Merge pull request #878 from grasdk/feature/metadata-extra-refactor

removed ts-node-iptc dependency.

package-lock.json

@@ -26,7 +26,6 @@
         "nodemailer": "6.9.4",
         "reflect-metadata": "0.1.13",
         "sharp": "0.31.3",
-        "ts-node-iptc": "1.0.11",
         "typeconfig": "2.2.11",
         "typeorm": "0.3.12",
         "xml2js": "0.6.2"

package.json

@@ -53,7 +53,6 @@
     "nodemailer": "6.9.4",
     "reflect-metadata": "0.1.13",
     "sharp": "0.31.3",
-    "ts-node-iptc": "1.0.11",
     "typeconfig": "2.2.11",
     "typeorm": "0.3.12",
     "xml2js": "0.6.2"

src/backend/model/fileaccess/MetadataLoader.ts

@@ -12,7 +12,6 @@ import { FfprobeData } from 'fluent-ffmpeg';
 import { FileHandle } from 'fs/promises';
 import * as util from 'node:util';
 import * as path from 'path';
-import { IptcParser } from 'ts-node-iptc';
 import { Utils } from '../../../common/Utils';
 import { FFmpegFactory } from '../FFmpegFactory';
 import { ExtensionDecorator } from '../extension/ExtensionDecorator';
@@ -181,7 +180,7 @@ export class MetadataLoader {
       icc: false,
       jfif: false, //not needed and not supported for png
       ihdr: true,
-      iptc: false, //exifr reads UTF8-encoded data wrongly, using IptcParser instead
+      iptc: true, 
       exif: true,
       gps: true,
       reviveValues: false, //don't convert timestamps
@@ -221,46 +220,6 @@ export class MetadataLoader {
         await fileHandle.close();
       }
       try {
-
-
-        try { //Parse iptc data using the IptcParser, which works correctly for both UTF-8 and ASCII
-          const iptcData = IptcParser.parse(data);
-          if (iptcData.country_or_primary_location_name) {
-            metadata.positionData = metadata.positionData || {};
-            metadata.positionData.country =
-              iptcData.country_or_primary_location_name
-                .replace(/\0/g, '')
-                .trim();
-          }
-          if (iptcData.province_or_state) {
-            metadata.positionData = metadata.positionData || {};
-            metadata.positionData.state = iptcData.province_or_state
-              .replace(/\0/g, '')
-              .trim();
-          }
-          if (iptcData.city) {
-            metadata.positionData = metadata.positionData || {};
-            metadata.positionData.city = iptcData.city
-              .replace(/\0/g, '')
-              .trim();
-          }
-          if (iptcData.object_name) {
-            metadata.title = iptcData.object_name.replace(/\0/g, '').trim();
-          }
-          if (iptcData.caption) {
-            metadata.caption = iptcData.caption.replace(/\0/g, '').trim();
-          }
-          if (Array.isArray(iptcData.keywords)) {
-            metadata.keywords = iptcData.keywords;
-          }
-
-          if (iptcData.date_time) {
-            metadata.creationDate = iptcData.date_time.getTime();
-          }
-        } catch (err) {
-          // Logger.debug(LOG_TAG, 'Error parsing iptc data', fullPath, err);
-        }
-
         try {
           const exif = await exifr.parse(data, exifrOptions);
           MetadataLoader.mapMetadata(metadata, exif);
@@ -370,20 +329,35 @@ export class MetadataLoader {
         }
       }
     }
+    if (exif.iptc &&
+      exif.iptc.Keywords &&
+      exif.iptc.Keywords.length > 0) {
+      const subj = Array.isArray(exif.iptc.Keywords) ? exif.iptc.Keywords : [exif.iptc.Keywords];
+      if (metadata.keywords === undefined) {
+        metadata.keywords = [];
+      }
+      for (let kw of subj) {
+        kw = Utils.asciiToUTF8(kw);
+        if (metadata.keywords.indexOf(kw) === -1) {
+          metadata.keywords.push(kw);
+        }
+      }
+    }
   }
 
   private static mapTitle(metadata: PhotoMetadata, exif: any) {
-    metadata.title = exif.dc?.title?.value || metadata.title || exif.photoshop?.Headline || exif.acdsee?.caption; //acdsee caption holds the title when data is saved by digikam. Used as last resort if iptc and dc do not contain the data
+    metadata.title = exif.dc?.title?.value || Utils.asciiToUTF8(exif.iptc?.ObjectName) || metadata.title || exif.photoshop?.Headline || exif.acdsee?.caption; //acdsee caption holds the title when data is saved by digikam. Used as last resort if iptc and dc do not contain the data
   }
 
   private static mapCaption(metadata: PhotoMetadata, exif: any) {
-    metadata.caption = exif.dc?.description?.value || metadata.caption || exif.ifd0?.ImageDescription || exif.exif?.UserComment?.value || exif.Iptc4xmpCore?.ExtDescrAccessibility?.value ||exif.acdsee?.notes;
+    metadata.caption = exif.dc?.description?.value || Utils.asciiToUTF8(exif.iptc?.Caption) || metadata.caption || exif.ifd0?.ImageDescription || exif.exif?.UserComment?.value || exif.Iptc4xmpCore?.ExtDescrAccessibility?.value ||exif.acdsee?.notes;
   }
 
