Exceptions, Warnings, and Log messages
pypdf makes use of 3 mechanisms to show that something went wrong:
Log messages are informative messages that can be used for post-mortem analysis. Most of the time, users can ignore them. They come in different levels, such as info / warning / error indicating the severity. Examples are non-standard compliant PDF files which pypdf can deal with.
Warnings are avoidable issues, such as using deprecated classes / functions / parameters. Another example is missing capabilities of pypdf. In those cases, pypdf users should adjust their code. Warnings are issued by the
warningsmodule - those are different from the log-level “warning”.
Exceptions are error-cases that pypdf users should explicitly handle. In the
strict=Truemode, most log messages with the warning level will become exceptions. This can be useful in applications where you can force to user to fix the broken PDF.
Exeptions need to be catched if you want to handle them. For example, you could want to read the text from a PDF as a part of a search function.
Most PDF files don’t follow the specifications. In this case pypdf needs to guess which kinds of mistakes were potentially done when the PDF file was created. See the robustness page for the related issues.
As a users, you likely don’t care about it. If it’s readable in any way, you want the text. You might use pdfminer.six as a fallback and do this:
from pypdf import PdfReader from pdfminer.high_level import extract_text as fallback_text_extraction text = "" try: reader = PdfReader("example.pdf") for page in reader.pages: text += page.extract_text() except Exception as exc: text = fallback_text_extraction("example.pdf")
You could also capture
if you prefer something more specific.
warnings module allows
you to ignore warnings:
import warnings warnings.filterwarnings("ignore")
In many cases, you actually want to start Python with the
-W flag so that you
see all warnings. This is especially true for Continuous Integration (CI).
Log messages can be noisy in some cases. pypdf hopefully is having a reasonable level of log messages, but you can reduce which types of messages you want to see:
import logging logger = logging.getLogger("pypdf") logger.setLevel(logging.ERROR)
defines six log levels: