Automatically populate previous metadata

frontend/src/components/tools/changeMetadata/ChangeMetadataSettings.tsx

@@ -1,12 +1,16 @@
 import { Stack, TextInput, Select, Checkbox, Button, Group, Divider, Text } from "@mantine/core";
 import { useTranslation } from "react-i18next";
-import { ChangeMetadataParameters, TrappedStatus } from "../../../hooks/tools/changeMetadata/useChangeMetadataParameters";
+import { useEffect, useState } from "react";
+import { ChangeMetadataParameters } from "../../../hooks/tools/changeMetadata/useChangeMetadataParameters";
+import { TrappedStatus } from "../../../types/metadata";
+import { PDFMetadataService } from "../../../services/pdfMetadataService";
+import { useSelectedFiles } from "../../../contexts/file/fileHooks";
 
 interface ChangeMetadataSettingsProps {
   parameters: ChangeMetadataParameters;
   onParameterChange: <K extends keyof ChangeMetadataParameters>(key: K, value: ChangeMetadataParameters[K]) => void;
   disabled?: boolean;
-  addCustomMetadata: () => void;
+  addCustomMetadata: (key?: string, value?: string) => void;
   removeCustomMetadata: (id: string) => void;
   updateCustomMetadata: (id: string, key: string, value: string) => void;
 }
@@ -24,8 +28,57 @@ const ChangeMetadataSettings = ({
   updateCustomMetadata
 }: ChangeMetadataSettingsProps) => {
   const { t } = useTranslation();
+  const { selectedFiles } = useSelectedFiles();
+  const [isExtractingMetadata, setIsExtractingMetadata] = useState(false);
+  const [hasExtractedMetadata, setHasExtractedMetadata] = useState(false);
+
   const isDeleteAllEnabled = parameters.deleteAll;
-  const fieldsDisabled = disabled || isDeleteAllEnabled;
+  const fieldsDisabled = disabled || isDeleteAllEnabled || isExtractingMetadata;
+
+  // Extract metadata from first file when files change
+  useEffect(() => {
+    const extractMetadata = async () => {
+      if (selectedFiles.length === 0 || hasExtractedMetadata) {
+        return;
+      }
+
+      const firstFile = selectedFiles[0];
+      if (!firstFile) {
+        return;
+      }
+
+      setIsExtractingMetadata(true);
+      try {
+        const result = await PDFMetadataService.extractMetadata(firstFile);
+
+        if (result.success) {
+          const metadata = result.metadata;
+
+          // Pre-populate all fields with extracted metadata
+          onParameterChange('title', metadata.title);
+          onParameterChange('author', metadata.author);
+          onParameterChange('subject', metadata.subject);
+          onParameterChange('keywords', metadata.keywords);
+          onParameterChange('creator', metadata.creator);
+          onParameterChange('producer', metadata.producer);
+          onParameterChange('creationDate', metadata.creationDate);
+          onParameterChange('modificationDate', metadata.modificationDate);
+          onParameterChange('trapped', metadata.trapped);
+
+          // Set custom metadata entries directly to avoid state update timing issues
+          onParameterChange('customMetadata', metadata.customMetadata);
+
+          setHasExtractedMetadata(true);
+        }
+      } catch (error) {
+        console.warn('Failed to extract metadata:', error);
+      } finally {
+        setIsExtractingMetadata(false);
+      }
+    };
+
+    extractMetadata();
+  }, [selectedFiles, hasExtractedMetadata, onParameterChange, addCustomMetadata, updateCustomMetadata, removeCustomMetadata, parameters.customMetadata]);
 
   return (
     <Stack gap="md">
@@ -149,7 +202,7 @@ const ChangeMetadataSettings = ({
             <Button
               size="xs"
               variant="light"
-              onClick={addCustomMetadata}
+              onClick={() => addCustomMetadata()}
               disabled={fieldsDisabled}
             >
               {t('changeMetadata.customFields.add', 'Add Field')}

frontend/src/hooks/tools/changeMetadata/useChangeMetadataOperation.test.ts

@@ -1,5 +1,6 @@
 import { buildChangeMetadataFormData } from './useChangeMetadataOperation';
-import { ChangeMetadataParameters, TrappedStatus } from './useChangeMetadataParameters';
+import { ChangeMetadataParameters } from './useChangeMetadataParameters';
+import { TrappedStatus } from '../../../types/metadata';
 import { describe, expect, test } from 'vitest';
 
