tools-and-devops

[urban pecan]

but i'm not sure about that yet

@tawdry needle does that answer your question?

[tawdry needle]

q?

from justuse import use seems reasonable

[urban pecan]

https://pypi.org/project/q/

i don't see anything that couldn't be put into the class and keep the simplistic import use though

any particular (technical) reason to take the less beautiful route?

[tawdry needle]

Callable imported thing is confusing and weird to me

[urban pecan]

so is importing things via a function instead of import 😉

[tawdry needle]

Meh

Modules are objects

It's ok to return them, cf importlib

It's not that weird but maybe something would break if import doesn't return a module... good q for #internals-and-peps

[urban pecan]

actually, i stumbled over a strange thing that reminds me of this, the only actual issue i had with use() is with dataclasses

it seems dataclasses also manipulate sys.path and does weird tricks on import, which kind of crosses streams with use()

but i never had a problem with use itself not being a module so far

i'll reconsider if someone shows me an example of a lib that operates on modules that couldn't work with use

[tawdry needle]

I imagine there could be other issues, i just don't see a reason for it

All to save the from?

[urban pecan]

if there was other stuff in the module the user might want, it would make sense to require the from, but as i see it, that's the most succinct form, so i prefer that

[supple venture]

Ah yes, thought so, thanks!

[rugged flame]

Does GitHub really not have a way to view markdown/rst files without rendering them, or am I missing something? 🤔

Context: I'd like to refer to a line of a rst file in an issue.

[burnt ibex]

You can click the "Raw" button on the file

[rugged flame]

Hmm, but then you get a text file, and you can't get a permanent link to a specific line.

[burnt ibex]

You can click "Blame" I think, but that has some extra stuff

[rugged flame]

Ah right. Thanks!

[finite fulcrum]

While not lines, if it has headings etc. you can go to those

[urban pecan]

i was just wondering if it might be useful to add p2p capability to pip? if i add an auto-install feature to use() it might be feasible to add a p2p layer?

or is there already p2p planned for pip itself maybe?

[urban pecan]

i'm just thinking that maybe with IoT this could be useful.. like if you had a swarm of drones, you could tell one to use a package, which pulls it from pypi, then the other drones could install that package without burdening pypi

[wooden ibex]

Nope

[ebon escarp]

lol

[urban pecan]

what nope?

[runic verge]

i am surpised with hearing how many companies store their repositories on github ... and was wondering why do they do that . Why put their trust into github

i saw theres Git .. which i saw explained as Git is software for tracking changes in any set of files, usually used for coordinating work among programmers collaboratively developing source code during software development. Its goals include speed, data integrity, and support for distributed, non-linear workflows And that makes a lot of sense to integrate that into your company .. rather than creating a system to do that for your company

[eternal flicker]

not that different from using cloud hosting instead of on-site

[runic verge]

i havent uploaded anything that is commercial to github, even in a private repository , becuase i see its one more company that can see all of my source code
so thats why its wild to me that larger companies with a lot more expensive software are doing it

[runic verge]

I see that

i have uploaded to github .. stuff that consider to be opensource

[eternal flicker]

so it's mostly a matter of companies employing the services of another company

[runic verge]

and Git sounds incredible ... you can upload all your source-code where all the engineers are able to reach it , AND it can be integrated with Git for seamless updating , i see trust is kind of out of the equation

[eternal flicker]

i see trust is kind of out of the equation
how so?

[runic verge]

Well becuase the gains are so high .

[finite fulcrum]

I really doubt private repos are readily available to employees of github or that a company as large as them would risk doing something with them

[runic verge]

to having it integrated . And they are a huge company , and of course are legally responsible for protecting the data they store . but i see that is overlooked.

