extract-pdf 0.1.21 → 0.1.22

package/src/transforms/calculate-global-stats.ts

@@ -1,132 +1,132 @@
-/**
- * @description First-pass transformation that scans all `TextItem`s to compute
- * document-wide statistics stored in `parseResult.globals`: most-frequent text
- * height (`mostUsedHeight`), most-frequent font, most-frequent line distance,
- * maximum height, and a font-to-format map (BOLD/OBLIQUE/BOLD_OBLIQUE) derived
- * from font names. Also deep-copies every page item so subsequent transforms
- * never mutate the original parsed data.
- */
-import ToTextItemTransformation from "./base/to-text-item-transform";
-import ParseResult from "../models/parse-result";
-import { WordFormat } from "../models/line-converter";
-
-export default class CalculateGlobalStats extends ToTextItemTransformation {
-  fontMap: Map<string, { name: string }> | undefined;
-
-  constructor(fontMap?: Map<string, { name: string }>) {
-    super("Calculate Global Stats");
-    this.fontMap = fontMap;
-  }
-
-  transform(parseResult: ParseResult): ParseResult {
-    const heightToOccurrence: Record<string, number> = {};
-    const fontToOccurrence: Record<string, number> = {};
-    var maxHeight = 0;
-    var maxHeightFont: string | undefined;
-    parseResult.pages.forEach((page) => {
-      page.items.forEach((item) => {
-        if (!item.height) return;
-        heightToOccurrence[item.height] = heightToOccurrence[item.height]
-          ? heightToOccurrence[item.height] + 1
-          : 1;
-        fontToOccurrence[item.font] = fontToOccurrence[item.font]
-          ? fontToOccurrence[item.font] + 1
-          : 1;
-        if (item.height > maxHeight) {
-          maxHeight = item.height;
-          maxHeightFont = item.font;
-        }
-      });
-    });
-    const mostUsedHeight = parseInt(getMostUsedKey(heightToOccurrence));
-    const mostUsedFont = getMostUsedKey(fontToOccurrence);
-
-    const distanceToOccurrence: Record<string, number> = {};
-    parseResult.pages.forEach((page) => {
-      var lastItemOfMostUsedHeight: any;
-      page.items.forEach((item) => {
-        if (item.height === mostUsedHeight && item.text.trim().length > 0) {
-          if (
-            lastItemOfMostUsedHeight &&
-            item.y !== lastItemOfMostUsedHeight.y
-          ) {
-            const distance = lastItemOfMostUsedHeight.y - item.y;
-            if (distance > 0) {
-              distanceToOccurrence[distance] = distanceToOccurrence[distance]
-                ? distanceToOccurrence[distance] + 1
-                : 1;
-            }
-          }
-          lastItemOfMostUsedHeight = item;
-        } else {
-          lastItemOfMostUsedHeight = null;
-        }
-      });
-    });
-    const mostUsedDistance = parseInt(getMostUsedKey(distanceToOccurrence));
-    const fontIdToName: string[] = [];
-    const fontToFormats = new Map<string, string>();
-    this.fontMap?.forEach(function (value, key) {
-      fontIdToName.push(key + " = " + value.name);
-      const fontName = value.name.toLowerCase();
-      var format: { name: string } | null = null;
-      if (key === mostUsedFont) {
-        format = null;
-      } else if (
-        fontName.includes("bold") &&
-        (fontName.includes("oblique") || fontName.includes("italic"))
-      ) {
-        format = WordFormat.BOLD_OBLIQUE;
-      } else if (fontName.includes("bold")) {
-        format = WordFormat.BOLD;
-      } else if (fontName.includes("oblique") || fontName.includes("italic")) {
-        format = WordFormat.OBLIQUE;
-      } else if (fontName === maxHeightFont) {
-        format = WordFormat.BOLD;
-      }
-      if (format) {
-        fontToFormats.set(key, format.name);
-      }
-    });
-    fontIdToName.sort();
-
-    const newPages = parseResult.pages.map((page) => {
-      return {
-        ...page,
-        items: page.items.map((textItem) => ({ ...textItem })),
-      };
-    });
-    return new ParseResult({
-      ...parseResult,
-      pages: newPages,
-      globals: {
-        mostUsedHeight,
-        mostUsedFont,
-        mostUsedDistance,
-        maxHeight,
-        maxHeightFont,
-        fontToFormats,
-      },
-      messages: [
-        "Items per height: " + JSON.stringify(heightToOccurrence),
-        "Items per font: " + JSON.stringify(fontToOccurrence),
-        "Items per distance: " + JSON.stringify(distanceToOccurrence),
-        "Fonts:" + JSON.stringify(fontIdToName),
-      ],
-    });
-  }
-}
-
-function getMostUsedKey(keyToOccurrence: Record<string, number>): string {
-  var maxOccurence = 0;
-  var maxKey = "";
-  Object.keys(keyToOccurrence).map((element) => {
-    if (!maxKey || keyToOccurrence[element] > maxOccurence) {
-      maxOccurence = keyToOccurrence[element];
-      maxKey = element;
-    }
-  });
-  return maxKey;
-}
-
-
+/**
+ * @description First-pass transformation that scans all `TextItem`s to compute
+ * document-wide statistics stored in `parseResult.globals`: most-frequent text
+ * height (`mostUsedHeight`), most-frequent font, most-frequent line distance,
+ * maximum height, and a font-to-format map (BOLD/OBLIQUE/BOLD_OBLIQUE) derived
+ * from font names. Also deep-copies every page item so subsequent transforms
+ * never mutate the original parsed data.
+ */
+import ToTextItemTransformation from "./base/to-text-item-transform";
+import ParseResult from "../models/parse-result";
+import { WordFormat } from "../models/line-converter";
+
+export default class CalculateGlobalStats extends ToTextItemTransformation {
+  fontMap: Map<string, { name: string }> | undefined;
+
+  constructor(fontMap?: Map<string, { name: string }>) {
+    super("Calculate Global Stats");
+    this.fontMap = fontMap;
+  }
+
+  transform(parseResult: ParseResult): ParseResult {
+    const heightToOccurrence: Record<string, number> = {};
+    const fontToOccurrence: Record<string, number> = {};
+    var maxHeight = 0;
+    var maxHeightFont: string | undefined;
+    parseResult.pages.forEach((page) => {
+      page.items.forEach((item) => {
+        if (!item.height) return;
+        heightToOccurrence[item.height] = heightToOccurrence[item.height]
+          ? heightToOccurrence[item.height] + 1
+          : 1;
+        fontToOccurrence[item.font] = fontToOccurrence[item.font]
+          ? fontToOccurrence[item.font] + 1
+          : 1;
+        if (item.height > maxHeight) {
+          maxHeight = item.height;
+          maxHeightFont = item.font;
+        }
+      });
+    });
+    const mostUsedHeight = parseInt(getMostUsedKey(heightToOccurrence));
+    const mostUsedFont = getMostUsedKey(fontToOccurrence);
+
+    const distanceToOccurrence: Record<string, number> = {};
+    parseResult.pages.forEach((page) => {
+      var lastItemOfMostUsedHeight: any;
+      page.items.forEach((item) => {
+        if (item.height === mostUsedHeight && item.text.trim().length > 0) {
+          if (
+            lastItemOfMostUsedHeight &&
+            item.y !== lastItemOfMostUsedHeight.y
+          ) {
+            const distance = lastItemOfMostUsedHeight.y - item.y;
+            if (distance > 0) {
+              distanceToOccurrence[distance] = distanceToOccurrence[distance]
+                ? distanceToOccurrence[distance] + 1
+                : 1;
+            }
+          }
+          lastItemOfMostUsedHeight = item;
+        } else {
+          lastItemOfMostUsedHeight = null;
+        }
+      });
+    });
+    const mostUsedDistance = parseInt(getMostUsedKey(distanceToOccurrence));
+    const fontIdToName: string[] = [];
+    const fontToFormats = new Map<string, string>();
+    this.fontMap?.forEach(function (value, key) {
+      fontIdToName.push(key + " = " + value.name);
+      const fontName = value.name.toLowerCase();
+      var format: { name: string } | null = null;
+      if (key === mostUsedFont) {
+        format = null;
+      } else if (
+        fontName.includes("bold") &&
+        (fontName.includes("oblique") || fontName.includes("italic"))
+      ) {
+        format = WordFormat.BOLD_OBLIQUE;
+      } else if (fontName.includes("bold")) {
+        format = WordFormat.BOLD;
+      } else if (fontName.includes("oblique") || fontName.includes("italic")) {
+        format = WordFormat.OBLIQUE;
+      } else if (fontName === maxHeightFont) {
+        format = WordFormat.BOLD;
+      }
+      if (format) {
+        fontToFormats.set(key, format.name);
+      }
+    });
+    fontIdToName.sort();
+
+    const newPages = parseResult.pages.map((page) => {
+      return {
+        ...page,
+        items: page.items.map((textItem) => ({ ...textItem })),
+      };
+    });
+    return new ParseResult({
+      ...parseResult,
+      pages: newPages,
+      globals: {
+        mostUsedHeight,
+        mostUsedFont,
+        mostUsedDistance,
+        maxHeight,
+        maxHeightFont,
+        fontToFormats,
+      },
+      messages: [
+        "Items per height: " + JSON.stringify(heightToOccurrence),
+        "Items per font: " + JSON.stringify(fontToOccurrence),
+        "Items per distance: " + JSON.stringify(distanceToOccurrence),
+        "Fonts:" + JSON.stringify(fontIdToName),
+      ],
+    });
+  }
+}
+
+function getMostUsedKey(keyToOccurrence: Record<string, number>): string {
+  var maxOccurence = 0;
+  var maxKey = "";
+  Object.keys(keyToOccurrence).map((element) => {
+    if (!maxKey || keyToOccurrence[element] > maxOccurence) {