 describe('buildChangeMetadataFormData', () => {

frontend/src/hooks/tools/changeMetadata/useChangeMetadataParameters.test.ts

@@ -1,5 +1,6 @@
 import { renderHook, act } from '@testing-library/react';
-import { useChangeMetadataParameters, TrappedStatus } from './useChangeMetadataParameters';
+import { useChangeMetadataParameters } from './useChangeMetadataParameters';
+import { TrappedStatus } from '../../../types/metadata';
 import { describe, expect, test } from 'vitest';
 
 describe('useChangeMetadataParameters', () => {

frontend/src/hooks/tools/changeMetadata/useChangeMetadataParameters.ts

@@ -1,18 +1,7 @@
 import { BaseParameters } from '../../../types/parameters';
+import { TrappedStatus, CustomMetadataEntry } from '../../../types/metadata';
 import { useBaseParameters, BaseParametersHook } from '../shared/useBaseParameters';
 
-export enum TrappedStatus {
-  TRUE = 'True',
-  FALSE = 'False',
-  UNKNOWN = 'Unknown'
-}
-
-export interface CustomMetadataEntry {
-  key: string;
-  value: string;
-  id: string; // Format: "custom1", "custom2", etc.
-}
-
 export interface ChangeMetadataParameters extends BaseParameters {
   // Standard PDF metadata fields
   title: string;
@@ -86,7 +75,7 @@ const validateParameters = (params: ChangeMetadataParameters): boolean => {
 };
 
 export type ChangeMetadataParametersHook = BaseParametersHook<ChangeMetadataParameters> & {
-  addCustomMetadata: () => void;
+  addCustomMetadata: (key?: string, value?: string) => void;
   removeCustomMetadata: (id: string) => void;
   updateCustomMetadata: (id: string, key: string, value: string) => void;
 };
@@ -98,10 +87,10 @@ export const useChangeMetadataParameters = (): ChangeMetadataParametersHook => {
     validateFn: validateParameters,
   });
 
-  const addCustomMetadata = () => {
+  const addCustomMetadata = (key: string = '', value: string = '') => {
     const newEntry: CustomMetadataEntry = {
-      key: '',
+      key,
-      value: '',
+      value,
       id: `custom${customMetadataIdCounter++}`,
     };
 

