@foxycape/core 1.1.3

package/dist/mediaTypes/pdf/shared/selection/PdfSelection.mjs

@@ -0,0 +1,237 @@
+import { removeCJCharWhitespace as d, isNullOrWhiteSpace as m } from "../../../../kernal/common/text.mjs";
+import { mergeClientRectsToLines as y } from "../geometry/textRects.mjs";
+import { checkIsCrossPage as w } from "../range.mjs";
+import { BrowserCapabilities as p } from "../../../../kernal/web/BrowserCapabilities.mjs";
+class L {
+  constructor(t) {
+    this.ranges = [], this.englishPunctuations = /[ !"'@#$%^&*()\-+=;:<>?,.~`/|\]\[}{]/, this.latinCharaters = /[\u0000-\u00FF]/, this.sentenceEndPunctuations = ["。", "！", "？", "…", "；", "：", ".", "!", "?", ";"], this.otherEndPunctuations = ["，", "、", ",", ":", "）", ")", "》", "」", "』", "]", "】"], this.checkRequireWhiteSpace = (e) => {
+      if (!e)
+        return !1;
+      const n = e[e.length - 1];
+      return n.match(this.latinCharaters) ? !n.match(this.englishPunctuations) : !1;
+    }, this.joinItemTexts = (e) => {
+      let n = "";
+      for (const s of e)
+        s && (n && this.checkRequireWhiteSpace(n) && !/\s$/.test(n) && !/^\s/.test(s) && (n += " "), n += s);
+      return n;
+    };
+  }
+  getText(t) {
+    if (!t)
+      return "";
+    if (w(t) || p.isSupportTouch())
+      return d(t.toString());
+    if (this.ranges.includes(t))
+      return t.computedText ?? "";
+    const n = this.getRects(t);
+    let s = "";
+    try {
+      return n.some((r) => r.isVertical) ? (s = d(t.toString()), s) : (s = this.joinLinesAsParagraphs(n), s = d(s).replace(/\n+$/g, ""), m(s) && (s = d(t.toString())), s);
+    } finally {
+      t.computedText = s, this.ranges.unshift(t), this.ranges.length > 10 && this.ranges.splice(10);
+    }
+  }
+  /**
+   * Join visual lines. Base: original alignment heuristic;
+   * refine with punctuation + distance to the selection union bbox.
+   */
+  joinLinesAsParagraphs(t) {
+    if (t.length === 0)
+      return "";
+    const e = t.map((r) => ({
+      ...r,
+      endX: r.endX ?? r.x + r.width,
+      endY: r.endY ?? r.y + r.height,
+      text: r.text ?? ""
+    }));
+    if (e.length === 1)
+      return e[0].text;
+    const n = this.calcMaxRect(e);
+    let s = e[0].text;
+    for (let r = 1; r < e.length; r++) {
+      const i = e[r - 1], h = e[r];
+      if (!this.checkAllowAppendToHorizontal(i, h, n)) {
+        s = s.trimEnd() + `
+` + h.text;
+        continue;
+      }
+      this.checkRequireWhiteSpace(i.text) && !/\s$/.test(s) && (s += " "), s += h.text;
+    }
+    return s;
+  }
+  calcMaxRect(t) {
+    const e = Math.min(...t.map((i) => i.x)), n = Math.min(...t.map((i) => i.y)), s = Math.max(...t.map((i) => i.endX ?? i.x + i.width)), r = Math.max(...t.map((i) => i.endY ?? i.y + i.height));
+    return {
+      x: e,
+      y: n,
+      endX: s,
+      endY: r,
+      width: s - e,
+      height: r - n
+    };
+  }
+  getLineEndKind(t) {
+    const e = (t ?? "").trimEnd();
+    if (!e)
+      return "none";
+    const n = e[e.length - 1];
+    return this.sentenceEndPunctuations.includes(n) ? "sentence" : this.otherEndPunctuations.includes(n) ? "other" : "none";
+  }
+  /**
+   * 是否应软换行拼接（true=不换行，false=换段）。
+   * 先按标点分类，再结合相对外接矩形的 rightGap / leftGap，最后回退到原对齐启发式。
+   */
+  checkAllowAppendToHorizontal(t, e, n) {
+    const s = t.endX ?? t.x + t.width, r = e.endX ?? e.x + e.width, i = n.x, h = n.endX ?? n.x + n.width, a = Math.max(0, h - s), o = Math.max(0, e.x - i), c = t.width / Math.max((t.text ?? "").length, 1), l = e.width / Math.max((e.text ?? "").length, 1), x = Math.max(Math.min(c, l), 1), g = this.getLineEndKind(t.text), u = a > x * 3, f = o > x * 1.5;
+    return g === "sentence" ? !(u || f) : g === "other" ? !(u && f) : u && f ? !1 : this.checkAlignmentAllowsAppend(t, e, s, r, c, l);
+  }
+  /** 原 PdfTextAssistant 对齐启发式（true=继续拼接） */
+  checkAlignmentAllowsAppend(t, e, n, s, r, i) {
+    const h = Math.abs(t.x - e.x), a = Math.abs(n - s), o = Math.abs(t.width - e.width);
+    if (o < r && o < i)
+      return !0;
+    if (Math.abs(a - h) < r && Math.abs(a - h) < i)
+      return !1;
+    if (h < r && h < i) {
+      if (n < s) {
+        const c = t.height / (e.height || 1), l = t.width / (e.width || 1);
+        return c >= 0.95 && c <= 1.05 && l >= 0.7 && l <= 1.3;
+      }
+      return !0;
+    }
+    if (h > i * 1.5) {
+      if (t.x < e.x)
+        return t.width >= e.width;
+      const c = t.width > e.width ? e.width / (t.width || 1) : t.width / (e.width || 1);
+      return c >= 0.7 ? !0 : !(c < 0.7);