+      maxOccurence = keyToOccurrence[element];
+      maxKey = element;
+    }
+  });
+  return maxKey;
+}
+
+

package/src/transforms/line-item/compact-lines.ts

@@ -1,92 +1,92 @@
-/**
- * @description Line-item transformation that collapses `TextItem`s sharing the
- * same y-coordinate into single `LineItem`s via `TextItemLineGrouper` (line
- * bucketing) and `LineConverter` (word/format detection). Simultaneously detects
- * and records footnote links, footnote definitions, hyperlinks, and bold/italic
- * word counts, tagging items as ADDED/REMOVED for debug rendering.
- */
-import ToLineItemTransformation from "../base/to-line-item-transform";
-import ParseResult from "../../models/parse-result";
-import LineItem from "../../models/line-item";
-import TextItemLineGrouper from "../../models/text-item-line-grouper";
-import LineConverter from "../../models/line-converter";
-import BlockType from "../../models/block-type";
-import { REMOVED_ANNOTATION, ADDED_ANNOTATION } from "../../models/annotation";
-
-// gathers text items on the same y line to one line item
-export default class CompactLines extends ToLineItemTransformation {
-  constructor() {
-    super("Compact To Lines");
-  }
-
-  transform(parseResult: ParseResult): ParseResult {
-    const { mostUsedDistance, fontToFormats } = parseResult.globals;
-    const foundFootnotes = [];
-    const foundFootnoteLinks = [];
-    var linkCount = 0;
-    var formattedWords = 0;
-
-    const lineGrouper = new TextItemLineGrouper({
-      mostUsedDistance: mostUsedDistance,
-    });
-    const lineCompactor = new LineConverter(fontToFormats);
-
-    parseResult.pages.forEach((page) => {
-      if (page.items.length > 0) {
-        const lineItems = [];
-        const textItemsGroupedByLine = lineGrouper.group(page.items);
-        textItemsGroupedByLine.forEach((lineTextItems) => {
-          const lineItem = lineCompactor.compact(lineTextItems);
-          if (lineTextItems.length > 1) {
-            lineItem.annotation = ADDED_ANNOTATION;
-            lineTextItems.forEach((item) => {
-              item.annotation = REMOVED_ANNOTATION;
-              lineItems.push(
-                new LineItem({
-                  ...item,
-                }),
-              );
-            });
-          }
-          if (lineItem.words.length === 0) {
-            lineItem.annotation = REMOVED_ANNOTATION;
-          }
-          lineItems.push(lineItem);
-
-          if (lineItem.parsedElements.formattedWords) {
-            formattedWords += lineItem.parsedElements.formattedWords;
-          }
-          if (lineItem.parsedElements.containLinks) {
-            linkCount++;
-          }
-          if (lineItem.parsedElements.footnoteLinks.length > 0) {
-            const footnoteLinks = lineItem.parsedElements.footnoteLinks.map(
-              (footnoteLink) => ({ footnoteLink, page: page.index + 1 }),
-            );
-            foundFootnoteLinks.push.apply(foundFootnoteLinks, footnoteLinks);
-          }
-          if (lineItem.parsedElements.footnotes.length > 0) {
-            lineItem.type = BlockType.FOOTNOTES;
-            const footnotes = lineItem.parsedElements.footnotes.map(
-              (footnote) => ({ footnote, page: page.index + 1 }),
-            );
-            foundFootnotes.push.apply(foundFootnotes, footnotes);
-          }
-        });
-        page.items = lineItems;
-      }
-    });
-
-    return new ParseResult({
-      ...parseResult,
-      messages: [
-        "Detected " + formattedWords + " formatted words",
-        "Found " + linkCount + " links",
-        "Detected " + foundFootnoteLinks.length + " footnotes links",
-        "Detected " + foundFootnotes.length + " footnotes",
-      ],
-    });
-  }
-}
-
-
+/**
+ * @description Line-item transformation that collapses `TextItem`s sharing the
+ * same y-coordinate into single `LineItem`s via `TextItemLineGrouper` (line
+ * bucketing) and `LineConverter` (word/format detection). Simultaneously detects
+ * and records footnote links, footnote definitions, hyperlinks, and bold/italic
+ * word counts, tagging items as ADDED/REMOVED for debug rendering.