[finite fulcrum]

Hosting git or other vcs is not that uncommon, but I doubt there's much more risk having it on github compared to hosting it yourself and exposing it to the internet

[supple venture]

you could always host your own git cloud server

many do that

[runic verge]

i currently have an auction bot on github. It provides valuable information about which items are available at auctions around the US . It's routed with an API . Working on the API and protecting the code .. and just the frontend and integration is opensource

[subtle horizon]

hey can I get some help regarding brython?

im trying to develop a chrome extension using brython. I can get the python console to open in chrome, but I can't get it to run MY script

[tawdry needle]

@runic verge lots of companies pay github for private repositories or even on-prem deployments hosted on internal networks

Version control is really a life-changing tool in programming, distributed version control like git even moreso

Git itself is a very powerful tool with an elegant underlying model... and a confusing command line interface that can be difficult to learn

[runic verge]

I appreciate the suggestion to look into DevOps @civic hound , i have heard from others that its important aswell , and i see it is important in a team environment.

[runic verge]

i've seen . I have seen already so many times where version control would be life-changing on a project that only i am working on . And then projects with other devs .. it's incredible.

[runic verge]

i dont like working on github files, on github (editting the saved code that is on github with the changes of the file on my computer or deleting and uploading the same file when updating) ... I see Git can connect to the repository and make those changes from CMD or from the Git GUI

[runic verge]

i am not sure about brython or whether its considered tools /devops . There may be a better channel for the subject . .. I see #❓｜how-to-get-help and creating a help channel could get some help (though im not sure of that either , because i see brython may be a specific thing that fewer people are aware of ) . Maybe talking about it in python-general , because there are more users there .

[subtle horizon]

thanks for replying. I am going through the documentation to see if I can find anything. It's unlikely that I'll get help for such a specific thing

[runic verge]

hey guys, if anyone knows anything about the http.server module, i would greatly appreciate any help in #help-chestnut
i just saw this phrased like this , in the python-general channel and i thought of the question i heard you had
i see this is a good way of doing it
as the http.server module is also a specialized thing
hope you find what youre looking for @subtle horizon

[opaque gorge]

How would I wait for another github workflow to finish, before continuing with my one?
I know I could do this: ```yml
on:
workflow_run:
workflows: ["Verification"]
branches: ["main"]
paths:
- pyproject.toml
- setup.cfg
- setup.py

which works well, except it could trigger multiple times if say, both `setup.py` and `setup.cfg` were edited, the workflow would run twice, which isn't what i want.

If I however use on push, I don't know how would I wait for the `Verification` workflow to finish, before continuing, and if it fails, this workflow should fail too

Please use reply, or ping me on response

[supple venture]

In the same workflow or two different workflow files?

[opaque gorge]

2 different files

[supple venture]

https://docs.github.com/en/free-pro-team@latest/actions/reference/events-that-trigger-workflows#workflow_run you can use the workflow_run trigger

name: PyPi Publish

# Only trigger, when the lint workflow succeeded
on:
  workflow_run:
    workflows: ["Lint"]
    types:
      - completed

Example

name of first file should be the exact match of the workflow_run: workflows: ["Lint"] in the second file.

[opaque gorge]

yeah, that's what I did in my example, but since I only want it to trigger on separate files, if they're specified, this triggers 3 times if all 3 files were changed in single commit

[supple venture]

Basically are u trying to run the workflow only on those file edits right?

[runic verge]

@opaque gorge

oh that was hours ago .

hearing about workflows i am wondering if that's how git projects can be seperated from eachother .. and each project can be worked on seperated . Right now when i do git add <filename> i see it adds it to the same project ... but i am working on something that doesn't belong in the project it is adding the file to . ( we want to have a seperate project .. is it a workflow?)

also looking for a way to link this to github , and separate github projects

[rapid sparrow]

nah
https://www.atlassian.com/git/tutorials/comparing-workflows/gitflow-workflow (eng)
https://www.atlassian.com/ru/git/tutorials/comparing-workflows/gitflow-workflow (rus, if necessary)

workflows are all about organizing work for one repository and their meanings

1. main one about actually collaboration with other developers
2. secondary one, to make job easier for devops guys to deploy releases of your project

if you want to have a separate project... it belongs to separate repository.
one repository should have only everything that belongs purely to that project and at maximum it can have... some additions that can be launched only together with the project. (for example my Django projects, have included into them docker-compose/docker scenarios to build Nginx / Redis / Celery containers together with main project,)

I usually answer for myself question, if it belongs to separate repository or not, but asking myself... does it need separate ci/cd, its own pipline runner? if yes -> then create new repository

btw, recommending to read atlassian guide about git from the beginning "Learn git" up to "Forking workflow" chapter

quite cool one in my opinion

p.s. wink, we are probably from the same country, or at least our origins are.