+    }
+    return a < r && a < i, !0;
+  }
+  /**
+   * Build visual line boxes for a selection (SVG text + pdf.js spans).
+   * Geometry merge is shared with mark overlay via {@link mergeClientRectsToLines}.
+   */
+  getRects(t) {
+    const e = this.collectLiveTextElementsInRange(t);
+    if (e.length === 0)
+      return [];
+    const n = [];
+    for (const s of e) {
+      const r = this.createItemRect(t, s);
+      r && n.push(r);
+    }
+    return n.length === 0 ? [] : y(n, this.joinItemTexts).map((s) => ({
+      x: s.x,
+      y: s.y,
+      width: s.width,
+      height: s.height,
+      endX: s.endX,
+      endY: s.endY,
+      text: s.text,
+      isVertical: s.isVertical
+    }));
+  }
+  /** Plain geometry + selected text for one live text element. */
+  createItemRect(t, e) {
+    const n = this.getTextIntersection(t, e);
+    if (!n)
+      return null;
+    const s = e.textContent ?? "", r = n !== s, i = (r ? this.measureIntersectionBounds(t, e) : null) ?? this.toBounds(e.getBoundingClientRect());
+    return i.width <= 0.5 || i.height <= 0.5 ? null : {
+      x: i.x,
+      y: i.y,
+      width: i.width,
+      height: i.height,
+      endX: i.x + i.width,
+      endY: i.y + i.height,
+      text: n,
+      isPartial: r
+    };
+  }
+  toBounds(t) {
+    return {
+      x: t.x,
+      y: t.y,
+      width: t.width,
+      height: t.height
+    };
+  }
+  /** Union of client rects for the selection ∩ element contents. */
+  measureIntersectionBounds(t, e) {
+    try {
+      const n = this.createIntersectionRange(t, e);
+      if (!n)
+        return null;
+      const s = Array.from(n.getClientRects()).filter((o) => o.width > 0.5 && o.height > 0.5);
+      if (s.length === 0) {
+        const o = n.getBoundingClientRect();
+        return o.width <= 0.5 || o.height <= 0.5 ? null : this.toBounds(o);
+      }
+      const r = Math.min(...s.map((o) => o.x)), i = Math.min(...s.map((o) => o.y)), h = Math.max(...s.map((o) => o.x + o.width)), a = Math.max(...s.map((o) => o.y + o.height));
+      return {
+        x: r,
+        y: i,
+        width: h - r,
+        height: a - i
+      };
+    } catch {
+      return null;
+    }
+  }
+  createIntersectionRange(t, e) {
+    try {
+      const n = e.ownerDocument.createRange();
+      n.selectNodeContents(e);
+      const s = t.cloneRange();
+      return s.compareBoundaryPoints(Range.START_TO_START, n) < 0 && s.setStart(n.startContainer, n.startOffset), s.compareBoundaryPoints(Range.END_TO_END, n) > 0 && s.setEnd(n.endContainer, n.endOffset), s;
+    } catch {
+      return null;
+    }
+  }
+  resolveTextLayerKind(t) {
+    const e = t.startContainer, n = e.nodeType === Node.TEXT_NODE ? e.parentElement : e;
+    if (!n)
+      return "unknown";
+    if (n.closest("svg.custom-text-layer"))
+      return "svg";
+    if (n.closest(".textLayer"))
+      return "span";
+    const s = n.closest(".page");
+    return s?.querySelector("svg.custom-text-layer text") ? "svg" : s?.querySelector(".textLayer span") ? "span" : "unknown";
+  }
+  collectPageTextCandidates(t, e) {
+    if (e === "svg")
+      return Array.from(t.querySelectorAll("svg.custom-text-layer text"));
+    if (e === "span")
+      return this.collectTextLayerSpans(t);
+    const n = Array.from(t.querySelectorAll("svg.custom-text-layer text"));
+    return n.length > 0 ? n : this.collectTextLayerSpans(t);
+  }
+  collectTextLayerSpans(t) {
+    return Array.from(t.querySelectorAll(".textLayer span")).filter((e) => !e.classList.contains("highlight") && !e.classList.contains("appended") && (e.textContent?.length ?? 0) > 0);
+  }
+  /**
+   * Collect live SVG `<text>` or `.textLayer span` elements intersected by the range.
+   * Span mode has no element ids, so we must walk the live DOM instead of cloneContents + getElementById.
+   */
+  collectLiveTextElementsInRange(t) {
+    const n = (t.startContainer.nodeType === Node.TEXT_NODE ? t.startContainer.parentElement : t.startContainer)?.closest(".page");
+    if (!n)
+      return [];
+    const s = this.resolveTextLayerKind(t), r = this.collectPageTextCandidates(n, s), i = [];
+    for (const h of r)
+      try {
+        t.intersectsNode(h) && i.push(h);
+      } catch {
+      }
+    return i;
+  }
+  /** Selected text within one live text element (partial for first/last). */
+  getTextIntersection(t, e) {
+    const n = this.createIntersectionRange(t, e);
+    return n ? n.toString() : e.textContent ?? "";
+  }
+}
+export {
+  L as PdfSelection
+};

