pdf-extractor-service

Uses Spring Boot Web, PDFBox and Tesseract to extract text, tables and images from PDF file

Steps

• Build project:

  mvn clean install

• In local:

  • Run service :

    mvn spring-boot:run

• In docker container:

  • Build project: will use the PDFExtractorService.jar created above

    docker build -f Dockerfile -t pdf-extractor-service .

  • Run service:

    docker run -p 8086:8086 pdf-extractor-service

• Configure TessData and Set Env Vars (for local environment only):

  cd src/main/resources/tessdata

  Either git pull or download data from: https://github.com/tesseract-ocr/tessdata

  Set env vars as:

  TESSDATA_PREFIX=src/main/resources/tessdata

  ENV=local

• Use REST client like Postman and make POST Call:

  • Below request helps to extract text content from PDF file:

    • Method: POST
    • URL: http://localhost:8086/api/v1/pdf/extract-text
    • Body: Param: file, Value: ERDM_PR_Test.pdf

  • Response:

    `{
    "textContent": "ERDM \r\n \r\nPerformance Return Domain \r\n \r\nStatistics: \r\nSN Portfolios EffectiveDate StdDev ReturnPercent \r\n1 ABC 2024-04-14 0.5 50.00% \r\n2 DEF 2024-04-14 0.05 95.00% \r\n3 DEF 2024-04-14 0.01 99.00% \r\n4 GHI 2024-04-15 0.04 96.00% \r\n5 JKL 2024-04-16 0.2 80.00% \r\n \r\n",
    "tableContent": null,
    "imageContent": null
    }`
    

• Similarly, there are endpoints for extract-tables, extract-images and extract-all

  • Below request helps to extract text from tables and charts:

    • Method: POST
    • URL: http://localhost:8086/api/v1/pdf/extract-tables
    • Body: Param: file, Value: ERDM_PR_Test.pdf

  • Below request helps to extract text from images:

    • Method: POST
    • URL: http://localhost:8086/api/v1/pdf/extract-images
    • Body: Param: file, Value: ERDM_PR_Test.pdf

  • Below request helps to extract text from normal text, tables, charts and images:

    • Method: POST
    • URL: http://localhost:8086/api/v1/pdf/extract-all
    • Body: Param: file, Value: ERDM_PR_Test.pdf

  • Below request helps to extract text from specific rectangular area or section:

    • Method: POST
    • URL: http://localhost:8086/api/v1/pdf/extract-section
    • Body:
      • Param: file, Value: ERDM_PR_Test.pdf
      • Param: x, Value: 50
      • Param: y, Value: 50
      • Param: width, Value: 200
      • Param: height, Value: 200