Generate PDF using reportlab and write to s3 bucket without saving locally

[Danny Blaker]

I'm trying to generate a PDF using reportlab and write to s3 bucket without saving locally

I know I'm missing something simple:

# create a stream

stream = io.BytesIO()

# generate PDF

doc = SimpleDocTemplate(stream, pagesize=letter,

                        rightMargin=72, leftMargin=72,

                        topMargin=72, bottomMargin=18)

Story = []

styles = getSampleStyleSheet()

Story.append(Paragraph('<font size=11>This is a PDF</font>', styles["Normal"]))

doc.build(Story)

# get buffer

pdf_buffer = stream.getbuffer()

filename = "new.pdf"

bucket_name = 'insert_bucket_name'

object_name = bucket_name

# here is where I get stuck - how should be passing the pdf_buffer to s3?

# how you typically write to s3 :

# Method 1

s3 = boto3.client('s3')

with open(filename, "rb") as f:

    s3.upload_fileobj(f, bucket_name, object_name)

# Method 2

s3.Bucket(bucket_name).put_object(Key=filename, Body=file)

Any ideas most appreciated!

Thanks!

here are some related resources:

https://stackoverflow.com/questions/43373006/django-reportlab-save-generated-pdf-directly-to-filefield-in-aws-s3

https://sciwiki.fredhutch.org/compdemos/aws-python/#about-pandas-and-dask

https://stackoverflow.com/questions/12570465/how-to-upload-a-file-to-s3-without-creating-a-temporary-local-file

[dzu...@gmail.com]

Figured it out: streams have this builtin position attr which bytesIO uses to keep track of to know where to insert data. Reportlab simpledoctemplate writes to the stream object, but when it completes pdf generation at doc.build(story), it doesn't reset the stream position back to the beginning of the file. as a result the s3 put_object method writes only the portion of the file from where the position is at currently, through to the end, resulting in corrupt file. all that was neeeded was to reset the stream position just before writing to s3 using stream.seek(0)