lgcosta/paperless-ngx
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: Update page size

Browse Source
This commit is contained in:
otxtan@gmail.com 2024-05-29 10:05:15 +07:00
parent 74af6d8853
commit 8a4f809d83
1 changed files with 8 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
src/paperless_ocr_custom/parsers.py
View File

@ -258,11 +258,14 @@ class RasterisedDocumentParser(DocumentParser):
first_page=1, first_page=1,
last_page=input_pdf.getNumPages()+1) last_page=input_pdf.getNumPages()+1)
can = canvas.Canvas(str(output_path), pagesize=letter) can = canvas.Canvas(str(output_path), pagesize=letter)
for page_num, image in enumerate(images): for page_num, page in enumerate(input_pdf.pages):
page_width, page_height = image.size page_height = page.mediabox.getHeight()
# page_height = input_pdf.pages[page_num].mediabox[3] page_width = page.mediabox.getWidth()
# page_width = input_pdf.pages[page_num].mediabox[2] width_api_img = data["pages"][page_num]["dimensions"][1]
height_api_img = data["pages"][page_num]["dimensions"][0]
# set size new page # set size new page
if width_api_img < height_api_img and page_height < page_width:
page_height,page_width = page_width,page_height
can.setPageSize((page_width, page_height)) can.setPageSize((page_width, page_height))
byte_image = io.BytesIO() byte_image = io.BytesIO()
images[page_num].save(byte_image, format='JPEG') images[page_num].save(byte_image, format='JPEG')
@ -273,8 +276,6 @@ class RasterisedDocumentParser(DocumentParser):
# height=float(page_height)) # height=float(page_height))
# set font size # set font size
pdfmetrics.registerFont(TTFont('Arial', font_path)) pdfmetrics.registerFont(TTFont('Arial', font_path))
width_api_img = data["pages"][page_num]["dimensions"][1]
height_api_img = data["pages"][page_num]["dimensions"][0]
# print(f'kich thuoc goc: height{page_height}, width{page_width}, kich thuoc api: height{height_api_img} width{width_api_img}') # print(f'kich thuoc goc: height{page_height}, width{page_width}, kich thuoc api: height{height_api_img} width{width_api_img}')
rolate_height = height_api_img /page_height rolate_height = height_api_img /page_height
rolate_width = width_api_img /page_width rolate_width = width_api_img /page_width
@ -282,7 +283,7 @@ class RasterisedDocumentParser(DocumentParser):
for line in block.get("lines", []): for line in block.get("lines", []):
y1 = (line.get("bbox")[0][1] / float(rolate_height)) y1 = (line.get("bbox")[0][1] / float(rolate_height))
y2 = (line.get("bbox")[1][1] / float(rolate_height)) y2 = (line.get("bbox")[1][1] / float(rolate_height))
font_size = math.floor((y2 - y1) * 72 / 96)-2 font_size = math.floor((y2 - y1) * 72 / 96)-1
y_center_coordinates = y2 - (y2 - y1)/2 y_center_coordinates = y2 - (y2 - y1)/2
for word in line.get("words", []): for word in line.get("words", []):
x1 = word["bbox"][0][0] / float(rolate_width) x1 = word["bbox"][0][0] / float(rolate_width)