseline.
+*/
+function isOnSameLine(a, b, angleThresh = 10, lineThresh = .5) {
+	if (Math.abs(a.angle - b.angle) > angleThresh) return false;
+	const avgHeight = (a.height + b.height) / 2;
+	if (avgHeight <= 0) return false;
+	const acx = a.x + a.width / 2;
+	const acy = a.y + a.height / 2;
+	const bcx = b.x + b.width / 2;
+	const bcy = b.y + b.height / 2;
+	if (Math.abs(a.angle) < angleThresh) return Math.abs(acy - bcy) < avgHeight * lineThresh;
+	const rad = a.angle * Math.PI / 180;
+	const nx = -Math.sin(rad);
+	const ny = Math.cos(rad);
+	return Math.abs((bcx - acx) * nx + (bcy - acy) * ny) < avgHeight * lineThresh;
+}
+/**
+* Greedily merge boxes that overlap and share a line. Port of Python
+* `Box.merge_overlapping_boxes`.
+*
+* Restarts the scan after each merge so a chain of boxes collapses fully in one
+* call, matching Python's behaviour.
+*/
+function mergeOverlappingBoxes(boxes, iouThreshold = .3, angleThresh = 10, lineThresh = .5) {
+	const merged = [];
+	const used = new Array(boxes.length).fill(false);
+	for (let i = 0; i < boxes.length; i++) {
+		if (used[i]) continue;
+		let current = boxes[i];
+		for (let j = i + 1; j < boxes.length; j++) {
+			if (used[j]) continue;
+			const other = boxes[j];
+			if (boxIou(current, other) > iouThreshold && isOnSameLine(current, other, angleThresh, lineThresh)) {
+				current = mergeBoxes(current, other);
+				used[j] = true;
+			}
+		}
+		merged.push(current);
+		used[i] = true;
+	}
+	return merged;
+}
+//#endregion
+//#region src/core/image.ts
+function assertCanvasSupport() {
+	if (typeof OffscreenCanvas === "undefined") throw new Error("OffscreenCanvas is unavailable. scaledp requires a browser or worker context.");
+}
+function createCanvas(width, height) {
+	assertCanvasSupport();
+	return new OffscreenCanvas(Math.max(1, Math.round(width)), Math.max(1, Math.round(height)));
+}
+function context2d(canvas) {
+	const ctx = canvas.getContext("2d", { willReadFrequently: true });
+	if (!ctx) throw new Error("Failed to acquire a 2D context");
+	return ctx;
+}
+async function decodeImage(data) {
+	if (data instanceof Blob) return createImageBitmap(data);
+	const bytes = new Uint8Array(data.byteLength);
+	bytes.set(data);
+	return createImageBitmap(new Blob([bytes.buffer]));
+}
+function toImageData(source) {
+	if (source instanceof OffscreenCanvas) return context2d(source).getImageData(0, 0, source.width, source.height);
+	const canvas = createCanvas(source.width, source.height);
+	context2d(canvas).drawImage(source, 0, 0);
+	return context2d(canvas).getImageData(0, 0, canvas.width, canvas.height);
+}
+function imageDataToCanvas(image) {
+	const canvas = createCanvas(image.width, image.height);
+	context2d(canvas).putImageData(image, 0, 0);
+	return canvas;
+}
+async function encodeImage(source, type = "image/png", quality) {
+	const blob = await (source instanceof OffscreenCanvas ? source : imageDataToCanvas(source)).convertToBlob({
+		type,
+		quality
+	});
+	return new Uint8Array(await blob.arrayBuffer());
+}
+/** Read the intrinsic size of encoded image bytes without keeping the bitmap. */
+async function probeImageSize(data) {
+	const bitmap = await decodeImage(data);
+	try {
+		return {
+			width: bitmap.width,
+			height: bitmap.height
+		};
+	} finally {
+		bitmap.close();
+	}
+}
+/**
+* Fit an image into `target` preserving aspect ratio.
+*
+* `padding: 'end'` pads bottom and right only, matching PaddleOCR's detection
+* preprocessing -- coordinates then restore by dividing by `scale`, with no
+* offset to subtract. `padding: 'center'` centres the image, matching the YOLO
+* preprocessing, where the pad offsets must be subtracted before unscaling.
