feat: make pdf2john cli executable

README.md

@@ -22,27 +22,22 @@ Create a virtual environment and install Python dependencies
 ```bash
 pyenv virtualenv 3.11.4 pdf2john
 pyenv shell pdf2john
-poetry install
-```
-
-or if you don't want to use a virtual environment, simply
-```bash
-pip install -r requirements.txt
+poetry install --only main
 ```
 
 ## Usage
-Run pdf2john.py with a PDF of choice
-```bash
-./pdf2john.py example.pdf
-```
-
 To pass the hash to john:
 ```bash
-./pdf2john.py example.pdf >> .hash
+pdf2john example.pdf >> .hash
 john .hash
 john --show --format=PDF .hash
 ```
 
+If you don't want to install pdf2john, you can run it as a standalone script:
+```bash
+./src/pdf2john/pdf2john.py example.pdf >> .hash
+```
+
 ## Features
 - Responsibility for PDF parsing and handling has been delegated to [pyHanko](https://github.com/MatthiasValvekens/pyHanko) (a crytography focused fork of PyPDF2)
 - CICD workflow that tests pdf2john against PDFs ranging from Security Handler Revision 2 -> 6

pyproject.toml

@@ -31,3 +31,6 @@ profile = "black"
 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
+
+[tool.poetry.scripts]
+pdf2john = "pdf2john.pdf2john:main"

src/pdf2john/__init__.py

@@ -0,0 +1,3 @@
+from .pdf2john import PdfHashExtractor
+
+__all__ = ["PdfHashExtractor"]

pdf2john.py → src/pdf2john/pdf2john.py

@@ -111,7 +111,7 @@ def get_passwords(self) -> str:
         return "*".join(passwords)
 
 
-if __name__ == "__main__":
+def main() -> None:
     parser = argparse.ArgumentParser(description="PDF Hash Extractor")
     parser.add_argument(
         "pdf_files", nargs="+", help="PDF file(s) to extract information from"
@@ -137,3 +137,7 @@ def get_passwords(self) -> str:
 
         except PdfReadError as error:
             logger.error("%s : %s", filename, error, exc_info=True)
+
+
+if __name__ == "__main__":
+    main()

tests/test_cli.py

@@ -1,9 +1,11 @@
 import subprocess
 
+from pytest import mark
 
-def run_cli(args, input_text=None):
+
+def run_cli(args: list, cmd: str = "pdf2john", input_text=None):
     process = subprocess.Popen(
-        ["python3", "pdf2john.py"] + args,
+        [cmd] + args,
         stdout=subprocess.PIPE,
         stderr=subprocess.PIPE,
         stdin=subprocess.PIPE,
@@ -13,9 +15,10 @@ def run_cli(args, input_text=None):
     return process, stdout, stderr
 
 
-def test_cli_extraction():
+@mark.parametrize("cmd", ["pdf2john", "./src/pdf2john/pdf2john.py"])
+def test_cli_extraction(cmd):
     pdf_file = "tests/pdf/r6-test-bad-password.pdf"
-    process, stdout, _ = run_cli([pdf_file])
+    process, stdout, _ = run_cli(cmd=cmd, args=[pdf_file])
 
     expected = "$pdf$5*6*256*-4*1*16*fce2fe96b7e142b4a0576f61e2e9c441*48*aef6c4bf5e8a0f3bb1adef2b8ac2367d1ce95ecc1ddc3243ce49786086a023aa310aa9f7d1d103f837e4d4f738ac913d*48*eabf37f3f1f1b208f7c8ddfad3b817c689889ecbadd30f4581382cfbf79806304fb438e9ca227a023138a38eadcf82f3*32*37afcbfcbb32d4e1bca1eb10165693a1633ebb742c00045177a284ba22196937*32*3459a644d5f4c4f7cee562b754b30df48d598e1911ea513ef29bb3928593caf3"
     assert process.returncode == 0