   private static mapTimestampAndOffset(metadata: PhotoMetadata, exif: any) {
     metadata.creationDate = Utils.timestampToMS(exif?.photoshop?.DateCreated, null) ||
     Utils.timestampToMS(exif?.xmp?.CreateDate, null) ||
     Utils.timestampToMS(exif?.xmp?.ModifyDate, null) ||
+    Utils.timestampToMS(Utils.toIsoTimestampString(exif?.iptc?.DateCreated, exif?.iptc?.TimeCreated), null) ||
     metadata.creationDate;
 
     metadata.creationDateOffset = Utils.timestampToOffsetString(exif?.photoshop?.DateCreated) ||
@@ -490,24 +464,15 @@ export class MetadataLoader {
 
   private static mapToponyms(metadata: PhotoMetadata, exif: any) {
     //Function to convert html code for special characters into their corresponding character (used in exif.photoshop-section)
-    const unescape = (tag: string) => {
+
-      return tag.replace(/&#([0-9]{1,3});/gi, function (match, numStr) {
+    metadata.positionData = metadata.positionData || {};
-        return String.fromCharCode(parseInt(numStr, 10));
+    metadata.positionData.country = Utils.asciiToUTF8(exif.iptc?.Country) || Utils.decodeHTMLChars(exif.photoshop?.Country);
-      });
+    metadata.positionData.state = Utils.asciiToUTF8(exif.iptc?.State) || Utils.decodeHTMLChars(exif.photoshop?.State);
-    }
+    metadata.positionData.city = Utils.asciiToUTF8(exif.iptc?.City) || Utils.decodeHTMLChars(exif.photoshop?.City);
-    //photoshop section sometimes has City, Country and State
+    if (metadata.positionData) {
-    if (exif.photoshop) {
+      Utils.removeNullOrEmptyObj(metadata.positionData);
-      if (!metadata.positionData?.country && exif.photoshop.Country) {
+      if (Object.keys(metadata.positionData).length === 0) {
-        metadata.positionData = metadata.positionData || {};
+        delete metadata.positionData;
-        metadata.positionData.country = unescape(exif.photoshop.Country);
-      }
-      if (!metadata.positionData?.state && exif.photoshop.State) {
-        metadata.positionData = metadata.positionData || {};
-        metadata.positionData.state = unescape(exif.photoshop.State);
-      }
-      if (!metadata.positionData?.city && exif.photoshop.City) {
-        metadata.positionData = metadata.positionData || {};
-        metadata.positionData.city = unescape(exif.photoshop.City);
       }
     }
   }

src/common/HTMLCharCodes.ts

@@ -0,0 +1,118 @@
+interface HTMLCharDictionary {
+  [key: string]: string;
+}
+
+export const HTMLChar: HTMLCharDictionary = {
+  """: "\"",
+  "&": "&",
+  "&lt;": "<",
+  "&gt;": ">",
+  "&nbsp;": " ",
+  "&iexcl;": "¡",
+  "&cent;": "¢",
+  "&pound;": "£",
+  "&curren;": "¤",
+  "&yen;": "¥",
+  "&brvbar;": "¦",
+  "&sect;": "§",
+  "&uml;": "¨",
+  "&copy;": "©",
+  "&reg;": "®",
+  "&trade;": "™",
+  "&ordf;": "ª",
+  "&laquo;": "«",
+  "&not;": "¬",
+  "&shy;": "­",
+  "&macr;": "¯",
+  "&deg;": "°",
+  "&plusmn;": "±",
+  "&sup2;": "²",
+  "&sup3;": "³",
+  "&acute;": "´",
+  "&micro;": "µ",
+  "&para;": "¶",
+  "&middot;": "·",
+  "&cedil;": "¸",
+  "&sup1;": "¹",
+  "&ordm;": "º",
+  "&raquo;": "»",
+  "&frac14;": "¼",
+  "&frac12;": "½",
+  "&frac34;": "¾",
+  "&iquest;": "¿",
+  "&times;": "×",
+  "&divide;": "÷",
+  "&ETH;": "Ð",
+  "&eth;": "ð",
+  "&THORN;": "Þ",
+  "&thorn;": "þ",
+  "&AElig;": "Æ",
+  "&aelig;": "æ",
+  "&OElig;": "Œ",
+  "&oelig;": "œ",
+  "&Aring;": "Å",
+  "&Oslash;": "Ø",
+  "&Ccedil;": "Ç",
+  "&ccedil;": "ç",
+  "&szlig;": "ß",
+  "&Ntilde;": "Ñ",
+  "&ntilde;": "ñ",
+  "&Aacute;": "Á",
+  "&Agrave;": "À",
+  "&Acirc;": "Â",
+  "&Auml;": "Ä",
+  "&Atilde;": "Ã",
+  "&aacute;": "á",
+  "&agrave;": "à",
+  "&acirc;": "â",
+  "&auml;": "ä",
+  "&atilde;": "ã",
+  "&aring;": "å",
+  "&Eacute;": "É",
+  "&Egrave;": "È",
+  "&Ecirc;": "Ê",
+  "&Euml;": "Ë",
+  "&Etilde;": "Ẽ",
+  "&eacute;": "é",
+  "&egrave;": "è",
+  "&ecirc;": "ê",
+  "&euml;": "ë",
+  "&Iacute;": "Í",
+  "&Igrave;": "Ì",
+  "&Icirc;": "Î",
+  "&Iuml;": "Ï",
+  "&Itilde;": "Ĩ",
+  "&iacute;": "í",
+  "&igrave;": "ì",
+  "&icirc;": "î",
+  "&iuml;": "ï",
+  "&itilde;": "ĩ",
+  "&Oacute;": "Ó",
+  "&Ograve;": "Ò",
+  "&Ocirc;": "Ô",
+  "&Ouml;": "Ö",
+  "&Otilde;": "Õ",
+  "&oacute;": "ó",
+  "&ograve;": "ò",
+  "&ocirc;": "ô",
+  "&ouml;": "ö",
+  "&otilde;": "õ",
+  "&Uacute;": "Ú",
+  "&Ugrave;": "Ù",
+  "&Ucirc;": "Û",
+  "&Uuml;": "Ü",
+  "&Utilde;": "Ũ",
+  "&Uring;": "Ů",
+  "&uacute;": "ú",
+  "&ugrave;": "ù",
+  "&ucirc;": "û",
+  "&uuml;": "ü",
+  "&utilde;": "ũ",
+  "&uring;": "ů",
+  "&Yacute;": "Ý",
+  "&Ycirc;": "Ŷ",
+  "&Yuml;": "Ÿ",
+  "&yacute;": "ý",
+  "&ycirc;": "ŷ",
+  "&yuml;": "ÿ"
+};

