xonsh / xontrib-template Goto Github PK

https://github.com/MichaelAquilina/zsh-you-should-use

ZSH plugin that reminds you to use existing aliases for commands you just typed

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

Hello! I have an idea!

Sort files or dirs in completion by the createion or change date i.e.:

cd /tmp
echo "hello" > 1
echo "world" > 2
#cat <Tab>
# 2
# 1 

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

After xonsh/xonsh#4827

Hello! I have an idea!

Example 1:

# Create file `1.txt` somewhere using GUI text editor
cd /tmp
echo "hello" > 1.txt
vim # Press key shortcut to get completer list
vim /path/to/somewhere/1.txt

Example 2:

cd /tmp
echo "hello" > 1
echo "world" > 2
#cat <Tab>
# 2
# 1
cat 2
# world 

Example of the xontrib name: xontrib-completer-last-file.

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

Hello! I have an idea!

Using DuckDB for xonsh history backend - https://dirk-petersen.medium.com/researchers-please-replace-sqlite-with-duckdb-now-f038044a2702

Use implementation of sqlite as start point - https://github.com/xonsh/xonsh/blob/5afb83d67e0846d3401b9bc0581b6c6a33ff5d76/xonsh/history/sqlite.py

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

Hey @jnoortheen !

I tried to run copier to create xontrib but got an error. It looks that current instructions is not working.

docker run -it --rm -v $PWD:/ddd python:3.9-alpine sh
pip install copier jinja2-time cookiecutter
copier gh:xonsh/xontrib-template .
# Unknown sub-command 'gh:xonsh/xontrib-template'

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

• completion definitions format

  • https://pypi.org/project/completions/
  • https://github.com/fvictorio/completely
  • fig autocompletions - typescript definitions can be used as the starting point

• autocompletion based on --help. we can write a cli that can run a command and generate completions from --help command

  • https://github.com/dim-an/cod
  • https://xon.sh/tutorial_completers.html
  • https://dvc.org/blog/shtab-completion-release
  • https://github.com/frostming/pycomplete

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

@anki-code does this template need requirements.txt ? we use PEP-621 meta data or poetry.

upd: This seems to be another issue with autoloading, the tests simply fail with xontrib_x folder structure this template creates by default, they can only load from xontrib/x.py due to some hardcoded names in the load_xontrib function and maybe other reasons

I've decided to check why the CI from this template was failing despite the installation of xontribs working fine and discovered that running poetry run pytest fails with tests/test_xontrib.py::test_it_loads - Exception: Failed to load xontrib: The following xontribs are enabled but not installed:['cd']

Strangely enough, when I load xonsh directly from the pytest path with ~/Library/Caches/pypoetry/virtualenvs/xontrib-cd--vcGhuku-py3.11/bin/xonsh --no-rc (from a non-xonsh shell to avoid any environment contamination) I can load it just fine

from xonsh.xontribs import xontribs_load
xontribs_load(['cd'])
# (None, None, <ExitCode.OK: 0>)

So I don't understand what's going on

Do you maybe have an example of a xontrib that uses this template and also has working tests? I've looked through a bunch of awesome xontrib list, but they either don't use the template, or don't use poetry.
Or maybe you know what's wrong

Hello! I have an idea!

The ability to create command template. Before executing the command the placeholders should be filled interactively.

First step - Ask to input argument and with autocomplete:

echo I am saying @(ask('What to say'))
# What to say: hello
# I am saying hello

echo Give @(ask('Fruit', ['apple', 'banana', 'orange'])) to @(ask('To', [$(whoami).strip()]))
# Fruit: <Tab>
# Fruit: apple
# To: <Tab>
# To: user
# Give apple to user

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

The template suggests disabling autoloading via
$XONTRIBS_AUTOLOAD_DISABLED = {"ptk_shell", }

There are a few of issues with this:

• It would override the same variable set by other xontribs instead of appending
• From the docs, the type of XONTRIBS_AUTOLOAD_DISABLED could either by a bool or a list, but you use a set, so actually appending this variable becomes a challenge due to a difference in updating lists and sets

Fish has this very nice trick of a functionality that allows you to CD to previous/next dir in history with ⌘◀/▶ WITHOUT creating any new prompts, so the existing prompt just gets updated in the background

In xonsh it's also possible to update the exising prompts, see the gist below for a sample xontrib that CDs up one folder on ⎈z and updates all 3 prompts (if they exist) "inline", i.e., without generating a new line like the carriage return workaround does

https://gist.github.com/eugenesvk/fb8078d7388fc4d722f2c4b88f63cda1

This xontrib would rely on this, but add env vars to allow changing keybinds and use os-relevant defaults (I think temrinals on Macs map ⌘◀/▶ to Send ^[ [1;5H Send ^[ [1;5F)

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

Links to search new ideas:

• https://jvns.ca/blog/2022/04/12/a-list-of-new-ish--command-line-tools/

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

We’ve created an improved version of OpenAI Codex, our AI system that translates natural language to code, and we are releasing it through our API in private beta starting today.