package/dist/mediaTypes/pdf/shared/selection/index.cjs

@@ -0,0 +1 @@
+"use strict";Object.defineProperty(exports,Symbol.toStringTag,{value:"Module"});const e=require("./PdfSelection.cjs"),c=require("../range.cjs");exports.PdfSelection=e.PdfSelection;exports.checkIsCrossPage=c.checkIsCrossPage;

package/dist/mediaTypes/pdf/shared/selection/index.d.ts

@@ -0,0 +1,3 @@
+export * from './IPdfSelection';
+export * from './PdfSelection';
+export * from '../range';

package/dist/mediaTypes/pdf/shared/selection/index.mjs

@@ -0,0 +1,6 @@
+import { PdfSelection as r } from "./PdfSelection.mjs";
+import { checkIsCrossPage as f } from "../range.mjs";
+export {
+  r as PdfSelection,
+  f as checkIsCrossPage
+};

package/dist/mediaTypes/pdf/shared/text/bbox.cjs

@@ -0,0 +1 @@
+"use strict";Object.defineProperty(exports,Symbol.toStringTag,{value:"Module"});const z=require("../../../../pdfjs/legacy/build/pdf.mjs");function D(t){const n=Object.create(null,{[Symbol.toStringTag]:{value:"Module"}});if(t){for(const o in t)if(o!=="default"){const e=Object.getOwnPropertyDescriptor(t,o);Object.defineProperty(n,o,e.get?e:{enumerable:!0,get:()=>t[o]})}}return n.default=t,Object.freeze(n)}const v=D(z),S=.8,j=t=>typeof t.str=="string",l=(t,n,o)=>({x:t[0]*n+t[2]*o+t[4],y:t[1]*n+t[3]*o+t[5]}),B=t=>{const n=t.map(e=>e.x),o=t.map(e=>e.y);return{x1:Math.min(...n),y1:Math.min(...o),x2:Math.max(...n),y2:Math.max(...o)}},C=(t,n)=>{if(!t||n<=0)return n*S;const o=t.ascent,e=Math.abs(t.descent??0);return o>0&&o<=1?n*o:o>1?o:e>0&&e<=1?n*(1-e):n*S},I=(t,n,o)=>{if(!t)return n-o;const e=Math.abs(t.descent??0);return e>0&&e<=1?n*e:e>1?e:Math.max(n-o,0)},V=(t,n,o)=>n<=0||o<=0?{x1:0,y1:0,x2:0,y2:0}:{x1:t.x1/n,y1:t.y1/o,x2:t.x2/n,y2:t.y2/o},p=t=>Math.max(0,t.x2-t.x1)*Math.max(0,t.y2-t.y1),b=(t,n)=>{const o=Math.max(t.x1,n.x1),e=Math.max(t.y1,n.y1),r=Math.min(t.x2,n.x2),a=Math.min(t.y2,n.y2);return r<=o||a<=e?null:{x1:o,y1:e,x2:r,y2:a}},F=t=>({x:(t.x1+t.x2)/2,y:(t.y1+t.y2)/2}),L=(t,n)=>t.x>=n.x1&&t.x<=n.x2&&t.y>=n.y1&&t.y<=n.y2,T=(t,n,o)=>{const e=o??t.rotate,r=t.getViewport({scale:1,rotation:e}),a=n/r.width;return t.getViewport({scale:a,rotation:e})},A=(t,n,o)=>{const e=v.Util.transform(o,t.transform),r=Math.hypot(e[2],e[3]);if(r<=0)return{x1:e[4],y1:e[5],x2:e[4],y2:e[5]};const a=C(n,r),i=I(n,r,a),s=t.transform,u=!!n?.vertical,x=u?t.height:t.width,f=u?Math.hypot(s[2],s[3]):Math.hypot(s[0],s[1]),c=f>0?x/f:x;if(u){const h=Math.hypot(e[0],e[1])||r,g=a/h,M=-i/h,R=[l(e,M,0),l(e,g,0),l(e,g,-c),l(e,M,-c)];return B(R)}const m=r,d=a/m,y=-i/m,P=[l(e,0,d),l(e,c,d),l(e,c,y),l(e,0,y)];return B(P)},O=(t,n,o)=>{const e=v.Util.transform(o,t.transform);let r=Math.atan2(e[1],e[0]);return n?.vertical&&(r+=Math.PI/2),r*(180/Math.PI)},w=(t,n,o,e,r,a=!1)=>{if(n<=0||o<=0)return t;if(a){const s=e/n,u=r/o;return{x1:t.x1*s,y1:t.y1*u,x2:t.x2*s,y2:t.y2*u}}const i=e/n;return{x1:t.x1*i,y1:t.y1*i,x2:t.x2*i,y2:t.y2*i}},N=async(t,n)=>{const{targetWidth:o,targetHeight:e,rotation:r,independentAxisScale:a=!1}=n,i=T(t,o,r),s=await t.getTextContent(),u=[];let x=0;for(const f of s.items){if(!j(f))continue;const c=f;if(!c.str&&c.width<=0&&c.height<=0)continue;const m=s.styles[c.fontName],d=A(c,m,i.transform),y=w(d,i.width,i.height,o,e,a);u.push({index:x,text:c.str,bounds:y,normalizedBounds:V(y,o,e),hasEOL:c.hasEOL,dir:c.dir,fontName:c.fontName,angleDeg:O(c,m,i.transform)}),x++}return u},_=(t,n=.005)=>[...t].sort((o,e)=>{const r=o.normalizedBounds.y1-e.normalizedBounds.y1;return Math.abs(r)>n?r:o.normalizedBounds.x1-e.normalizedBounds.x1}),E=(t,n,o)=>{const e=o?.minOverlapRatio??.05,r=o?.minSpanCoverage??.4,a=p(n);return a<=0?"":t.filter(s=>{if(!s.text)return!1;if(L(F(s.bounds),n))return!0;const u=b(s.bounds,n);if(!u)return!1;const x=p(u),f=p(s.bounds),c=x/a,m=f>0?x/f:0;return!(c<e&&m<r)}).map(s=>s.text).join("")};exports.collectTextInBounds=E;exports.computeSpanAngleDeg=O;exports.computeSpanBoundsInViewport=A;exports.computeViewport=T;exports.extractFromPage=N;exports.mapBoundsToTargetSize=w;exports.sortByReadingOrder=_;

