zackees / zcmds

Swiss army knife of insanely productive CLI and AI tools. Cross platform.

zcmds

Cross platform(ish) productivity commands written in python. Tools for doing media manipulation through ffmpeg and AI. On Windows ls, rm and other common unix file commands are installed. Whenever there is something that doesn't work on Windows but does on Mac/Linux, I will apply a tool to make it work here. This toolset is ever-evolving and it's going to get insane in 2024 with all the AI that I'm now integrating.

Install

> pip install zcmds
> zcmds  # shows all commands
> diskaudit  # audits the disk usage from the current directory.

Commands

Remember that typing in zcmds at the terminal will show you all the commands

• archive
  • Zips up the specified directory or file.
• askai
  • Asks a question to OpenAI from the terminal command. Requires an openai token which will be requested and saved on first use.
  • Prefix your query with ! to run command directly.
• aicode
  • A front end for Aider, an AI pair programming tool. This is the future the sci fi writers promised you.
• audnorm
  • Normalizes audio in a media file to a standard volume.
• codeup
  • If your current git repo has ./lint, ./test, then this tool will run them in this order. If they both pass then git add . followed by git commit -m or aicommits will be invoked.
• comports
  • Shows all the ports that are in use at the current computer (useful for Arduino debugging).
• diskaudit
  • walks the directory from the current directory and catalogs which of the child folders take up the most space.
• git-bash (win32)
  • launches git-bash terminal (windows only).
• gitconfig
  • Configures git so that it's in "easy-to-use-mode".
• gitsummary
  • Generates a summary of the git repository commits, useful for invoicing
• findfiles
  • finds a file with the given glob.
• img2webp
  • Conversion tool for converting images into webp format.
• img2vid
  • Converts a series of images to a video.
• obs_organize
  • organizes the files in your default obs directory.
• merge-to
  • Merges a clean git repo (no untracked files) to the target branch, pushes that target branch, then switches back to the original branch.
• new
  • Opens a new terminal command window from the current terminal command window.
• printenv
  • prints the current environment variables, including path. Everything is sorted
• pdf2png
  • Converts a pdf to a series of images
• pdf2txt
  • Converts a pdf to a text file.
• push
  • A safer way to git push, checks if the rebase is dirty.
• removbackground
  • Launches an AI tool in the browser to remove the background from an Image. Can also generate video with background removed. Front end for rembg backend.
• search_and_replace
  • Search all the files from the current directory and applies exact text search and replace.
• search_in_files
  • Search all files from current working directory for exact string matches matches.
• sharedir
  • takes the current folder and shares it via a reverse proxy using ngrok.
• stereo2mono
  • Reduces a stereo audio / video to a single mono track.
• sudo (win32 only)
  • Runs a command as in sudo, using the gsudo tool.
• vidcat
  • Concatenates two videos together, upscaling a lower resolution video.
• vidmute
  • Strips out the audio in a video file and saves it as a new file.
• vidinfo
  • Uses ffprobe to find the information from a video file.
• vid2gif
  • A video is converted into an animated gif.
• vid2jpg
  • A video is converted to a series of jpegs.
• vid2mp3
  • A video is converted to an mp3.
• vid2mp4
  • A video is converted to mp4. Useful for obs which saves everything as mkv. Extremely fast with mkv -> mp4 converstion.
• vidclip
  • Clips a video using timestamps.
• viddur
  • Get's the during, use vidinfo instead.
• vidshrink
  • Shrinks a video. Useful for social media posts.
• vidspeed
  • Changes the speed of a video.
• vidvol
  • Changes the volume of a video.
• ytclip
  • Download and clip a video from a url from youtube, rumble, bitchute, twitter... The timestamps are prompted by this program.
• trash
  • Sends the folder or files to the trash. This sometimes works better than deleting files on Windows.
• whichall
  • Finds all the executables in the path.
• unzip
  • unzip the provided file
• fixinternet
  • Attempts to fix the internet connection by flushing the dns and resetting the network adapter.
• fixvmmem (win32 only)
  • Fixes the vmmem consuming 100% cpu on windows 10 after hibernate.
• transcribe-anything
  • Transcribe media content using state of the art insanely-fast-whisper
• tx
  • Easily send files over the internet. tx README.md
    • Front end to womrhole send file, but gives you the code upfront so the client can auto connect.

