暫定版陸自イラク日報テキスト付PDF生成スクリプト

create_pdf_with_text.py

# -*- coding: utf-8 -*-
from reportlab.pdfgen import canvas
from reportlab.pdfbase import pdfmetrics
# from reportlab.rl_config import defaultPageSize
from reportlab.lib.pagesizes import A4, landscape
from reportlab.pdfbase.cidfonts import UnicodeCIDFont
from reportlab.lib.colors import red,black
from reportlab.lib.units import cm

from pdf2image import convert_from_path, convert_from_bytes

import json
from pathlib import Path

import tempfile
import natsort
import math

# [PDF生成 (ReportLab) | Python-izm](https://www.python-izm.com/third_party/pdf/repotlab/)
# [PythonでKindle用自炊PDFを作ろうと実験 - Qiita](https://qiita.com/Gen6/items/5ace54481d66b653fa15)


def create_pdf(pdffile, output, jsonlist ):

    images_from_path = []


    print("processing pdf: {0}".format(pdffile))


    text_pages = []

    for i, path in enumerate(jsonlist) :
        raw_data = json.load(path.open())

        text_pages.append(raw_data['textAnnotations'])


    print("json data loaded.")

    with tempfile.TemporaryDirectory() as path:
        images_from_path = convert_from_path(pdffile, output_folder=path)



    print(len(images_from_path))


    newPdfPage = canvas.Canvas(output)
    newPdfPage.setPageSize(A4)

    newPdfPage.saveState() # 必要？

    newPdfPage.setAuthor('Provide by Information disclosure')
    newPdfPage.setTitle('陸上自衛隊イラク日報PDF')
    newPdfPage.setSubject('')

    pdfmetrics.registerFont(UnicodeCIDFont('HeiseiKakuGo-W5'))
    # pdfmetrics.registerFont(UnicodeCIDFont('HeiseiMin-W3'))
    # pdfmetrics.registerFont(UnicodeCIDFont('KozMinPro-Regular'))

    for i, image in enumerate(images_from_path):
        print(f"page: {i}")

        print(image.size)
        image_width , image_height = image.size

        ratio = image_width / image_height

        landscape_mode = False
        page_size = {}


        if ratio > 1.0 :
            landscape_mode = True
            newPdfPage.setPageSize(landscape(A4))
            page_size['width'], page_size['height'] = landscape(A4)
        else:
            newPdfPage.setPageSize(A4)
            page_size['width'], page_size['height'] = A4




        with tempfile.NamedTemporaryFile(mode='w+b',suffix='jpg') as fp:

            image.save(fp.name,'JPEG')


            newPdfPage.drawImage(fp.name,0,0,width=page_size['width'], height=page_size['height'], preserveAspectRatio=True)

        newPdfPage.setFont('HeiseiKakuGo-W5', 10)

        newPdfPage.setFillColor(red, alpha=0.7)

        print(len(text_pages))
        for j, textblob in enumerate(text_pages[i]) :

            # 最初のデータはページ全体なのでスキップ
            if j == 0 :
                continue


            #for node in textblob:
            node = textblob

            #print(node)
            #print(node['boundingPoly'])
            #print(node['boundingPoly']['vertices'])



            anchor_y = image_height - int(node['boundingPoly']['vertices'][3]['y'])
            anchor_x =  int(node['boundingPoly']['vertices'][3]['x'])

            scale = 1.0


            if landscape_mode :
                scale = min(page_size['height'] / image_height, page_size['width'] / image_width)

            else:
                scale = min(page_size['height'] / image_height, page_size['width'] / image_width)


            text_height = int(node['boundingPoly']['vertices'][3]['y']) -  int(node['boundingPoly']['vertices'][0]['y'])

            fontsize = math.ceil(text_height / (200 / 72 ) )

            newPdfPage.setFont('HeiseiKakuGo-W5', fontsize)

            newPdfPage.drawString(scale * anchor_x, scale * anchor_y, node['description'])

        # newPdfPage.drawString(250, 650, 'これはテストです') #　Y座標は左下原点

        newPdfPage.showPage()

    newPdfPage.save()

if __name__ == "__main__" :

    # ToDo ファイル名の指定 or 一括処理

    file = "pdf/イラク後送業務隊/060720.pdf"

    json_file_st = "./json/イラク後送業務隊/060720_*.json"

    p = Path('.')

    unsort  = list(p.glob(json_file_st))

    json_path_list = natsort.natsorted(unsort)


    create_pdf(file, output="test_out.pdf",jsonlist=json_path_list )