PDF Font Autosize Not Working On Firefox

[kyl293]

What happened? What were you trying to achieve?
PDF's that are generated do not auto size font for inputs on Firefox but look normal on Microsoft Edge and Google Chrome. If I then save it using acrobat it fixes the font size issue. If there is a way (code wise) to fix the scaling issue?

Environment

Which environment were you using when you encountered the problem?
I am running the software on AWS Lambda (Put pypdf in a layer so that Lambda can access it)

$ python -m platform
{
  "statusCode": 200,
  "body": {
    "file_name": "1182_insert_id_card.pdf"
  }
}
# TODO: Your output goes here

Code + PDF

This is a minimal, complete example that shows the issue:

import os
import json
import boto3
from pypdf import PdfReader, PdfWriter
from io import BytesIO

def generate_pdf(event, context):
    # Extract input data from the Lambda event
    #print(json.dumps(event))
 # Retrieve the raw JSON payload from the request body
    raw_body = event.get("body")
    input_data = raw_body
    #print("A")
    # Check if the body is present
    """
    if raw_body is not None:
        
        # Parse the JSON from the request body
        try:
           input_data = json.loads(raw_body)
        except json.JSONDecodeError as e:
            return {
                'statusCode': 400,
                'body': json.dumps({'error': 'Invalid JSON payload'}),
            }
    else:
        return {
            'statusCode': 400,
            'body': json.dumps({'error': 'Request body is missing'}),
        }
"""
    # Now input_data should contain the parsed JSON payload
    #print(f"Received Input Data: {input_data}")

    # Path to the template PDF file (assuming it's included in the Lambda deployment package)
    #print("B")
    template_path = getTemplate(input_data)
    #print(f"{template_path=}")
    #print("C")
    # Fill in the form fields in the template PDF
    #print("d")
    filled_pdf_buffer = fill_pdf_template(template_path, input_data)
    #print("e")
    # Store the filled PDF in the /tmp directory

    # Upload the filled PDF to the destination S3 bucket
    #print('f')
    destination_bucket = input_data.get('dest_bucket')
    output_key = get_output_key(input_data)
    #print('g')
    upload_to_s3(filled_pdf_buffer, destination_bucket, output_key)
    #print('h')
    #print("b4 return")

    return {
            'statusCode': 200,
            'body': {'file_name': f'{filled_pdf_buffer.split("/")[-1]}'}
        }
        
        
def fill_pdf_template(template_path, input_data):

    reader = PdfReader(template_path,strict=True)
    metadata=reader.metadata
    writer = PdfWriter()
    writer.clone_reader_document_root(reader)
    writer.add_metadata(metadata)    
    pagecount=0
    writer.set_need_appearances_writer(False)
   
    
    writer.append(reader)
    
    for page in range(len(reader.pages)):
        fields = reader.get_form_text_fields()
        print(fields)
        if fields is not None:
            writer.update_page_form_field_values(
                writer.pages[page],
                input_data.get('input_data'),
                flags=1,
                auto_regenerate=True,
           )

    #print("run" ,page)
    # write "output" to pypdf-output.pdf
    tmp_filename=f'/tmp/{input_data.get("uniqueID", "Unknown")}_{input_data.get("file_name", "Unknown")}'
    with open(tmp_filename, "wb") as output_stream:
        writer.write(output_stream)
 
    return tmp_filename
    
    
def get_output_key(input_data):
    # Customize how you want to generate the output key based on input data
    # This is just a simple example, adjust it as needed
    return f'{input_data.get("uniqueID", "Unknown")}_{input_data.get("file_name", "Unknown")}'

def upload_to_s3(tmp_filename,dest_bucket,output_key):
    s3 = boto3.client('s3')
    s3.upload_file(tmp_filename,dest_bucket,output_key)


def getTemplate(input_data):
    s3= boto3.client('s3')
    bucket=input_data.get("bucket_name")
    print(f"{bucket=}")
    path=input_data.get("file_path")
    print(f"{path=}")
    name='/tmp/'+input_data.get("file_name")
    print(f"{name=}")
    try:
        print("yes")
        s3.download_file(bucket,path,name)
        return name
    except  Exception as error:
        print("error during getTemplate")
        print(str(error))
        return {
            'statusCode': 400,
            'body': {'error': error},
        }

Share here the PDF file(s) that cause the issue. The smaller they are, the
better. Let us know if we may add them to our tests!
Card with filled dummy data
1182_insert_id_card.pdf
Card with unfilled data
insert_id_card (1).pdf

[pubpub-zz]

I suspect the way you generate writer.
Can you try to replace your fill_pdf_template function:

def fill_pdf_template(template_path, input_data):

    reader = PdfReader(template_path,strict=True)
    writer = PdfWriter(clone_from=reader)
    pagecount=0
    for page in range(len(reader.pages)):
        fields = reader.get_form_text_fields()
        print(fields)
        if fields is not None:
            writer.update_page_form_field_values(
                writer.pages[page],
                input_data.get('input_data'),
                flags=1,
                auto_regenerate=True,
           )
    tmp_filename=f'/tmp/{input_data.get("uniqueID", "Unknown")}_{input_data.get("file_name", "Unknown")}'
    writer.write(tmp_filename)
 
    return tmp_filename

[kyl293]

I just tried your code and it did not fix the issue unfortunately. The policy number is still not complete on firefox

[pubpub-zz]

this is my test code:

from pypdf import PdfReader, PdfWriter

reader = PdfReader("insert_id_card.1.pdf",strict=True)
writer = PdfWriter(clone_from=reader)

f = reader.get_form_text_fields()
for k in f:
    f[k]="!!!"+k+"!!!"

writer.update_page_form_field_values(writer.pages[0],f,1,True)

writer.write("tt.pdf")

and the result:

[kyl293]

this is my test code:

from pypdf import PdfReader, PdfWriter

reader = PdfReader("insert_id_card.1.pdf",strict=True)
writer = PdfWriter(clone_from=reader)

f = reader.get_form_text_fields()
for k in f:
    f[k]="!!!"+k+"!!!"

writer.update_page_form_field_values(writer.pages[0],f,1,True)

writer.write("tt.pdf")

and the result:

Try making the Policy Number absurdly long. When that is done the Font size stays the same on Firefox. This happens with every PDF I generate. So for other forms that have tighter boxes the cut off is much smaller.

Here is what it looks like on chrome using your code snippet:

And here is what the PDF looks like when I open it in Firefox:

Policy Number get's cut off.

[pubpub-zz]

The points deals with the font size.
in the DA field (default appearance initialisation) you can see that the size (number before Tf) is set to 0 : this means the font size is dynamic. if you look at the pdf1.7 reference page 435 you will read:

From my interpretation Firefox is in accordance with the standard.

Further, pypdf does not have yet a capability to change the height of the font (see #2064 still in progress).
meanwhile you should be able to compute manually the proper font size and substitute the number before Tf in the annotation content.
Once done you should not ask the appearance to regenerated, using auto_regenerate=False

[pubpub-zz]

@kyl293
any update ?

[pubpub-zz]

I close this issue as solved. Feel free to send update to reopen-it