Uncaught exception raised when assembling requirements

[RobbinDG]

Hey!

I've taken an interest in this tool because I think it could be useful for quickly and efficiently deploying cloud functions in a python mono-repo. I encountered the following uncaught exception when attempting to extract a simple

Traceback (most recent call last):
  File "venv/bin/python-object-extractor", line 8, in <module>
    sys.exit(main())
  File "[...]/venv/lib/python3.8/site-packages/python_object_extractor/main.py", line 115, in main
    output(
  File "[...]/venv/lib/python3.8/site-packages/python_object_extractor/output.py", line 32, in output
    output_requirements(sys.stdout, imports)
  File "[...]/venv/lib/python3.8/site-packages/python_object_extractor/output.py", line 89, in output_requirements
    requirements = {
TypeError: 'NoneType' object is not iterable

Steps to reproduce

1. Setup:
   OS: arch linux
   Python version: 3.8 (venv)
   Object extractor version: 1.0.0

2. use the CLI to use the object extractor on a module with a single function and no import statements.

3. Output: The contents of the file written to STDOUT, followed by the exception above.

Expected output
A proper error message indicating what the issue is and optionally how to resolve it, if this is correct behaviour.

Expected cause
I looked into the output.py file briefly, and it appears that the .third_party attribute is None. I did not find the origins of this attribute and what the contents are supposed to represent, but I assume that in my case, an empty list or similar is expected (as I have no imports, third party or otherwise).

Let me know if I can be of any help solving this, I'm willing to contribute to the repo if necessary.

[o3bvv]

Hello! Thank you for your detailed feedback! It is a pleasure to find this tool useful.

I admit, this repo contains a setup.py with an outdated implementation of parse_requirements() function, which can cause issues, sorry for that.

I am planning to fix this in the near future. Meanwhile, I place an updated version of the function for those who might want to fix the issue locally:

from pathlib import Path


def parse_requirements(file_path: Path) -> tuple[list[str], list[str]]:
  requirements = list()
  dependencies = list()

  if not file_path.exists():
    return requirements, dependencies

  with file_path.open("rt") as f:
    for line in f:
      line = line.strip()

      # check if comment or empty
      if not line or line.startswith("#"):
        continue

      # check if is an inclusion of other requirements file
      elif line.startswith("-r"):
        name = Path(line.split(" ", 1)[1])
        path = file_path.parent / name
        subrequirements, subdependencies = parse_requirements(path)
        requirements.extend(subrequirements)
        dependencies.extend(subdependencies)

      # check if is a link, e.g. to a private repository
      elif "://" in line:
        if line.startswith("-e"):
          line = line[3:].lstrip()

        link, package_name = line.split("#egg=", maxsplit=1)

        requirements.append(f"{package_name} @ {link}")
        dependencies.append(line)

      # assume standard requirement
      else:
        requirements.append(line)

  return requirements, dependencies