Python library to parse the pcap-ng format used by newer versions of dumpcap & similar tools (wireshark, winpcap, ...).

If you prefer the RTD theme, or want documentation for any version other than the latest, head here:

http://python-pcapng.readthedocs.org/en/latest/

If you prefer the more comfortable, page-wide, default sphinx theme, a documentation mirror is hosted on GitHub pages:

http://rshk.github.io/python-pcapng/

Branch	Status
master
develop

Source, issue tracker etc. on GitHub: https://github.com/rshk/python-pcapng

Get the source from git:

git clone https://github.com/rshk/python-pcapng

Download zip of the latest version:

https://github.com/rshk/python-pcapng/archive/master.zip

Install from pypi:

pip install python-pcapng

The official page on the Python Package Index is: https://pypi.python.org/pypi/python-pcapng

• I need to decently extract some information from a bunch of pcap-ng files, but apparently tcpdump has some problems reading those files,

  I couldn't find other nice tools nor Python bindings to a library able to parse this format, so..

• In general, it appears there are (quite a bunch of!) Python modules to parse the old (much simpler) format, but nothing for the new one.
• And, they usually completely lack any form of documentation.

Yes, I guess it would be much slower than something written in C, but I'm much better at Python than C.

..and I need to get things done, and CPU time is not that expensive :)

(Maybe I'll give a try porting the thing to Cython to speed it up, but anyways, pure-Python libraries are always useful, eg. for PyPy).

Basic usage is as simple as:

from pcapng import FileScanner

with open('/tmp/mycapture.pcap', 'rb') as fp:
    scanner = FileScanner(fp)
    for block in scanner:
        pass  # do something with the block...

Have a look at the blocks documentation to see what they do; also, the examples directory contains some example scripts using the library.

Format specification is here:

https://github.com/pcapng/pcapng/

Contributions are welcome, please contact me if you're planning to do some big change, so that we can sort out the best way to integrate it.

Or even better, open an issue so the whole world can participate in the discussion :)

Write support exists as of version 2.0.0. See the file examples/generate_pcapng.py for an example of the minimum code needed to generate a pcapng file.

In most cases, this library will prevent you from creating broken data. If you want to create marginal pcapng files, e.g. as test cases for other software, you can do that by adjusting the "strictness" of the library, as in:

from pcapng.strictness import Strictness, set_strictness
set_strictness(Strictness.FIX)

Recognized values are Strictness.FORBID (the default), Strictness.FIX (warn about problems, fix if possible), Strictness.WARN (warn only), and Strictness.NONE (no warnings). Circumstances that will result in strictness warnings include:

• Adding multiples of a non-repeatable option to a block
• Adding a SPB to a file with more than one interface
• Writing a PB (PBs are obsolete and not to be used in new files)
• Writing EPB/SPB/PB/ISB before writing any IDBs

1. Create a tag for the new version:

   git tag v2.0.0 -m 'Version 2.0.0'

2. Install build dependencies in a virtualenv:

   python -m venv ./.build-venv
   ./.build-venv/bin/python -m pip install build twine

3. Build source and wheel distributions:

   rm -rf ./dist *.egg-info
   ./.build-venv/bin/python -m build

4. Use Twine to upload to pypi:

   twine upload dist/*

If you get some crazy version number like 2.0.1.dev0+g7bd8575.d20220310 instead of what you expect (eg 2.0.0), it's because you have uncommitted or untracked files in your local working copy, or you created more commits after creating the tag. Such a version number will be refused by pypi (and it's not a good version number anyways), so make sure you have a clean working copy before building.