[Python] Empty PDF attachment using fpdf2 with boto3 sesv2

I'm facing an issue with empty PDF attachments sent with boto3 SESv2 client using Templated Email type.

I think this is something related to encoding or specific boto3 send_email rules that I'm not following, but IDK which exactly. Saving file locally works totally correct.

docs I'm using: https://boto3.amazonaws.com/v1/docu...ference/services/sesv2/client/send_email.html

https://py-pdf.github.io/fpdf2/index.html

class AttachmentDict(TypedDict):
RawContent: bytes
ContentDisposition: Literal["ATTACHMENT", "INLINE"]
FileName: str
ContentId: str
ContentType: Literal["application/pdf", "image/jpeg"]


def send_templated_email(email_request: EmailTemplateRequest) -> str:
"""
Sends an email based on a template. A template contains replaceable tags
each enclosed in two curly braces, such as {{name}}. The template data passed
in this function contains key-value pairs that define the values to insert
in place of the template tags.

aram email_request: data that contains necessary details for email
:return: The ID of the message, assigned by Amazon SES.
"""
send_args: TemplatedEmailArgs = {
"FromEmailAddress": email_request.source,
"Destination": {"ToAddresses": email_request.recipients},
"Content": {
"Template": {
"TemplateName": email_request.template_name,
"TemplateData": email_request.template_data.model_dump_json(),
"Attachments": [],
},
},
}

# Add attachments if any
for attachment in email_request.attachments:
attachment_arg: AttachmentDict = {
"RawContent": attachment.content,
"ContentDisposition": "ATTACHMENT",
"FileName": attachment.filename,
"ContentType": "application/pdf",
}
send_args["Content"]["Template"]["Attachments"].append(attachment_arg)

response = email_client.send_email(**send_args)



The content for attachment is assembled for fpdf2 library like this

So on the line "RawContent": attachment.content attachment.content is file_content (bytes type)

pdf_file = fpdf.FPDF(orientation="P", unit="mm", format="A4")
pdf_file.add_page()
pdf_file.set_auto_page_break(auto=True, margin=15)
# ... Generation of content
file_content = pdf_file.output() # This file content goes to "RawContent" as bytes



But in the end I see PDF attachment with an empty body, but correct page number.

Continue reading...