Reading a PDF file from a S3 bucket location?

[feelosophy13]

I need to read PDF files that are stored in a S3 bucket. I tried the following approaches, but none of them worked. Does anyone know how I can read PDFs off of a S3 bucket?

######################################################################
## attempt 1: supplying S3 URL directly to pdfplumber.open() method ##
######################################################################

import pdfplumber

s3_url = "s3://example-s3-bucket/example.pdf"

with pdfplumber.open(s3) as pdf:
  pages = pdf.pages

## ERROR -- AttributeError: 'S3' object has no attribute 'seek'




#####################################################
## attempt 2: opening content using pre-signed URL ##
#####################################################

import pdfplumber
import boto3

s3 = boto3.client('s3')

pdf_url = s3.generate_presigned_url(
    'get_object',
    Params={'Bucket': 'example-s3-bucket', 'Key': 'example.pdf'},
    ExpiresIn=3600  # Set an expiration time for the URL if needed
)

with pdfplumber.open(pdf_url) as pdf:
  pages = pdf.pages

## ERROR -- OSError: [Errno 63] File name too long:




############################################################
## attempt 3: supplying bytes to pdfplumber.open() method ##
############################################################

import pdfplumber
import boto3

bucket_name = "example-s3-bucket"
file_key = "example.pdf"
response = s3.get_object(Bucket=bucket_name, Key=file_key)
file_bytes = response['Body'].read()

with pdfplumber.open(file_bytes) as pdf:
  pages = pdf.pages

## ERROR -- AttributeError: 'bytes' object has no attribute 'seek'

[jsvine]

Hi @feelosophy13; pdfplumber will not load PDFs directly from a URL, but the bytes approach should work. What version of pdfplumber are you using?

If, for some reason, you're on the latest version and the bytes approach still isn't working, you can try:

import io
pdfplumber.open(io.BytesIO(my_bytes))

[feelosophy13]

Hi @jsvine. Thank you! Your solution worked!

I'm on pdfplumber version 0.10.3 and Python 3.9.0. With my original bytes approach, I got the following error: AttributeError: 'bytes' object has no attribute 'seek'