Install (dev):

• git clone https://github.com/zackees/zcmds
• cd zcmds
• python -pip install -e .
• Test by typing in zcmds

Additional install

For the pdf2image use:

• win32: choco install poppler
• ... ?

Note:

Running tox will install hooks into the .tox directory. Keep this in my if you are developing. TODO: Add a cleanup function to undo this.

Release Notes

• 1.4.87: aicode has been externalized into package advanced-aicode, but can still be invoked using aicode.
• 1.4.86: New tool git-diff
• 1.4.85: aicode now defaults to use openai/gpt-4o whenever possible, because it's that much better that claude3.
• 1.4.84: Fixes aicode issue where the saved version number could become corrupted with a KeyboardInterrupt
• 1.4.83: removebackground now uses -b for bitrate for mp4 like it does webm. Bitrates can now be specified in other units like 500k.code
• 1.4.82: removebackground now generates an mp4 encoded in HEVC for yuva420p support as well as webm with vp9 yuva420p.
• 1.4.81: aicode now defaults to --claude3 if anthropic key is set. removebackground now supports parallel processing.
• 1.4.80: aicode now supports --claude3, use --set-anthropic-key to set the key.
• 1.4.79: removebackground now allows --fps
• 1.4.78: removebackground now allows --height
• 1.4.77: removebackground now allows video/image as input.
• 1.4.76: askai now had --input-file, for better tooling.
• 1.4.75: Fix #13 in img2webp
• 1.4.74: Fix https://github.com/zackees/zcmds/issues/901
• 1.4.73: aicode now accepts Windows paths and converts them to posix paths prior to sending them to Aider.
• 1.4.72: aicode is now 2x faster to load because checking update version is now a delayed background task.
• 1.4.71: codeup now finds .git directory a few parents up, also allows --no-test
• 1.4.70: askai now has --assistant-prompt-file
• 1.4.69: codeup now implies --push. If you don't want to push then use `--no-push``
• 1.4.68: codeup now has --push to allow pushes to the repo if everything passes.
• 1.4.67: Adds new tool codeup which will run ./lint (if it exists) then ./test (if it exists) then aicommits (if it exists)
• 1.4.66: askai now allows optional --assistant-prompt to tell the AI what it is. This is super useful for data scientists, you can use this in your Jupyter Notebooks quite easily!
• 1.4.65: Adds gitconfig
• 1.4.64: Adds new to open a new terminal command window from the current terminal command window.
• 1.4.63: Adds gitconfigure to give sane defaults to your git.
• 1.4.62: askai can now run commands by prefixing with !
• 1.4.61: Fix bug in tx
• 1.4.60: New tool tx, a wrapper around wormhole send but easier to use.
• 1.4.59: New tool push, a safe way to git push
• 1.4.58: Fixes askai with positional args (asking a question and then immediatly exiting.)
• 1.4.57: Bring in new zcmds_win32 include sshpass
• 1.4.56: Fixes aicode on first run crash.
• 1.4.54: Bring in new zcmds_win32 fixes and improvements.
• 1.4.53: Fixes transcribe-anything with python 3.11 for --device insane
• 1.4.52: Update transcribe-anything for bug fix 2.7.23
• 1.4.51: Updates transcribe-anything to 2.7.22
• 1.4.50: Uses git-bash version of ssh for windows.
• 1.4.49: Adds trash which sends files to the trash.
• 1.4.48: Adds removebackground which uses AI to remove a background image. Uses rembg backend
• 1.4.47: Adds transcribe-anything to the command stack.
• 1.4.46: Fix merge-to with missing push step from target step.
• 1.4.45: Adds new tool merge-to, which streamlines merge a current branch into the other and then pushing.
• 1.4.44: Fixes vidwebmaster (Qt6 pinned version just stopped working!!)
• 1.4.43: Adds aicode which is the same as askai --code
• 1.4.42: Adds imgshrink
• 1.4.41: aider now installed with pipx to avoid package conflicts because of it's pinned deps.
• 1.4.40: Fix askai in python 3.11 with linux.
• 1.4.39: aider is now part of this command set. An awesome ai pair programmer. Enable it with askai --code
• 1.4.37: askai now streams output to the console.