[runic verge]

i figured out what i was looking for , in the meantime . I was looking for pushing to different repositories with minimal reconfiguration.. turned out to be very simple .

git remote add origin2 https://github.com/swordysrepo/usegit2.git
git origin2 branch random
git push -u origin2 random```

jumping between repos checking that out.

[runic verge]

this i see is advanced for where i am at . It's what i'm looking for though. I appreciate the mention . Seeing it's what connects Git on someones desktop to a network of "Gits on desktops" and organizes them so they can be committed with less overlapping .

[runic verge]

https://github.com/swordysrepo/Windows-10-Toast-Notifications/pull/1
the first github PR request
it was merged! 😄
Such a simple fix also . (but i looked into it for hours .. to discover where the problem was originating from) @civic hound

[runic verge]

the profile picture wasn't me . Just a guy i respect . And haven't meant to use this account for the purpose of discussions/joining technical discord when i made the account years ago , so i hadnt changed the profile until now, after a while of using discord for that purpose.

[foggy token]

for readthedocs how do i make it not automatically install requirements.txt?

nvm

[ebon swift]

Thanks, I see that it checks the version info from a clone in /tmp

[ripe junco]

Anyone know how to correctly install all azure SDKs for Python? I'm trying to install azure and I get this error

ERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output

[supple venture]

how are you installing it? i mean whats the command

[ripe junco]

pip install azure as mentioned https://pypi.org/project/azure/

[supple venture]

pip install -U setuptools does this hlp?

[ripe junco]

Nop, still get the same error

[supple venture]

umm, do pip --version

[ripe junco]

Even if I specify the version like this pip install azure==5.0.0 it doesn't work

[ripe junco]

20.0.2

[supple venture]

try updating it

[ripe junco]

Still get the error

Even with the upgraded pip

[supple venture]

could you send the full output of when you do pip install azure

[rancid schooner]

Hey @ripe junco!

Uh-oh! It looks like your message got zapped by our spam filter. We currently don't allow .txt attachments, so here are some tips to help you travel safely:

• If you attempted to send a message longer than 2000 characters, try shortening your message to fit within the character limit or use a pasting service (see below)

• If you tried to show someone your code, you can use codeblocks
(run !code-blocks in #bot-commands for more information) or use a pasting service like:

https://paste.pythondiscord.com

[ripe junco]

https://paste.pythondiscord.com/aquzerezut.sql

Ok, I think I've solved it

I've installed version 4.0.0

[supple venture]

Starting with v5.0.0, the 'azure' meta-package is deprecated and cannot be installed anymore.

[ripe junco]

Yeah, I did not pay full attention to the text 😅

Still, thanks a lot @supple venture

[supple venture]

https://pypi.org/project/azure/ you can read the README

[wooden ibex]

There is no reason to grab base azure package, they have packages for all their services including rest api. Which is pretty bad

[floral smelt]

I'm not sure it's the good section but I don't find anything ANaconda-related

conda has a bug with dependencies. Soemtimes it takes eons before checking on the conflicts. I abort with Ctrl+C and relaunch the command

And it works Idg why. Anybody has this bug?

[trim solstice]

i have a django powered website that needs to be hosted for a fairly affordable price, ive been given the target of <$30/mo but im not sure if that will happen.

where might i get a decent VPS for this price?

[burnt ibex]

I believe this community hosts their site on Linode

[trim solstice]

im not sure what that is, is it a VPS provider?

[burnt ibex]

You could also use something such as AWS, Digitial Ocean, etc.

[burnt ibex]

That's one thing it offers, yes

[trim solstice]

aws seems good, too many options for me to wrap my head around tho lol

im really not a devops person, more frontend tbh. we're a small company so ive just been loaded with everything to do with the website.

what would aws's webhosting service be? or at least the best fit

[burnt ibex]

https://docs.aws.amazon.com/elasticbeanstalk/latest/dg/create-deploy-python-django.html

[trim solstice]

ah awesome! i'll have a read through. thank you

[burnt ibex]

Np

[trim solstice]

just to check, what im getting from that is i need an EC2 server instance, and then i use EB to manage moving the the project from my computer to the server as well as linking to a domain, did i get that right?

[burnt ibex]

Uh, not sure, I haven't read through that whole thing

But that sounds right

[trim solstice]

okay cool, i'll have a play around with this leading up to the end of the project and hopefully get everything sorted. thanks again!

[twin valley]

I just stumbled upon https://deepnote.com/home, does anyone use this? Would love some thoughts if you do

[short peak]

isn't this what jupyter does too?

[tawdry needle]

all these not-jupyter notebooks are too focused on "cloud" and "collaboration" and not enough on "being more usable" imo

it's the same feature set rehashed over and over

data science ux generally still sucks

[twin valley]

I guess their focus is a little bit more "publishing" slanted and trying to make the ui/ux a little cleaner looking but yeah; I'm a fan of https://observablehq.com/ and I think this is someone trying to do something similar for python? 🤷

[frozen anchor]

Any1 here using GitHub SuperLinter? I started using it a while ago and it works - showing me that I have linting errors on PR.
Question is: Is there a way to automatically fix the (easy) lint problems - something like x= 1 to x = 1? I didnt find an option to autoformat the code.
I think black allows that - however I use other languages as well in my project so a tool like "SuperLinter with Autoformatting" would be amazing.

[urban pecan]

@frozen anchor just use black 😉

or brunette for a little less black

[urban pecan]

is there a way to use pip inside a script as an API from within a script without calling it as a subprocess?

i can call it as a subprocess, but python -m pip download requests==2.6.0 for instance puts the file in the cwd with a name that isn't 100% predictable

i could glob for that with the name-version as a pattern, but it would be nicer to have a function that would just give me the file name

after downloading it

having asked this, i just discovered https://pypi.org/project/pip-api/

i'll try that

[tawdry needle]

it's unofficial though

pip doesn't guarantee any stable python api (annoyingly)

you could parse the output i guess

import re
import subprocess

proc = subprocess.run(['pip', 'download', name], stdout=subprocess.PIPE)
m = re.search(r'(?m)^Saved (.*)$', proc.stdout)
save_location = m.group(1)

very annoying that you can't configure this

maybe use a temporary directory, and run pip in that directory? using the cwd= option

[urban pecan]

cwd option of subprocess?

or?

@tawdry needle

[tawdry needle]

yeah, subprocess.run has a cwd kwarg

!d subprocess.run

[rancid schooner]

subprocess.run

subprocess.run(args, *, stdin=None, input=None, stdout=None, stderr=None, capture_output=False, shell=False, cwd=None, timeout=None, check=False, encoding=None, errors=None, ...)```
Run the command described by *args*. Wait for command to complete, then return a [`CompletedProcess`](https://docs.python.org/3/library/subprocess.html#subprocess.CompletedProcess "subprocess.CompletedProcess") instance.

The arguments shown above are merely the most common ones, described below in [Frequently Used Arguments](https://docs.python.org/3/library/subprocess.html#frequently-used-arguments) (hence the use of keyword-only notation in the abbreviated signature). The full function signature is largely the same as that of the [`Popen`](https://docs.python.org/3/library/subprocess.html#subprocess.Popen "subprocess.Popen") constructor - most of the arguments to this function are passed through to that interface. (*timeout*, *input*, *check*, and *capture\_output* are not.)

[urban pecan]

okay

[tawdry needle]

!d tempfile

[rancid schooner]

tempfile

Source code: Lib/tempfile.py

This module creates temporary files and directories. It works on all supported platforms. TemporaryFile, NamedTemporaryFile, TemporaryDirectory, and SpooledTemporaryFile are high-level interfaces which provide automatic cleanup and can be used as context managers. mkstemp() and mkdtemp() are lower-level functions which require manual cleanup.

[tawdry needle]

use TemporaryDirectory

[urban pecan]

yeah, i planned to use that for the "not-in-a-virtual-env" case.. i figured out how to detect whether i'm in a virtual env or not, should i use TemporaryDirectory either way?

it might be a bit awkward (but probably safest) to use it in all cases

it will certainly be slowest to re-install it every time the program is run, though.. on the other hand, i'm planning on incorporating some p2p distribution of packages, which might alleviate this problem somewhat

let's try to make it work for all cases first and have it save and slow before making it fast 🙂

safe*

[tawdry needle]

This is just for controlling where pip download dumps its files right?

Maybe you can integrate with the existing pip cache somehow

Pip already has a download cache

Would be silly imo not to be able to use it for this

Why do you even use pip download? I believe pip install has options to control where exactly the package is installed to

[unique imp]

s

[urban pecan]

@tawdry needle i need to first download the file in order to check its hash before installing anything malicious

having looked at pip's code, i think it might be most reasonable to simply utilize pypi's pretty slick json API

via requests

so far it's no big trouble to get the fitting wheel, the bigger next hurdle is how to install it in a way it can be directly loaded

i fear only packages with a setup.py (whose requirements are already fulfilled or that can be auto-installed themselves) can be auto-installed

which wouldn't be a big issue for me, but others might complain 🙂

with auto-install requiring a proper hash with version-pinning, it might even be a more future-proof way overall

i feel no pressure to match or replace pip in all its features, though. i think it's a great alternative as a command line tool

but thinking of typoes and malicious packages on pypi, insisting on version pinning and hashes is the most straight forward way to get to a really secure supply chain

which, in my opinion, is done best directly in the code doing the import to be 100% clear on what exactly is used

[urban pecan]

if that only contains the wheels, i could use that, sure

honestly, i feared getting the package from pypi would involve a lot more jumping through hoops, but it was very straight forward after i discovered https://warehouse.readthedocs.io/api-reference/json.html

it's also easy to check the available links for a matching hash, so that's even simpler for the user than what pip hashing mode suggests - it asks the user for a list of possible hashes

installing and importing a package might become as simple as use("yourpackage", version="1.2.0", hash_value="7a7a8b91086deccc54cac8d631e33f6a0e232ce5775c6be3dc44f86c2154019d", auto_install=True)

copy&paste-able anywhere 🙂

[lament compass]

does someone mind helping me with a problem that ive facing while working on a python script for the back end?

ModuleNotFoundError: No module named 'sklearn.linear_model.logistic'
this is the error that m getting and Ive already installed sklearn

dm me if someone thinks they can help me out

[rapid sparrow]

did you import it

try import sklearn before where you call it

[lament compass]

Yes I did

I am in dire need for some help I am a beginner and I have a very important submission tomorrow I've been working on this for 2 weeks and I can't solve this error

[frozen anchor]

@lament compass this is a tools/devops channel and your question is not related to tools/devops at all. I suggest you open a help channel with:
The code you want to execute
The error message

[lament compass]

M trying to develop a flask app and having trouble with a python module that's why I ended up posting the quesry here

[shy locust]

I think you have to import the specific inner module of sklearn

From sklearn import linear_model then call linear_model.logistic_regression()

[wintry iris]

what should one do to have a perfectly structured python project on github?

i've looked around at repos, but can't find a certain common way

[tawdry needle]

there's no "perfect" structure

and don't worry about github so much - whatever is good in general is good for github

i generally follow this layout, except i also include a pyproject.toml file https://blog.ionelmc.ro/2014/05/25/python-packaging/

[short peak]

same ☝️

[sacred dragon]

does any 1 here has uploaded python package in pypi

i'm getting this error even tho my token is correct

[runic verge]

actually yes .. theres a way to pass in a file .. that has the credentails

i forget whether i found out how to do this without the file ..