package/dist/mediaTypes/pdf/shared/text/bbox.d.ts

@@ -0,0 +1,35 @@
+import { TextItem, TextStyle } from '../../../../pdfjs/types/src/display/api';
+import * as pdfjsLib from '../../../../pdfjs/legacy/build/pdf.mjs';
+export type PdfAxisAlignedRect = {
+    x1: number;
+    y1: number;
+    x2: number;
+    y2: number;
+};
+export type PdfTextSpan = {
+    index: number;
+    text: string;
+    bounds: PdfAxisAlignedRect;
+    normalizedBounds: PdfAxisAlignedRect;
+    hasEOL: boolean;
+    dir: string;
+    fontName: string;
+    angleDeg: number;
+};
+export type PdfTextSpanExtractOptions = {
+    targetWidth: number;
+    targetHeight: number;
+    rotation?: number;
+    independentAxisScale?: boolean;
+};
+export type PdfTextSpanCollectOptions = {
+    minOverlapRatio?: number;
+    minSpanCoverage?: number;
+};
+export declare const computeViewport: (page: pdfjsLib.PDFPageProxy, targetWidth: number, rotation?: number) => pdfjsLib.PageViewport;
+export declare const computeSpanBoundsInViewport: (textItem: TextItem, style: TextStyle | undefined, viewportTransform: number[]) => PdfAxisAlignedRect;
+export declare const computeSpanAngleDeg: (textItem: TextItem, style: TextStyle | undefined, viewportTransform: number[]) => number;
+export declare const mapBoundsToTargetSize: (bounds: PdfAxisAlignedRect, viewportWidth: number, viewportHeight: number, targetWidth: number, targetHeight: number, independentAxisScale?: boolean) => PdfAxisAlignedRect;
+export declare const extractFromPage: (page: pdfjsLib.PDFPageProxy, options: PdfTextSpanExtractOptions) => Promise<PdfTextSpan[]>;
+export declare const sortByReadingOrder: (spans: PdfTextSpan[], lineTolerance?: number) => PdfTextSpan[];
+export declare const collectTextInBounds: (spans: PdfTextSpan[], query: PdfAxisAlignedRect, options?: PdfTextSpanCollectOptions) => string;