• Open AI Codex: https://openai.com/blog/openai-codex
• How it could works in zsh: https://github.com/tom-doerr/zsh_codex
• How to create xontrib: https://github.com/xonsh/xontrib-template
• The example of adding text into terminal

The example of code for catching key press and add text:

from prompt_toolkit import prompt
from prompt_toolkit.key_binding import KeyBindings

# Define key bindings to replace text on space key press
kb = KeyBindings()
@kb.add(' ')
def _(event):
    event.current_buffer.text = event.current_buffer.text + ' space :)'
    event.current_buffer.cursor_position += len(event.current_buffer.text)  # move coursor to the end

# Start a prompt session with key bindings
text = prompt('> ', key_bindings=kb)
print('You entered:', text)

Draft 1:

import openai

openai.api_key = "YOUR_API_KEY_HERE"  # https://platform.openai.com/account/api-keys

# Example prompt
prompt = "I love Python because"

# Example completion using the GPT-3 model
response = openai.Completion.create(
    engine="davinci", prompt=prompt, max_tokens=50
)

# Print the response
print(response.choices[0].text)

Draft 2:

import os
import requests
import json
import time

from prompt_toolkit.keys import Keys

$OPENAI_API_KEY = '...' # https://platform.openai.com/account/api-keys

@events.on_ptk_create
def custom_keybindings(bindings, **kw):
    @bindings.add('c-k')  # control+k
    def run(event):
        input = event.current_buffer.text
        print('\n# ...')
        
        headers = {
            'Content-Type': 'application/json',
            'Authorization': 'Bearer ' + __xonsh__.env.get('OPENAI_API_KEY'),
        }

        json_data = {
            'model': __xonsh__.env.get('OPENAI_MODEL', 'text-davinci-003'),
            'max_tokens': __xonsh__.env.get('OPENAI_MAX_TOKENS', 1000),
            'temperature': __xonsh__.env.get('OPENAI_TEMPERATURE', 0),
            'prompt': input,            
        }

        output_delay = __xonsh__.env.get('OPENAI_OUTPUT_DELAY', 0.01)
        
        try:
            response = requests.post('https://api.openai.com/v1/completions', headers=headers, json=json_data)     
        except Exception as e:
            print(f'\n# Error request: {str(e)}')
            return        
        
        if response.status_code != 200:
            print(f'\n# Error response: {response.status_code}')
            return
        
        try:
            j = response.json()
        except Exception as e:
            print(f'\n# Error response: {response}')
            print(f'\n# Error exception: {str(e)}')
            return
        
        output = j['choices'][0]['text']
        for w in output.split(' '):
            event.current_buffer.insert_text(w + ' ')
            event.current_buffer.validate()
            time.sleep(output_delay)
            
# git commit <control+k> 

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

Hello! I have an idea!

https://asciinema.org

A xontrib that automatically records every session using aschiinema, ofc written in Python already 😄
Could their PTY implementation be useful for capturing output in general? If ran around xonsh it doesn't miss a byte pushed to the screen! 😸

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment

Hi @jnoortheen!

I'm trying to publish xontrib based on xontrib-template to pypi - https://github.com/anki-code/xontrib-jump-to-dir/blob/4c505cc37b7e426f82ccda21564f94b5971f5f02/.github/workflows/release.yml#L58-L61

And I'm facing with issue:

How I should fix this?
Please fix xontrib-template to avoid this error for the new template users.

Thanks!

Thanks for the fantastic work.

It's very likely that I'm just not good enough in searching issues here/missing the correct keywords so apologies if that's the case in advance.

The one feature that keeps me from switching to xonsh from ipython is the availability of a %autoreload like functionality when using it as a REPL when developing, that recursively imports the modules and its dependencies.

Is there such a function available/in planning?

Thanks!

For community

⬇️ Please click the 👍 reaction instead of leaving a +1 or 👍 comment