It is a common issue

what are you using to upload to pypi? @sacred dragon

I heard twine is a package that is the one that is supposed to work . And heard other packages might throw errors like this . From what i've gathered about uploading to pypi .. it is a fragile process , but once you've got it working it should work consistently.

[sacred dragon]

i'm using twine

consistently

it's my third release

and it shows like this for the first time

[heavy knot]

Hi this is my post here, I am trying to load and work with DMN files. Can anyone suggest me a Library to use?

[wintry iris]

do these not work?

[heavy knot]

I've created dmn xml based file in signavio and then trying to import in python. by pyDMNrules i can't do it. i tried cDMN API it doesn't work

[wintry iris]

ill look around

never used a dmn file tho 🤷‍♂️

[opaque tiger]

    ERROR: Command errored out with exit status 10:
     command: 'c:\users\co\appdata\local\programs\python\python38\python.exe' -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'C:\\Users\\Co\\AppData\\Local\\Temp\\pip-install-l_p5bva2\\pycurl_ca46745b7aa24090bf02fc06186d5660\\setup.py'"'"'; __file__='"'"'C:\\Users\\Co\\AppData\\Local\\Temp\\pip-install-l_p5bva2\\pycurl_ca46745b7aa24090bf02fc06186d5660\\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' egg_info --egg-base 'C:\Users\Co\AppData\Local\Temp\pip-pip-egg-info-1q1wejsn'
         cwd: C:\Users\Co\AppData\Local\Temp\pip-install-l_p5bva2\pycurl_ca46745b7aa24090bf02fc06186d5660\
    Complete output (1 lines):
    Please specify --curl-dir=/path/to/built/libcurl
    ----------------------------------------
ERROR: Command errored out with exit status 10: python setup.py egg_info Check the logs for full command output.
``` When trying to install pycurl.

Anyone? Would be very apprciated.

[heavy knot]

would be great help if you find around a solution 🙂

[wintry iris]

can't find anything 😔

[regal ledge]

Maybe this is the correct channel..

Hello.
I have been struggling with this issue for a long time now and I can't find answers online.

I'm trying to deploy a Flask app to an Apache server, with fcgi.
The file that runs the app, dispatch.fcgi, is below.

The issue is that when deployed, it can only find the routes that are @idle salmon.route in the same file.
It does not find any of my blueprints or blueprint routes. Nor can I create blueprint routes in this file after creating the app.

The application works on my local. Only difference is that instead of this file I run a run.py which just creates the app without extra config that Apache wants.

How can I get it to find my blueprint routes?
PS: It's a shared host, therefore I don't have root access so the fix I'm looking for should probably be on py side.

#!/usr/bin/env python3.7

import os, sys
sys.path.insert(0, "/www/apache/domains/www.somedomain.com/hinnang/hinnangapp")
from hinnangapp import create_app

os.environ['REQUEST_METHOD'] = "GET"
os.environ['SERVER_NAME'] = "somedomain.com"
os.environ['SERVER_PORT'] = "80"
os.environ['SERVER_PROTOCOL'] = "HTTP/1.1"
DOMAIN = "somedomain.com"
PREFIX = "/www/apache/domains/www.somedomain.com/"

sys.path.insert(0, os.path.join(PREFIX, ".virtualenvs/hinnangapp/lib/python3.7/site-packages/"))

from flup.server.fcgi import WSGIServer
from flask import Flask

app = create_app()
@app.route("/")
def hello(): #<- this works
    return "Hello World! <a href=\"test/\">to test</a>"


class ScriptNameStripper(object):
   def __init__(self, app):
       self.app = app

   def __call__(self, environ, start_response):
       environ['SCRIPT_NAME'] = ''
       return self.app(environ, start_response)

app = ScriptNameStripper(app)

#@main.route("/home") #<- tried to create a route for a blueprint here, didn't work
#def home():
#    return "home"


if __name__ == '__main__':
    WSGIServer(app).run()

[runic verge]

having a problem where git says everything up-to-date but its not

tried another way and i get GitHub - error: failed to push some refs to '[e-mail]:myrepo.git'

and found on google ... ```In my case git push was trying to push more that just the current branch, therefore, I got this error since the other branches were not in sync.

To fix that you could use: git config --global push.default simple That will make git to only push the current branch.