package/dist/mediaTypes/pdf/shared/text/bbox.mjs

@@ -0,0 +1,144 @@
+import * as B from "../../../../pdfjs/legacy/build/pdf.mjs";
+const v = 0.8, T = (t) => typeof t.str == "string", f = (t, o, e) => ({
+  x: t[0] * o + t[2] * e + t[4],
+  y: t[1] * o + t[3] * e + t[5]
+}), A = (t) => {
+  const o = t.map((n) => n.x), e = t.map((n) => n.y);
+  return {
+    x1: Math.min(...o),
+    y1: Math.min(...e),
+    x2: Math.max(...o),
+    y2: Math.max(...e)
+  };
+}, w = (t, o) => {
+  if (!t || o <= 0)
+    return o * v;
+  const e = t.ascent, n = Math.abs(t.descent ?? 0);
+  return e > 0 && e <= 1 ? o * e : e > 1 ? e : n > 0 && n <= 1 ? o * (1 - n) : o * v;
+}, C = (t, o, e) => {
+  if (!t)
+    return o - e;
+  const n = Math.abs(t.descent ?? 0);
+  return n > 0 && n <= 1 ? o * n : n > 1 ? n : Math.max(o - e, 0);
+}, z = (t, o, e) => o <= 0 || e <= 0 ? {
+  x1: 0,
+  y1: 0,
+  x2: 0,
+  y2: 0
+} : {
+  x1: t.x1 / o,
+  y1: t.y1 / e,
+  x2: t.x2 / o,
+  y2: t.y2 / e
+}, h = (t) => Math.max(0, t.x2 - t.x1) * Math.max(0, t.y2 - t.y1), D = (t, o) => {
+  const e = Math.max(t.x1, o.x1), n = Math.max(t.y1, o.y1), s = Math.min(t.x2, o.x2), c = Math.min(t.y2, o.y2);
+  return s <= e || c <= n ? null : {
+    x1: e,
+    y1: n,
+    x2: s,
+    y2: c
+  };
+}, I = (t) => ({
+  x: (t.x1 + t.x2) / 2,
+  y: (t.y1 + t.y2) / 2
+}), O = (t, o) => t.x >= o.x1 && t.x <= o.x2 && t.y >= o.y1 && t.y <= o.y2, P = (t, o, e) => {
+  const n = e ?? t.rotate, s = t.getViewport({
+    scale: 1,
+    rotation: n
+  }), c = o / s.width;
+  return t.getViewport({
+    scale: c,
+    rotation: n
+  });
+}, V = (t, o, e) => {
+  const n = B.Util.transform(e, t.transform), s = Math.hypot(n[2], n[3]);
+  if (s <= 0)
+    return {
+      x1: n[4],
+      y1: n[5],
+      x2: n[4],
+      y2: n[5]
+    };
+  const c = w(o, s), i = C(o, s, c), r = t.transform, x = !!o?.vertical, u = x ? t.height : t.width, y = x ? Math.hypot(r[2], r[3]) : Math.hypot(r[0], r[1]), a = y > 0 ? u / y : u;
+  if (x) {
+    const p = Math.hypot(n[0], n[1]) || s, M = c / p, g = -i / p, S = [f(n, g, 0), f(n, M, 0), f(n, M, -a), f(n, g, -a)];
+    return A(S);
+  }
+  const m = s, d = c / m, l = -i / m, R = [f(n, 0, d), f(n, a, d), f(n, a, l), f(n, 0, l)];
+  return A(R);
+}, F = (t, o, e) => {
+  const n = B.Util.transform(e, t.transform);
+  let s = Math.atan2(n[1], n[0]);
+  return o?.vertical && (s += Math.PI / 2), s * (180 / Math.PI);
+}, L = (t, o, e, n, s, c = !1) => {
+  if (o <= 0 || e <= 0)
+    return t;
+  if (c) {
+    const r = n / o, x = s / e;
+    return {
+      x1: t.x1 * r,
+      y1: t.y1 * x,
+      x2: t.x2 * r,
+      y2: t.y2 * x
+    };
+  }
+  const i = n / o;
+  return {
+    x1: t.x1 * i,
+    y1: t.y1 * i,
+    x2: t.x2 * i,
+    y2: t.y2 * i
+  };
+}, E = async (t, o) => {
+  const {
+    targetWidth: e,
+    targetHeight: n,
+    rotation: s,
+    independentAxisScale: c = !1
+  } = o, i = P(t, e, s), r = await t.getTextContent(), x = [];
+  let u = 0;
+  for (const y of r.items) {
+    if (!T(y))
+      continue;
+    const a = y;
+    if (!a.str && a.width <= 0 && a.height <= 0)
+      continue;
+    const m = r.styles[a.fontName], d = V(a, m, i.transform), l = L(d, i.width, i.height, e, n, c);
+    x.push({
+      index: u,
+      text: a.str,
+      bounds: l,
+      normalizedBounds: z(l, e, n),
+      hasEOL: a.hasEOL,
+      dir: a.dir,
+      fontName: a.fontName,
+      angleDeg: F(a, m, i.transform)
+    }), u++;
+  }
+  return x;
+}, N = (t, o = 5e-3) => [...t].sort((e, n) => {
+  const s = e.normalizedBounds.y1 - n.normalizedBounds.y1;
+  return Math.abs(s) > o ? s : e.normalizedBounds.x1 - n.normalizedBounds.x1;
+}), U = (t, o, e) => {
+  const n = e?.minOverlapRatio ?? 0.05, s = e?.minSpanCoverage ?? 0.4, c = h(o);
+  return c <= 0 ? "" : t.filter((r) => {
+    if (!r.text)
+      return !1;
+    if (O(I(r.bounds), o))
+      return !0;
+    const x = D(r.bounds, o);
+    if (!x)
+      return !1;
+    const u = h(x), y = h(r.bounds), a = u / c, m = y > 0 ? u / y : 0;
+    return !(a < n && m < s);
+  }).map((r) => r.text).join("");
+};
+export {
+  U as collectTextInBounds,
+  F as computeSpanAngleDeg,
+  V as computeSpanBoundsInViewport,
+  P as computeViewport,
+  E as extractFromPage,
+  L as mapBoundsToTargetSize,
+  N as sortByReadingOrder
+};

