Query PDFs Flexibly: Ask Questions or Extract JSON with UniPDF & AI

import (
	"bytes"
	"encoding/json"
	"fmt"
	"io"
	"mime/multipart"
	"net/http"
	"os"
	"strings"
	"github.com/unidoc/unipdf/v3/common/license"
	"github.com/unidoc/unipdf/v3/model"
)

func init() {
	// Make sure to load your metered License API key prior to using the library.
	// If you need a key, you can sign up and create a free one at https://cloud.unidoc.io
	err := license.SetMeteredKey(os.Getenv(`UNIDOC_LICENSE_API_KEY`))
	if err != nil {
		panic(err)
	}
}

func main() {
	if len(os.Args) < 2 {
		fmt.Printf("Usage: go run main.go input.pdf\n")
		os.Exit(1)
	}
	inputPath := os.Args[1]
	question := "What is the latest experience?"
	pages, err := splitPdfPages(inputPath)
	if err != nil {
		fmt.Printf("Error: %v\n", err)
		os.Exit(1)
	}
	answers := make([][]string, len(pages))
	for i, page := range pages {
		resp, err := processPageToAI(page, question)
		if err != nil {
			fmt.Printf("Error: %v\n", err)
			os.Exit(1)
		}
		defer resp.Body.Close()
		if resp.StatusCode != http.StatusOK {
			fmt.Printf("failed to upload file: %s", resp.Status)
			os.Exit(1)
		}
		responseBody, err := io.ReadAll(resp.Body)
		if err != nil {
			panic(err)
		}
		var jsonResponse map[string]interface{}
		err = json.Unmarshal(responseBody, &jsonResponse)
		if err != nil {
			panic(err)
		}
		jsonFormattedResponse, err := json.MarshalIndent(jsonResponse, "", "  ")
		if err != nil {
			panic(err)
		}

		// We print the JSON response for each page.
		fmt.Printf("Page %d:\n", i+1)
		fmt.Println(string(jsonFormattedResponse))
		var aiRes []string
		if files, err := jsonResponse["files"].([]interface{}); err {
			for _, file := range files {
				if fileMap, ok := file.(map[string]interface{}); ok {
					if cvContent, ok := fileMap["cv_content"].(map[string]interface{}); ok {
						if fullName, ok := cvContent["full_name"].(string); ok {
							aiRes = append(aiRes, fullName)
						}
						if answer, ok := cvContent["answer"].(string); ok {
							aiRes = append(aiRes, question, answer)
						}
					}
				}
			}
		}
		answers[i] = aiRes
	}
    // Print out the full name, question and answer
	for i, answer := range answers {
		fmt.Printf("%d. %+s\n", i+1, strings.Join(answer, ";"))
	}
}

// splitPdfPages splits the input PDF file.
func splitPdfPages(inputPath string) ([]bytes.Buffer, error) {
	pdfReader, f, err := model.NewPdfReaderFromFile(inputPath, nil)
	if err != nil {
		return nil, err
	}
	defer f.Close()
	numPages, err := pdfReader.GetNumPages()
	if err != nil {
		return nil, err
	}
	pdfPages := make([]bytes.Buffer, numPages)
	for i := 1; i <= numPages; i++ {
		pageNum := i
		page, err := pdfReader.GetPage(pageNum)
		if err != nil {
			return nil, err
		}
		var buf bytes.Buffer
		pdfWriter := model.NewPdfWriter()
		pdfWriter.AddPage(page)
		pdfWriter.Write(&buf)
		pdfPages[i-1] = buf
	}
	return pdfPages, nil
}

// processPageToAI processes the page and sends it to the AI.
func processPageToAI(pageBuf bytes.Buffer, question string) (*http.Response, error) {
	// write http curl request to send to AI
	var requestBody bytes.Buffer
	writer := multipart.NewWriter(&requestBody)
	part, err := writer.CreateFormFile("files", "cv.pdf")
	if err != nil {
		return nil, err
	}
	_, err = io.Copy(part, &pageBuf)
	if err != nil {
		return nil, err
	}
	err = writer.WriteField("question", question)
	if err != nil {
		return nil, err
	}
	err = writer.Close()
	if err != nil {
		return nil, err
	}
	req, err := http.NewRequest("POST", "https://ai.unidoc.io/extract_text", &requestBody)
	if err != nil {
		return nil, err
	}
	req.Header.Set("Content-Type", writer.FormDataContentType())
	client := &http.Client{}
	return client.Do(req)
}

Page 1:
{
  "files": [
    {
      "cv_content": {
        "achievements": "2001 Became CEO of Wayne Enterprises; 2005 Established the Wayne Foundation; 2010 Launched Wayne Tech's R\u0026D division; 2015 Introduced Wayne Enterprises' renewable energy initiative; 2020 Awarded Gotham Humanitarian of the Year",
        "answer": "CEO, Wayne Enterprises (2010 - Present) - Spearheaded corporate expansions into renewable energy, biotech, and security technologies. Led R\u0026D in advanced transportation and defense technologies.",
        "certifications": "",
        "education": "Bachelor's in Business Administration \u0026 Criminology from Gotham University",
        "email": "[email protected]",
        "experience": "CEO, Wayne Enterprises (2010 - Present) - Spearheaded corporate expansions into renewable energy, biotech, and security technologies. Led R\u0026D in advanced transportation and defense technologies. Founder, Wayne Foundation (2005 - 2010) - Funded social welfare programs focusing on education, healthcare, and crime prevention. Developed the Gotham Outreach Program.",
        "full_name": "Bruce Wayne",
        "languages": [],
        "phone": "+1 555-1940",
        "skills": [],
        "summary": "Visionary businessman, philanthropist, and CEO of Wayne Enterprises with expertise in corporate management, technological innovation, and security solutions.",
        "type": "CV"
      },
      "filename": "cv.pdf",
      "type": "CV"
    }
  ]
}
Page 2:
{
  "files": [
    {
      "cv_content": {
        "achievements": "1992 Graduated from MIT at age 17, 1999 Revolutionized the defense industry, 2008 Created first Arc Reactor-powered exosuit, 2010 Transitioned Stark Industries to clean energy, 2015 Developed AI-driven robotics innovations",
        "answer": "Founder, Stark Industries (2011 - Present) - Led groundbreaking research in clean energy, AI, and defense. Developed advanced exosuits and military-grade technology.",
        "certifications": "",
        "education": "Bachelor's in Electrical Engineering \u0026 Physics from MIT",
        "email": "[email protected]",
        "experience": "Founder, Stark Industries (2011 - Present) - Led groundbreaking research in clean energy, AI, and defense. Developed advanced exosuits and military-grade technology.",
        "full_name": "Tony Stark",
        "languages": [],
        "phone": "+1 555-3000",
        "skills": [],
        "summary": "Genius entrepreneur, inventor, and former CEO of Stark Industries. Innovator in robotics, AI, and defense technology with expertise in high-tech engineering and business strategy.",
        "type": "CV"
      },
      "filename": "cv.pdf",
      "type": "CV"
    }
  ]
}

1. Bruce Wayne;What is the latest experience?;CEO, Wayne Enterprises (2010 - Present) - Spearheaded corporate expansions into renewable energy, biotech, and security technologies. Led R&D in advanced transportation and defense technologies.
2. Tony Stark;What is the latest experience?;Founder, Stark Industries (2011 - Present) - Led groundbreaking research in clean energy, AI, and defense. Developed advanced exosuits and military-grade technology.

go run main.go <your_cv_pdf_file.pdf>