This will only work on more recent versions of git. i.e.: won't work on 1.7.9.5``` My branches are also not in sync , going to try the git config adjustment git config --global push.default simple and report back
heard it from this link https://stackoverflow.com/questions/10753268/github-error-failed-to-push-some-refs-to-gitgithub-commyrepo-git

it didn't change . hadn't noticed that i was getting this error before though error: src refspec database does not match any

[runic verge]

my branches are kind of a mess . Have been switching between them on git not knowing what it actually does . (it says certain files need to be removed before switching . )
and when i run git add (some files) after removing cached with git rm --cached . -r , and then commit the adds with git commit -m "database initial commit" , it says nothing changed . (even though there should be no record of the file , after it has been removed from cache)

(myactivity) [documentation +8 ~0 -0 !] C:\github\api-side-twitch
> git ls-files
.gitignore
README.md
(myactivity) [documentation +8 ~0 -0 !] C:\github\api-side-twitch
> git rm --cached .
fatal: not removing '.' recursively without -r
(myactivity) [documentation +8 ~0 -0 !] C:\github\api-side-twitch
rm '.gitignore'
rm 'README.md'
rm 'database.ini.example'
(myactivity) [documentation +0 ~0 -3 | +11 ~0 -0 !] C:\github\api-side-twitch
> git ls-files
(myactivity) [documentation +0 ~0 -3 | +11 ~0 -0 !] C:\github\api-side-twitch
> git add .gitignore
(myactivity) [documentation +0 ~0 -2 | +10 ~0 -0 !] C:\github\api-side-twitch
> git add README.md
(myactivity) [documentation +0 ~0 -1 | +9 ~0 -0 !] C:\github\api-side-twitch
> git add database.ini.example
(myactivity) [documentation +8 ~0 -0 !] C:\github\api-side-twitch
> git commit -m "initial database commit"
On branch documentation
Untracked files:
        code
        config.py
        connectdb.py
        create_api.py
        handle_requests_api.py
        models.py
        recieve_as_api.py
        templates/

nothing added to commit but untracked files present
(myactivity) [documentation +8 ~0 -0 !] C:\github\api-side-twitch
> git ls-files
.gitignore
README.md
database.ini.example
(myactivity) [documentation +8 ~0 -0 !] C:\github\api-side-twitch
> git push origin database
error: src refspec database does not match any
error: failed to push some refs to '[e-mail]:swordysrepo/api-side-twitch.git'

this is also a private repo

and i just set up SSH

i think i broke git 😄

[runic verge]

#help-dumpling
edit:
resolved #help-dumpling message

it wasn't SSH , i see it's becuase git was in the documentation branch and trying to push to the database branch (throwing the src refspec error) . creating a local database branch would have worked as git push origin database . but i decided to revert back to just 2 branches , main and staging

[heavy knot]

https://docs.python-requests.org/en/master/user/quickstart/?highlight=cookies#cookies

[shy mountain]

Any library I can use to detect languages used in a repository? like cloc but native within python?

[burnt ibex]

You could make a GET request to https://api.github.com/repos/{user}/{repostitory}/languages

https://docs.github.com/en/rest/reference/repos#list-repository-languages

[shy mountain]

I was more thinking of a local directory that might not even be git

[burnt ibex]

Oh, my bad

[shy mountain]

no problem, its a good thought!

I think I can just manually parse the output of cloc --json by using a JSON parser

[opaque tiger]

If you know how to code in Pycurl/pyurl shoot me a dm right now!

[reef lake]

Hi, has anyone done text extraction using pytesseract or worked with image pre-processing?

[deft trout]

Hey, can you help me? I just want to make a simple youtube downloader and I get this error every time 🙂

[burnt ibex]

Youtube downloaders are against its ToS, so we won't discuss that here.

[oblique finch]

hello folks, does someone know is there a way to strip JAVA_OPTIONS from output of java -jar...

IK it's python discord but it's messing up my python tool

[deft trout]

Oh, okay :)

[rancid schooner]

Hey @wise tiger!

It looks like you tried to attach file type(s) that we do not allow (.bat). We currently allow the following file types: .gif, .jpg, .jpeg, .mov, .mp4, .mpg, .png, .mp3, .wav, .ogg, .webm, .webp, .flac, .m4a.