src/common/Utils.ts

@@ -1,3 +1,5 @@
+import { HTMLChar } from './HTMLCharCodes';
+
 export class Utils {
   static GUID(): string {
     const s4 = (): string =>
@@ -97,6 +99,25 @@ export class Utils {
     return d.getUTCFullYear() + '-' + d.getUTCMonth() + '-' + d.getUTCDate();
   }
 
+  static toIsoTimestampString(YYYYMMDD: string, hhmmss: string): string {
+    if (YYYYMMDD && hhmmss) {
+      // Regular expression to match YYYYMMDD format
+      const dateRegex = /^(\d{4})(\d{2})(\d{2})$/;
+      // Regular expression to match hhmmss+/-ohom format
+      const timeRegex = /^(\d{2})(\d{2})(\d{2})([+-]\d{2})?(\d{2})?$/;
+      const [, year, month, day] = YYYYMMDD.match(dateRegex);
+      const [, hour, minute, second, offsetHour, offsetMinute] = hhmmss.match(timeRegex);
+      const isoTimestamp = `${year}-${month}-${day}T${hour}:${minute}:${second}`;
+      if (offsetHour && offsetMinute) {
+        return isoTimestamp + `${offsetHour}:${offsetMinute}`;
+      } else {
+        return isoTimestamp;
+      }
+    } else {
+      return undefined;
+    }
+  }
+
 
   static makeUTCMidnight(d: number | Date) {
     if (!(d instanceof Date)) {
@@ -125,7 +146,7 @@ export class Utils {
   }
 
   //function to convert timestamp into milliseconds taking offset into account
-  static timestampToMS(timestamp: string, offset: string) {
+  static timestampToMS(timestamp: string, offset: string): number {
     if (!timestamp) {
       return undefined;
     }
@@ -371,6 +392,31 @@ export class Utils {
     return curr;
   }
 
+  public static asciiToUTF8(text: string): string {
+    if (text) {
+      return Buffer.from(text, 'ascii').toString('utf-8');
+    } else {
+      return text;
+    }
+  }
+
+
+
+  public static decodeHTMLChars(text: string): string {
+    if (text) {
+      const newtext = text.replace(/&#([0-9]{1,3});/gi, function (match, numStr) {
+        return String.fromCharCode(parseInt(numStr, 10));
+      });
+      return newtext.replace(/&[^;]+;/g, function (match) {
+        const char = HTMLChar[match];
+        return char ? char : match;
+      });
+    } else {
+      return text;
+    }
+  }
+
+
   public static isUInt32(value: number, max = 4294967295): boolean {
     value = parseInt('' + value, 10);
     return !isNaN(value) && value >= 0 && value <= max;

test/backend/assets/sidecar/testimagedesc1.json

@@ -41,9 +41,9 @@
       "latitude": 37.871093,
       "longitude": -122.25678
     },
-    "city": "test city őúéáűóöí-.,)(=",
+    "city": "test city őúéáűóöí-.,)(=/%!+\"'",
     "country": "test country őúéáűóöí-.,)(=/%!+\"'",
-    "state": "test state őúéáűóöí-.,)("
+    "state": "test state őúéáűóöí-.,)(=/%!+\"'"
   },
   "rating": 3,
   "size": {