• 1.4.36: losslesscut (on windows) can now be executed on other drivers and doesn't block the current terminal.
• 1.4.35: askai now assumed --fast. You can use gpt4 vs --slow
• 1.4.34: Fixes geninvoice
• 1.4.32: OpenAI now requires version 1.3.8 or higher (fixes breaking changes from OpenAI)
• 1.4.31: Improve audnorm so that it uses sox instead of ffmpeg-normalize. Fix bug where not all commands were installed. Fixes openai api changes.
• 1.4.30: Fix error in diskaudit when no files found in protected dir.
• 1.4.29: Fix img2webp.
• 1.4.28: Bug fix
• 1.4.27: askai now has --fast
• 1.4.26: vid2jpg now has --no-open-folder
• 1.4.24: Adds archive
• 1.4.23: Bump zcmds-win32
• 1.4.21: askai handles pasting text that has double lines in it.
• 1.4.20: askai is now at gpt-4
• 1.4.19: Adds losslesscut for win32.
• 1.4.18: Fix win32 zcmds_win32
• 1.4.17: vid2mp4 now adds --nvenc and --height --crf
• 1.4.16: Fixes img2webp.
• 1.4.15: Adds img2webp utility.
• 1.4.13: Add --no-fast-start to vidwebmaster.
• 1.4.12: Fixes a bug in find files when an exception is thrown during file inspection.
• 1.4.11: findfiles now has --start --end --larger-than --smaller-then
• 1.4.10: zcmds now uses get_cmds.py to get all of the commands from the exe list.
• 1.4.8: audnorm now encodes in mp3 format (improves compatibility). vid2mp3 now allows --normalize
• 1.4.7: Fixes broken build.
• 1.4.6: Adds say command to speak out the text you give the program
• 1.4.5: Adds saved settings for gitsummary
• 1.4.4: Adds pdf2txt command
• 1.4.3: Adds gitsummary command
• 1.4.2: Bump up zcmds_win32 to 1.0.17
• 1.4.1: Adds 'whichall' command
• 1.4.0: Askai now supports question-answer-question-... interactive mode
• 1.3.17: Adds syntax highlighting to open askai tool
• 1.3.16: Improves openai by using gpt 3.5
• 1.3.15: Improve vidinfo for more data and be a lot faster with single pass probing.
• 1.3.14: Improve vidinfo to handle non existant streams and bad files.
• 1.3.13: Added img2vid command.
• 1.3.12: Added fixinternet command.
• 1.3.11: Fix badges.
• 1.3.10: Suppress spurious warnings with chardet in openai
• 1.3.9: Changes sound driver, should eliminate the runtime dependency on win32.
• 1.3.8: Adds askai tool
• 1.3.7: findfile -> findfiles
• 1.3.6: zcmds[win32] is now at 1.0.2 (includes unzip)
• 1.3.5: zcmds[win32] is now at 1.0.1 (includes nano and pico)
• 1.3.4: Adds printenv utility
• 1.3.3: Adds findfile utility.
• 1.3.2: Adds comports to display all comports that are active on the computer.
• 1.3.1: Nit improvement in search_and_replace to improve ui
• 1.3.0: vidwebmaster now does variable rate encoding. --crf and --heights has been replaced by --encodings
• 1.2.1: Adds improvements to vidhero for audio fade and makes vidclip improves usability
• 1.2.0: stripaudio -> vidmute
• 1.1.30: Improves vidinfo with less spam on the console and allows passing height list
• 1.1.29: More improvements to vidinfo
• 1.1.28: vidinfo now has more encoding information
• 1.1.27: Fix issues with spaces in vidinfo
• 1.1.26: Adds vidinfo
• 1.1.26: Vidclip now supports start_time end_time being omitted.
• 1.1.25: Even better performance of diskaudit. 50% reduction in execution time.
• 1.1.24: Fixes diskaudit from double counting
• 1.1.23: Fixes test_net_connection
• 1.1.22: vid2mp4 - if file exists, try another name.
• 1.1.21: Adds --fps option to vidshrink utility
• 1.1.19: Using pyprojec.toml build system now.
• 1.1.17: vidwebmaster fixes heights argument for other code path
• 1.1.16: vidwebmaster fixes heights argument
• 1.1.15: vidwebmaster fixed
• 1.1.14: QT5 -> QT6
• 1.1.13: vidwebmaster fixes () bash-bug in linux
• 1.1.12: vidwebmaster now has a gui if no file is supplied
• 1.1.11: Adds vidlist
• 1.1.10: Adds vidhero
• 1.1.9: adds vidwebmaster
• 1.1.8: adds vidmatrix to test out different settings.
• 1.1.7: vidshrink and vidclip now both feature width argument
• 1.1.6: Adds touch to win32
• 1.1.5: Adds unzip to win32
• 1.1.4: Fix home cmd.
• 1.1.3: Fix up cmds so it returns int
• 1.1.2: Fix git-bash on win32
• 1.1.1: Release