package/dist/mediaTypes/pdf/shared/text/pageText.cjs

@@ -0,0 +1,5 @@
+"use strict";Object.defineProperty(exports,Symbol.toStringTag,{value:"Module"});const x=require("../../../../kernal/common/text.cjs"),C=require("./bbox.cjs"),k=[".","。","？","?","!","！"],I=.2,j=.15,v=/[。！？；：」》）"'\]]$/,B=.85,d=t=>typeof t=="object"&&t!==null&&typeof t.str=="string",M=(t,n,e)=>{for(let r=n;r>=0&&r<t.length;r+=e){const s=t[r];if(d(s)&&s.str&&!/^\s+$/.test(s.str))return s}return null},E=(t,n)=>{const e=Math.max(t.height,n.height,1)*.5;return Math.abs(t.transform[5]-n.transform[5])<=e},O=t=>t.width/Math.max(t.str.length,1),R=(t,n,e)=>{const r=n.transform[4]-(t.transform[4]+t.width);return e?Math.max(e.width,r,0):Math.max(r,0)},L=(t,n,e)=>{const r=t.str.trimEnd(),s=n.str.trimStart();if(!x.endsWithCjkIdeograph(r)||!x.startsWithCjkIdeograph(s)||e?.str.includes("　")||t.hasEOL||t.fontName!==n.fontName||v.test(r))return!0;if(!E(t,n))return!1;const o=O(t),i=R(t,n,e);return r.length>1&&s.length>1&&i>o*j?!0:r.length===1&&s.length===1&&i<=o*I?!1:i>o*I},A=(t,n,e)=>{if(!/ +$/.test(t))return t;const r=t.replace(/ +$/,"");return!x.endsWithCjkIdeograph(r)||!x.startsWithCjkIdeograph(e.str)||L(n,e)?t:r},y=(t,n,e,r)=>r&&e&&!t.endsWith("-")?t+n+`
+`:t+n,S=(t,n,e)=>!e||!n.hasEOL||t.endsWith("-")?t:t+`
+`,P=(t,n)=>{let e="",r=null;for(let s=0;s<t.length;s++){const o=t[s];if(!d(o))continue;if(!o.str){e=S(e,o,n),r=o;continue}if(/^\s+$/.test(o.str)){const f=r??M(t,s-1,-1),a=M(t,s+1,1);!f||!a?e+=o.str:L(f,a,o)&&(e+=" ");continue}const i=x.removeCJCharWhitespace(o.str);r&&(e=A(e,r,o)),e=y(e,i,o.hasEOL,n),r=o}return e},F=(t,n)=>({x1:Math.min(t.x1,n.x1),y1:Math.min(t.y1,n.y1),x2:Math.max(t.x2,n.x2),y2:Math.max(t.y2,n.y2)}),m=t=>Math.max(0,t.x2-t.x1)*Math.max(0,t.y2-t.y1),N=(t,n)=>{const e=Math.max(t.x1,n.x1),r=Math.max(t.y1,n.y1),s=Math.min(t.x2,n.x2),o=Math.min(t.y2,n.y2);return s<=e||o<=r?null:{x1:e,y1:r,x2:s,y2:o}},$=(t,n)=>{const e=N(t,n);if(!e)return 0;const r=m(e),s=Math.min(m(t),m(n));return s>0?r/s:0},W=t=>t.replace(/\s+/g,""),z=(t,n)=>{const e=W(t),r=W(n);return!e||!r?!1:e===r||e.includes(r)||r.includes(e)},w=t=>!!(!t.str||/^\s+$/.test(t.str)||t.width<=0&&t.height<=0),G=t=>{const n=t.transform[4],e=t.transform[5],r=Math.max(t.width,0),s=Math.max(t.height,0);return{x1:Math.min(n,n+r),y1:Math.min(e,e+s),x2:Math.max(n,n+r),y2:Math.max(e,e+s)}},b=t=>{const n=[];for(const e of t){if(w(e.item)){n.push(e);continue}const r=m(e.bounds);let s=-1;for(let a=0;a<n.length;a++){const c=n[a];if(!w(c.item)&&z(c.item.str,e.item.str)&&!($(c.bounds,e.bounds)<B)){s=a;break}}if(s<0){n.push(e);continue}const o=n[s],i=m(o.bounds);(r>i*1.01||Math.abs(r-i)<=i*.01&&e.item.str.length>o.item.str.length)&&(n[s]=e)}return n},J=t=>{const n=[];for(let r=0;r<t.length;r++){const s=t[r];d(s)&&n.push({item:s,bounds:G(s),sourceIndex:r})}const e=new Set(b(n).map(r=>r.sourceIndex));return t.filter((r,s)=>!d(r)||e.has(s))},D=(t,n)=>{let r="";if(n?.removeCJWhitespace)r=P(t,!1);else for(const s of t)s.str&&(r=y(r,s.str,!1,!1));return n?.removeConsecutiveWhitespaceCharacters&&(r=r.trim(),r=r.replace(/[\f\t\v]+/g,"")),r},H=(t,n,e,r,s)=>{const o=C.computeSpanBoundsInViewport(t,n,e.transform);return C.mapBoundsToTargetSize(o,e.width,e.height,r,s)},q=async(t,n,e)=>{const r=C.computeViewport(t,n),s=r.height,o=await t.getTextContent(),i=[];for(const h of o.items){if(!d(h))continue;const u=h;if(!u.str&&u.width<=0&&u.height<=0&&!u.hasEOL)continue;const l=o.styles[u.fontName],g=H(u,l,r,n,s);i.push({item:u,bounds:g})}const f=b(i),a=[];let c=[],p=null;const T=()=>{if(c.length===0)return;const h=D(c.map(l=>l.item),e),u=c.map(l=>l.bounds).reduce((l,g)=>F(l,g));h&&a.push({text:h,rect:u}),c=[]};for(const h of f)p?.hasEOL&&T(),c.push(h),p=h.item;return T(),a},V=async(t,n)=>{let e="";const r=await t.getTextContent(),s=J(r.items),o=n?.combineLines||n?.convertEOLToCRLF||n?.removeConsecutiveBlankLine||n?.removeConsecutiveWhitespaceCharacters||n?.convertLFToWhitespace||n?.removeCJWhitespace;if(n?.removeCJWhitespace?e=P(s,o):s.forEach(i=>{if(d(i)){if(!i.str){e=S(e,i,o);return}e=y(e,i.str,i.hasEOL,o)}}),n?.combineLines){let i="";const f=e.split(`
+`);for(const a of f){const c=a.trim();if(!c)continue;i+=c;const p=c.substring(c.length-1);k.includes(p)&&(i+=`
+`)}e=i}return n?.removeConsecutiveWhitespaceCharacters&&(e=e.trim(),e=e.replace(/[\f\t\v]+/g,""),e=e.replace(/^\s*[\r\n]\s*/gm,"")),e};exports.getPageText=V;exports.getPageTextItems=q;