+*/
+function letterbox(source, target, opts = {}) {
+	const padding = opts.padding ?? "end";
+	const canvas = createCanvas(target.width, target.height);
+	const ctx = context2d(canvas);
+	ctx.fillStyle = opts.fill ?? "#ffffff";
+	ctx.fillRect(0, 0, canvas.width, canvas.height);
+	const scale = Math.min(target.width / source.width, target.height / source.height);
+	const width = Math.trunc(source.width * scale);
+	const height = Math.trunc(source.height * scale);
+	const dx = padding === "center" ? Math.trunc((target.width - width) / 2) : 0;
+	const dy = padding === "center" ? Math.trunc((target.height - height) / 2) : 0;
+	ctx.drawImage(source, dx, dy, width, height);
+	return {
+		canvas,
+		scale,
+		resized: {
+			width,
+			height
+		},
+		source: {
+			width: source.width,
+			height: source.height
+		}
+	};
+}
+/** Uniform resize by a scale factor. */
+function resize(source, factor) {
+	const canvas = createCanvas(source.width * factor, source.height * factor);
+	context2d(canvas).drawImage(source, 0, 0, canvas.width, canvas.height);
+	return canvas;
+}
+function cropGeometry(box, opts = {}) {
+	const scaled = scaleBox(box, opts.scaleFactor ?? 1, opts.padding ?? 0);
+	const width = Math.max(1, scaled.width);
+	const height = Math.max(1, scaled.height);
+	const axisAligned = {
+		scaled,
+		width,
+		height,
+		map: (x, y) => [scaled.x + x, scaled.y + y]
+	};
+	if (Math.abs(scaled.angle) < 3) return axisAligned;
+	const [, tl, tr, br] = boxPoints({
+		center: [scaled.x + width / 2, scaled.y + height / 2],
+		size: [width, height],
+		angle: scaled.angle
+	});
+	const bl = [tl[0] + (br[0] - tr[0]), tl[1] + (br[1] - tr[1])];
+	const ex = [(tr[0] - tl[0]) / width, (tr[1] - tl[1]) / width];
+	const ey = [(bl[0] - tl[0]) / height, (bl[1] - tl[1]) / height];
+	const det = ex[0] * ey[1] - ey[0] * ex[1];
+	if (Math.abs(det) < 1e-9) return axisAligned;
+	return {
+		scaled,
+		width,
+		height,
+		map: (x, y) => [tl[0] + x * ex[0] + y * ey[0], tl[1] + x * ex[1] + y * ey[1]]
+	};
+}
+function cropBox(source, box, opts = {}) {
+	const { scaled, width, height, map } = cropGeometry(box, opts);
+	const canvas = createCanvas(width, height);
+	const ctx = context2d(canvas);
+	const [ox, oy] = map(0, 0);
+	const [x1, y1] = map(1, 0);
+	const [x2, y2] = map(0, 1);
+	const ex = [x1 - ox, y1 - oy];
+	const ey = [x2 - ox, y2 - oy];
+	if (ex[1] === 0 && ey[0] === 0 && ex[0] === 1 && ey[1] === 1) {
+		ctx.drawImage(source, scaled.x, scaled.y, width, height, 0, 0, width, height);
+		return canvas;
+	}
+	const det = ex[0] * ey[1] - ey[0] * ex[1];
+	const a = ey[1] / det;
+	const b = -ex[1] / det;
+	const c = -ey[0] / det;
+	const d = ex[0] / det;
+	ctx.setTransform(a, b, c, d, -(a * ox + c * oy), -(b * ox + d * oy));
+	ctx.drawImage(source, 0, 0);
+	ctx.setTransform(1, 0, 0, 1, 0, 0);
+	return canvas;
+}
+/**
+* Build an NCHW float32 tensor from an image.
+*
+* `bgr` exists because ScaleDP's DBNet path converts RGB->BGR and never swaps
+* back, so the model is fed BGR channels against RGB ImageNet statistics.
+* Replicating that quirk is required for box parity.
+*/
+function toNchwFloat32(image, opts = {}) {
+	const { width, height, data } = image;
+	const scale = opts.scale ?? 1 / 255;
+	const mean = opts.mean ?? [
+		0,
+		0,
+		0
+	];
+	const std = opts.std ?? [
+		1,
+		1,
+		1
+	];
+	const order = opts.bgr ? [
+		2,
+		1,
+		0
+	] : [
+		0,
+		1,
+		2
+	];
+	const plane = width * height;
+	const out = new Float32Array(3 * plane);
+	for (let i = 0; i < plane; i++) for (let c = 0; c < 3; c++) {
+		const value = data[i * 4 + order[c]] * scale;
+		out[c * plane + i] = (value - mean[c]) / std[c];
+	}
+	return out;
+}
+const IMAGENET_MEAN = [
+	.485,
+	.456,
+	.406
+];
+const IMAGENET_STD = [
+	.229,
+	.224,
+	.225
+];
+//#endregion
+export { polygonArea as A, mergeBoxes as C, boxPoints as D, shape as E, convexHull as O, isRotated as S, scaleBox as T, boxFromBBox as _, cropBox as a, createBox as b, encodeImage as c, probeImageSize as d, resize as f, bbox as g, ROTATION_EPSILON_DEGREES as h, createCanvas as i, polygonPerimeter as j, minAreaRect as k, imageDataToCanvas as l, toNchwFloat32 as m, IMAGENET_STD as n, cropGeometry as o, toImageData as p, context2d as r, decodeImage as s, IMAGENET_MEAN as t, letterbox as u, boxFromPolygon as v, mergeOverlappingBoxes as w, isOnSameLine as x, boxIou as y };
+
+//# sourceMappingURL=image-CAH2rLv9.js.map