TODO:

• Add silence remover:
  • https://github.com/bambax/Remsi
• Add lossless cut to vidclip
  • https://github.com/mifi/lossless-cut

zcmds's Issues

scale up not work using img2webp

execute
!img2webp 'maxresdefault.jpg' --format 'jpg' --scale 2

result
scale up not work (e.g. --scale 2 (double up) )
it only work when scale down (e.g. --scale 0.5)

here the prove (the image before execute img2webp is the original (maxresdefault.jpg)

ModuleNotFoundError: imgai

I started looking at the images and videos tools. It seems like there's an error related to a missing module called inputimeout within my Python environment. I ran pip install inputimeout. It is already installed.
Traceback (most recent call last):
File "", line 198, in _run_module_as_main
File "", line 88, in run_code
File "C:\Users\xx\AppData\Local\Programs\Python\Python312\Scripts\imgai.exe_main.py", line 4, in
File "C:\Users\xx\zcmds\src\zcmds\cmds\common\imgai.py", line 16, in
from .inputimeout import TimeoutOccurred, inputimeout
ModuleNotFoundError: No module named 'zcmds.cmds.common.inputimeout'

askai - Entering 'Exit' creates a loop

I was chatting with Askai (nice piece of tool) but I enter exit as I was leaving the chat (not sure how to end it yet) and the following loop forever... "Exited due to 'exit' command..... See screenshot.

I close abruptly my CMD window.

Accessibility for Newbies

I discovered your repo thru your interview with Mike Adams.

As you know writing code is (in the minds of most people) an esoteric dark science.
To make it more accessible to a wider group and to get newbies started . . .
and in the interests of reducing the significant barriers to having fun with repositories - is there any chance you could expand your notes to your readme.md about 1) how to install and 2) how to use one or two of the tools.

INSTALLATION IN WINDOWS
Chose a location on your hard drive to install the repository.
It does not have to be on your C:\drive - it could be on a separate drive.
It makes sense to have a folder for Git Repositories - lets call it "Git"
Highlight the folder in windows and in the address field type "cmd"
This opens windows command-line.

In the command-line window Type>pip install zcmds - NOT zmcds!!!!
This installs the repo and pulls all the requirements needed to run the sub-programs

EXAMPLE: CONVERT AN IMAGE TO WEBP FORMAT
For example if you wanted to convert an image from a .png to .webp
In Windows

1. Create a folder in a convenient location - say E:\Git\zcmds - lets call the new folder "input"

2. Load image files you want to convert. (Consider renaming them or not)

3. Create an output folder - lets call it "output"

4. Navigate to the repository and in the address-bar type "cmd"
   In the command-line E:\Git\zcmds>

5. type >img2webp input E:\Git\zcmds\input\image-001.jpg --format webp --output E:\Git\zcmds\output

6. To access Help with this command type >img2webp -h

aicode - General question and detail instrutions

Hi Zack,
Nice little tool. I am working on a matching learning program to add to Streamlit and it is giving warnings and errors. I used chatgpt and google ai directly but while chatgpt re-wrote the entire code a bit differently, it is still giving warnings and errors.

A little bit of more clarification on how to use it would be helpful. Sorry, I am crazy for cheatsheets and documentations...