frontend/src/services/pdfMetadataService.ts

@@ -0,0 +1,175 @@
+import { pdfWorkerManager } from './pdfWorkerManager';
+import { FileAnalyzer } from './fileAnalyzer';
+import { TrappedStatus, CustomMetadataEntry, ExtractedPDFMetadata } from '../types/metadata';
+
+export interface MetadataExtractionResult {
+  success: true;
+  metadata: ExtractedPDFMetadata;
+}
+
+export interface MetadataExtractionError {
+  success: false;
+  error: string;
+}
+
+export type MetadataExtractionResponse = MetadataExtractionResult | MetadataExtractionError;
+
+/**
+ * Utility to format PDF date strings to required format (yyyy/MM/dd HH:mm:ss)
+ * Handles PDF date format: "D:YYYYMMDDHHmmSSOHH'mm'" or standard date strings
+ */
+function formatPDFDate(dateString: unknown): string {
+  if (!dateString || typeof dateString !== 'string') {
+    return '';
+  }
+
+  try {
+    let date: Date;
+
+    // Check if it's a PDF date format (starts with "D:")
+    if (dateString.startsWith('D:')) {
+      // Parse PDF date format: D:YYYYMMDDHHmmSSOHH'mm'
+      const dateStr = dateString.substring(2); // Remove "D:"
+
+      // Extract date parts
+      const year = parseInt(dateStr.substring(0, 4));
+      const month = parseInt(dateStr.substring(4, 6));
+      const day = parseInt(dateStr.substring(6, 8));
+      const hour = parseInt(dateStr.substring(8, 10)) || 0;
+      const minute = parseInt(dateStr.substring(10, 12)) || 0;
+      const second = parseInt(dateStr.substring(12, 14)) || 0;
+
+      // Create date object (month is 0-indexed)
+      date = new Date(year, month - 1, day, hour, minute, second);
+    } else {
+      // Try parsing as regular date string
+      date = new Date(dateString);
+    }
+
+    if (isNaN(date.getTime())) {
+      return '';
+    }
+
+    const year = date.getFullYear();
+    const month = String(date.getMonth() + 1).padStart(2, '0');
+    const day = String(date.getDate()).padStart(2, '0');
+    const hours = String(date.getHours()).padStart(2, '0');
+    const minutes = String(date.getMinutes()).padStart(2, '0');
+    const seconds = String(date.getSeconds()).padStart(2, '0');
+
+    return `${year}/${month}/${day} ${hours}:${minutes}:${seconds}`;
+  } catch {
+    return '';
+  }
+}
+
+/**
+ * Convert PDF.js trapped value to TrappedStatus enum
+ * PDF.js returns trapped as { name: "True" | "False" } object
+ */
+function convertTrappedStatus(trapped: unknown): TrappedStatus {
+  if (trapped && typeof trapped === 'object' && 'name' in trapped) {
+    const name = (trapped as Record<string, string>).name?.toLowerCase();
+    if (name === 'true') return TrappedStatus.TRUE;
+    if (name === 'false') return TrappedStatus.FALSE;
+  }
+  return TrappedStatus.UNKNOWN;
+}
+
+/**
+ * Extract custom metadata fields from PDF.js info object
+ * Custom metadata is nested under the "Custom" key
+ */
+function extractCustomMetadata(info: Record<string, unknown>): CustomMetadataEntry[] {
+  const customMetadata: CustomMetadataEntry[] = [];
+  let customIdCounter = 1;
+
+
+  // Check if there's a Custom object containing the custom metadata
+  if (info.Custom && typeof info.Custom === 'object' && info.Custom !== null) {
+    const customObj = info.Custom as Record<string, unknown>;
+
+    Object.entries(customObj).forEach(([key, value]) => {
+      if (value != null && value !== '') {
+        const entry = {
+          key,
+          value: String(value),
+          id: `custom${customIdCounter++}`
+        };
+        customMetadata.push(entry);
+      }
+    });
+  }
+
+  return customMetadata;
+}
+
+/**
+ * Service to extract metadata from PDF files using PDF.js
+ */
+export class PDFMetadataService {
+  /**
+   * Extract all metadata from a PDF file
+   * Returns a result object with success/error state
+   */
+  static async extractMetadata(file: File): Promise<MetadataExtractionResponse> {
+    // Use existing PDF validation
+    const isValidPDF = await FileAnalyzer.isValidPDF(file);
+    if (!isValidPDF) {
+      return {
+        success: false,
+        error: 'File is not a valid PDF'
+      };
+    }
+
+    let pdfDoc: any = null;
+
+    try {
+      const arrayBuffer = await file.arrayBuffer();
+      pdfDoc = await pdfWorkerManager.createDocument(arrayBuffer, {
+        disableAutoFetch: true,
+        disableStream: true
+      });
+
+      const metadata = await pdfDoc.getMetadata();
+      const info = metadata.info || {};
+
+      // Safely extract metadata with proper type checking
+      const extractedMetadata: ExtractedPDFMetadata = {
+        title: typeof info.Title === 'string' ? info.Title : '',
+        author: typeof info.Author === 'string' ? info.Author : '',
+        subject: typeof info.Subject === 'string' ? info.Subject : '',
+        keywords: typeof info.Keywords === 'string' ? info.Keywords : '',
+        creator: typeof info.Creator === 'string' ? info.Creator : '',
+        producer: typeof info.Producer === 'string' ? info.Producer : '',
+        creationDate: formatPDFDate(info.CreationDate),
+        modificationDate: formatPDFDate(info.ModDate),
+        trapped: convertTrappedStatus(info.Trapped),
+        customMetadata: extractCustomMetadata(info)
+      };
+
+      return {
+        success: true,
+        metadata: extractedMetadata
+      };
+
+    } catch (error) {
+      const errorMessage = error instanceof Error ? error.message : 'Unknown error';
+
+      return {
+        success: false,
+        error: `Failed to extract PDF metadata: ${errorMessage}`
+      };
+
+    } finally {
+      // Ensure cleanup even if extraction fails
+      if (pdfDoc) {
+        try {
+          pdfWorkerManager.destroyDocument(pdfDoc);
+        } catch (cleanupError) {
+          console.warn('Failed to cleanup PDF document:', cleanupError);
+        }
+      }
+    }
+  }
+}

frontend/src/types/metadata.ts

@@ -0,0 +1,24 @@
+export enum TrappedStatus {
+  TRUE = 'True',
+  FALSE = 'False',
+  UNKNOWN = 'Unknown'
+}
+
+export interface CustomMetadataEntry {
+  key: string;
+  value: string;
+  id: string; // For React uniqueness
+}
+
+export interface ExtractedPDFMetadata {
+  title: string;
+  author: string;
+  subject: string;
+  keywords: string;
+  creator: string;
+  producer: string;
+  creationDate: string;
+  modificationDate: string;
+  trapped: TrappedStatus;
+  customMetadata: CustomMetadataEntry[];
+}