+ */
+import ToLineItemTransformation from "../base/to-line-item-transform";
+import ParseResult from "../../models/parse-result";
+import LineItem from "../../models/line-item";
+import TextItemLineGrouper from "../../models/text-item-line-grouper";
+import LineConverter from "../../models/line-converter";
+import BlockType from "../../models/block-type";
+import { REMOVED_ANNOTATION, ADDED_ANNOTATION } from "../../models/annotation";
+
+// gathers text items on the same y line to one line item
+export default class CompactLines extends ToLineItemTransformation {
+  constructor() {
+    super("Compact To Lines");
+  }
+
+  transform(parseResult: ParseResult): ParseResult {
+    const { mostUsedDistance, fontToFormats } = parseResult.globals;
+    const foundFootnotes = [];
+    const foundFootnoteLinks = [];
+    var linkCount = 0;
+    var formattedWords = 0;
+
+    const lineGrouper = new TextItemLineGrouper({
+      mostUsedDistance: mostUsedDistance,
+    });
+    const lineCompactor = new LineConverter(fontToFormats);
+
+    parseResult.pages.forEach((page) => {
+      if (page.items.length > 0) {
+        const lineItems = [];
+        const textItemsGroupedByLine = lineGrouper.group(page.items);
+        textItemsGroupedByLine.forEach((lineTextItems) => {
+          const lineItem = lineCompactor.compact(lineTextItems);
+          if (lineTextItems.length > 1) {
+            lineItem.annotation = ADDED_ANNOTATION;
+            lineTextItems.forEach((item) => {
+              item.annotation = REMOVED_ANNOTATION;
+              lineItems.push(
+                new LineItem({
+                  ...item,
+                }),
+              );
+            });
+          }
+          if (lineItem.words.length === 0) {
+            lineItem.annotation = REMOVED_ANNOTATION;
+          }
+          lineItems.push(lineItem);
+
+          if (lineItem.parsedElements.formattedWords) {
+            formattedWords += lineItem.parsedElements.formattedWords;
+          }
+          if (lineItem.parsedElements.containLinks) {
+            linkCount++;
+          }
+          if (lineItem.parsedElements.footnoteLinks.length > 0) {
+            const footnoteLinks = lineItem.parsedElements.footnoteLinks.map(
+              (footnoteLink) => ({ footnoteLink, page: page.index + 1 }),
+            );
+            foundFootnoteLinks.push.apply(foundFootnoteLinks, footnoteLinks);
+          }
+          if (lineItem.parsedElements.footnotes.length > 0) {
+            lineItem.type = BlockType.FOOTNOTES;
+            const footnotes = lineItem.parsedElements.footnotes.map(
+              (footnote) => ({ footnote, page: page.index + 1 }),
+            );
+            foundFootnotes.push.apply(foundFootnotes, footnotes);
+          }
+        });
+        page.items = lineItems;
+      }
+    });
+
+    return new ParseResult({
+      ...parseResult,
+      messages: [
+        "Detected " + formattedWords + " formatted words",
+        "Found " + linkCount + " links",
+        "Detected " + foundFootnoteLinks.length + " footnotes links",
+        "Detected " + foundFootnotes.length + " footnotes",
+      ],
+    });
+  }
+}
+
+