Thanks, Aaliyah

Unrecognized option 'cq'

execute
!vid2mp4 --rencode --nvenc --preset 'medium' --crf 23 'ForBiggerBlazes.mp4'

result

Running:
  static_ffmpeg -hide_banner -i "ForBiggerBlazes.mp4"  -vcodec h264_nvenc -preset medium -cq 23 -c:a copy -y "ForBiggerBlazes_converted.mp4"
Unrecognized option 'cq'.
Error splitting the argument list: Option not found
Generated ForBiggerBlazes_converted.mp4

ValueError: rate must be specified when data is a numpy array or list of audio samples

execute
!vid2mp3 -o 'vid2mp3.mp3' -n 'ForBiggerBlazes.mp4'

result

-o does not exist
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
[<ipython-input-83-be798a73c399>](https://localhost:8080/#) in <cell line: 3>()
      1 get_ipython().system("vid2mp3 -o 'vid2mp3.mp3' -n 'ForBiggerBlazes.mp4'")
      2 
----> 3 Audio('vid2mp3.mp3')

[/usr/local/lib/python3.10/dist-packages/IPython/lib/display.py](https://localhost:8080/#) in __init__(self, data, filename, url, embed, rate, autoplay, normalize, element_id)
    114         if self.data is not None and not isinstance(self.data, bytes):
    115             if rate is None:
--> 116                 raise ValueError("rate must be specified when data is a numpy array or list of audio samples.")
    117             self.data = Audio._make_wav(data, rate, normalize)
    118 

ValueError: rate must be specified when data is a numpy array or list of audio samples.```

got the various files when executre audnorm

execute
!audnorm 'ForBiggerBlazes.mp4' 'audnorm.mp4'

result
got the various files
e.g.

-rw-r--r-- 1 root root  367168 Apr 12 00:49 audnorm.m4a
-rw-r--r-- 1 root root 2503095 Apr 12 00:39 audnorm.mp4
-rw-r--r-- 1 root root 2650156 Apr 12 00:49 audnorm_norm.wav
-rw-r--r-- 1 root root 2650190 Apr 12 00:49 audnorm.wav

expected result
just the mention file when execute the command 'audnorm.mp4'

transcribe-anything torch version error

Hi ,

I was trying to use the transcrcribe tool and it gave me an error. Not even chapgpt was able to resolve it. I searched other platforms.

I upgraded to the latest pip and py versions. It doesn't like the latest py version.

Running: pip install torch==2.1.2
ERROR: Could not find a version that satisfies the requirement torch==2.1.2 (from versions: 2.2.0, 2.2.1)
ERROR: No matching distribution found for torch==2.1.2

My current py Python 3.12.1.
torch - 2.2.1+cu118

got the various files when executre vidclip

execute
!vidclip --start_timestamp 0 --end_timestamp 3 --outname 'vidclip.mp4' --crf 23 'ForBiggerBlazes.mp4'

result
got the various files
e.g.

rw-r--r-- 1 root root  320950 Apr 12 01:19 vidclip_alt.mp4
-rw-r--r-- 1 root root  320950 Apr 12 01:18 vidclip.mp4

expected result
just the mention file when execute the command 'vidclip.mp4'

pipx install zcmds isn't working

It might be worth updating the instructions to pipx install zcmds. I received the following:

$ pip install zcmds

error: externally-managed-environment

× This environment is externally managed
╰─> To install Python packages system-wide, try apt install
    python3-xyz, where xyz is the package you are trying to
    install.
    
    If you wish to install a non-Debian-packaged Python package,
    create a virtual environment using python3 -m venv path/to/venv.
    Then use path/to/venv/bin/python and path/to/venv/bin/pip. Make
    sure you have python3-full installed.
    
    If you wish to install a non-Debian packaged Python application,
    it may be easiest to use pipx install xyz, which will manage a
    virtual environment for you. Make sure you have pipx installed.
    
    See /usr/share/doc/python3.11/README.venv for more information.

note: If you believe this is a mistake, please contact your Python installation or OS distribution provider. You can override this, at the risk of breaking your Python installation or OS, by passing --break-system-packages.
hint: See PEP 668 for the detailed specification.

System specs:

$ lsb_release -a
No LSB modules are available.
Distributor ID:	Debian
Description:	Debian GNU/Linux 12 (bookworm)
Release:	12
Codename:	bookworm
$ python -V
Python 3.11.2
$ pip -V
pip 23.0.1 from /usr/lib/python3/dist-packages/pip (python 3.11)

pip install zmcds

Your README file says:

pip install zmcds
zcmds
diskaudit

It does not work with me.
But it works with
pip install zcmds