package/dist/mediaTypes/pdf/shared/text/pageText.d.ts

@@ -0,0 +1,9 @@
+import { TextFormatOptions } from '../../../../kernal';
+import { PdfAxisAlignedRect } from './bbox';
+
+export type PdfPageTextLineItem = {
+    text: string;
+    rect: PdfAxisAlignedRect;
+};
+export declare const getPageTextItems: (page: any, width: number, options?: TextFormatOptions) => Promise<PdfPageTextLineItem[]>;
+export declare const getPageText: (page: any, options?: TextFormatOptions) => Promise<string>;

package/dist/mediaTypes/pdf/shared/text/pageText.mjs

@@ -0,0 +1,186 @@
+import { removeCJCharWhitespace as v, endsWithCjkIdeograph as I, startsWithCjkIdeograph as L } from "../../../../kernal/common/text.mjs";
+import { computeViewport as B, computeSpanBoundsInViewport as E, mapBoundsToTargetSize as R } from "./bbox.mjs";
+const b = [".", "。", "？", "?", "!", "！"], y = 0.2, j = 0.15, O = /[。！？；：」》）"'\]]$/, A = 0.85, m = (t) => typeof t == "object" && t !== null && typeof t.str == "string", T = (t, n, e) => {
+  for (let r = n; r >= 0 && r < t.length; r += e) {
+    const s = t[r];
+    if (m(s) && s.str && !/^\s+$/.test(s.str))
+      return s;
+  }
+  return null;
+}, F = (t, n) => {
+  const e = Math.max(t.height, n.height, 1) * 0.5;
+  return Math.abs(t.transform[5] - n.transform[5]) <= e;
+}, N = (t) => t.width / Math.max(t.str.length, 1), $ = (t, n, e) => {
+  const r = n.transform[4] - (t.transform[4] + t.width);
+  return e ? Math.max(e.width, r, 0) : Math.max(r, 0);
+}, W = (t, n, e) => {
+  const r = t.str.trimEnd(), s = n.str.trimStart();
+  if (!I(r) || !L(s) || e?.str.includes("　") || t.hasEOL || t.fontName !== n.fontName || O.test(r))
+    return !0;
+  if (!F(t, n))
+    return !1;
+  const o = N(t), i = $(t, n, e);
+  return r.length > 1 && s.length > 1 && i > o * j ? !0 : r.length === 1 && s.length === 1 && i <= o * y ? !1 : i > o * y;
+}, z = (t, n, e) => {
+  if (!/ +$/.test(t))
+    return t;
+  const r = t.replace(/ +$/, "");
+  return !I(r) || !L(e.str) || W(n, e) ? t : r;
+}, g = (t, n, e, r) => r && e && !t.endsWith("-") ? t + n + `
+` : t + n, S = (t, n, e) => !e || !n.hasEOL || t.endsWith("-") ? t : t + `
+`, P = (t, n) => {
+  let e = "", r = null;
+  for (let s = 0; s < t.length; s++) {
+    const o = t[s];
+    if (!m(o))
+      continue;
+    if (!o.str) {
+      e = S(e, o, n), r = o;
+      continue;
+    }
+    if (/^\s+$/.test(o.str)) {
+      const f = r ?? T(t, s - 1, -1), a = T(t, s + 1, 1);
+      !f || !a ? e += o.str : W(f, a, o) && (e += " ");
+      continue;
+    }
+    const i = v(o.str);
+    r && (e = z(e, r, o)), e = g(e, i, o.hasEOL, n), r = o;
+  }
+  return e;
+}, G = (t, n) => ({
+  x1: Math.min(t.x1, n.x1),
+  y1: Math.min(t.y1, n.y1),
+  x2: Math.max(t.x2, n.x2),
+  y2: Math.max(t.y2, n.y2)
+}), d = (t) => Math.max(0, t.x2 - t.x1) * Math.max(0, t.y2 - t.y1), J = (t, n) => {
+  const e = Math.max(t.x1, n.x1), r = Math.max(t.y1, n.y1), s = Math.min(t.x2, n.x2), o = Math.min(t.y2, n.y2);
+  return s <= e || o <= r ? null : {
+    x1: e,
+    y1: r,
+    x2: s,
+    y2: o
+  };
+}, D = (t, n) => {
+  const e = J(t, n);
+  if (!e)
+    return 0;
+  const r = d(e), s = Math.min(d(t), d(n));
+  return s > 0 ? r / s : 0;
+}, M = (t) => t.replace(/\s+/g, ""), H = (t, n) => {
+  const e = M(t), r = M(n);
+  return !e || !r ? !1 : e === r || e.includes(r) || r.includes(e);
+}, w = (t) => !!(!t.str || /^\s+$/.test(t.str) || t.width <= 0 && t.height <= 0), V = (t) => {
+  const n = t.transform[4], e = t.transform[5], r = Math.max(t.width, 0), s = Math.max(t.height, 0);
+  return {
+    x1: Math.min(n, n + r),
+    y1: Math.min(e, e + s),
+    x2: Math.max(n, n + r),
+    y2: Math.max(e, e + s)
+  };
+}, k = (t) => {
+  const n = [];
+  for (const e of t) {
+    if (w(e.item)) {
+      n.push(e);
+      continue;
+    }
+    const r = d(e.bounds);
+    let s = -1;
+    for (let a = 0; a < n.length; a++) {
+      const c = n[a];
+      if (!w(c.item) && H(c.item.str, e.item.str) && !(D(c.bounds, e.bounds) < A)) {
+        s = a;
+        break;
+      }
+    }
+    if (s < 0) {
+      n.push(e);
+      continue;
+    }
+    const o = n[s], i = d(o.bounds);
+    (r > i * 1.01 || Math.abs(r - i) <= i * 0.01 && e.item.str.length > o.item.str.length) && (n[s] = e);
+  }
+  return n;
+}, K = (t) => {
+  const n = [];
+  for (let r = 0; r < t.length; r++) {
+    const s = t[r];
+    m(s) && n.push({
+      item: s,
+      bounds: V(s),
+      sourceIndex: r
+    });
+  }
+  const e = new Set(k(n).map((r) => r.sourceIndex));
+  return t.filter((r, s) => !m(r) || e.has(s));
+}, U = (t, n) => {
+  let r = "";
+  if (n?.removeCJWhitespace)
+    r = P(t, !1);
+  else
+    for (const s of t)
+      s.str && (r = g(r, s.str, !1, !1));
+  return n?.removeConsecutiveWhitespaceCharacters && (r = r.trim(), r = r.replace(/[\f\t\v]+/g, "")), r;
+}, q = (t, n, e, r, s) => {
+  const o = E(t, n, e.transform);
+  return R(o, e.width, e.height, r, s);
+}, Y = async (t, n, e) => {
+  const r = B(t, n), s = r.height, o = await t.getTextContent(), i = [];
+  for (const h of o.items) {
+    if (!m(h))
+      continue;
+    const u = h;
+    if (!u.str && u.width <= 0 && u.height <= 0 && !u.hasEOL)
+      continue;
+    const l = o.styles[u.fontName], p = q(u, l, r, n, s);
+    i.push({
+      item: u,
+      bounds: p
+    });
+  }
+  const f = k(i), a = [];
+  let c = [], x = null;
+  const C = () => {
+    if (c.length === 0)
+      return;
+    const h = U(c.map((l) => l.item), e), u = c.map((l) => l.bounds).reduce((l, p) => G(l, p));
+    h && a.push({
+      text: h,
+      rect: u
+    }), c = [];
+  };
+  for (const h of f)
+    x?.hasEOL && C(), c.push(h), x = h.item;
+  return C(), a;
+}, Z = async (t, n) => {
+  let e = "";
+  const r = await t.getTextContent(), s = K(r.items), o = n?.combineLines || n?.convertEOLToCRLF || n?.removeConsecutiveBlankLine || n?.removeConsecutiveWhitespaceCharacters || n?.convertLFToWhitespace || n?.removeCJWhitespace;
+  if (n?.removeCJWhitespace ? e = P(s, o) : s.forEach((i) => {
+    if (m(i)) {
+      if (!i.str) {
+        e = S(e, i, o);
+        return;
+      }
+      e = g(e, i.str, i.hasEOL, o);
+    }
+  }), n?.combineLines) {
+    let i = "";
+    const f = e.split(`
+`);
+    for (const a of f) {
+      const c = a.trim();
+      if (!c)
+        continue;
+      i += c;
+      const x = c.substring(c.length - 1);
+      b.includes(x) && (i += `
+`);
+    }
+    e = i;
+  }
+  return n?.removeConsecutiveWhitespaceCharacters && (e = e.trim(), e = e.replace(/[\f\t\v]+/g, ""), e = e.replace(/^\s*[\r\n]\s*/gm, "")), e;
+};
+export {
+  Z as getPageText,
+  Y as getPageTextItems
+};

package/dist/pdfjs/CODE_OF_CONDUCT.md

@@ -0,0 +1,15 @@
+# Community Participation Guidelines
+
+This repository is governed by Mozilla's code of conduct and etiquette guidelines. 
+For more details, please read the
+[Mozilla Community Participation Guidelines](https://www.mozilla.org/about/governance/policies/participation/). 
+
+## How to Report
+For more information on how to report violations of the Community Participation Guidelines, please read our '[How to Report](https://www.mozilla.org/about/governance/policies/participation/reporting/)' page.
+
+<!--
+## Project Specific Etiquette
+
+In some cases, there will be additional project etiquette i.e.: (https://bugzilla.mozilla.org/page.cgi?id=etiquette.html).
+Please update for your project.
+-->