Feel free to ask in #community-meta if you think this is a mistake.

[wise tiger]

where is a channel to learn phyton

[ancient socket]

Trying to visualize the way I do this project.
I have a piZero with a camera. Was going to do facial recognition.
its Headless to I have to SSH into it.
What is the testing like for this? I just need to make the pi create pictures,
send to a folder and have my desktop running a program to do facial recognition?
or is the piZero supposed to do it all? which its hard to test what it sees over SSH.

[wet stag]

If you have enough storage on the device, then you can keep your trained model there and try to recognize faces per-image based on that model.

That being said, modern systems such as Amazon Alexa have such massive always-changing-models, that data is definitely going to the cloud for processing

You can probably go either way

Either way, my recommendation will be getting the simplest solution working first

Then any related problems will be clear, and you'll have something that works that you can maybe redesign or expand on

I also recommend at having a look at publicly available cloud services, such as https://cloud.google.com/vision/docs/detecting-faces

This is a strong route to go, unless your intention is to understand the recognition and training algo's yourself, potentially to get closer to generic machine learning

[ancient socket]

Thanks, ya i was just trying to understand how it works to build it myself.

[mystic void]

It depends on the type of testing and what you are trying to optimize for:

• Create a dataset (ie. pictures from your camera) you can use to test on your desktop?
• Make it available through http so you can visualize it on your browser?

[shy locust]

What's the implication of terraform hitting 1.0 on version

Also I'm being sent for RHEL training, anyone know a good discord for that

[whole talon]

py charm vs vs code which is better

[lyric crown]

hey guys
I accidentally overwrote friend's package-lock.json file with my file upload (I added by file-upload a file with the same name)
I didn't know that it will do that, I thought that it will merge somehow (newbie mistake)
is it okay if I just copy the content of my friend's package-lock.json into mine?

[supple venture]

Better suited for #editors-ides, it ends up on personal preference both are great editors.

[rapid sparrow]

people usually use git for that

restoring file from previous commit

git checkout <path to file>

[timber hawk]

Hi, I m coding a program and I want it launch with windows, how can i do ?

[agile relic]

does anyone know of a good CLI python script i can run which generates prompts for filling in DESCRIPTION, README, puts init into a directory, etc? basically a quick package generator

[pseudo moon]

Sounds like you want cookiecutter

[agile relic]

this looks exactly like what i was looking for...THANK YOU. you saved me from spinning my wheels

[clear burrow]

NOTE: You must install tkinter on Linux to use MouseInfo. Run the following: sudo apt-get install python3-tk python3-dev

that's a weird note

[tawdry needle]

that will depend on your os of course... not all linux distros are the same

debian takes the approach of breaking out parts of the python stdlib into separate installable packages

[clear burrow]

I use Ubuntu until not

[rapid sparrow]

wow, there is venv for docker. It is named docker-machine.

[tawdry needle]

what would such a thing do?

[rapid sparrow]

apperently it is a shortcut to create virtual machine with running docker inside

so, a layer of isolation. Instead of running containers directly in your server, to run them in a quickly raised virtual machine

it would be useful for situations when your application is made of many microservices

in order to make isolated sandbox environment to test them without port and other conflicts

$ docker-machine create -d virtualbox testenv
Creating VirtualBox VM...
Creating SSH key...
Starting VirtualBox VM...
Starting VM...
To see how to connect Docker to this machine, run: docker-machine env testenv

$ docker-machine ls
NAME      ACTIVE   DRIVER       STATE     URL                         SWARM
testenv            virtualbox   Running   tcp://192.168.99.101:2376

To see how to connect Docker to this machine, run: docker-machine env testenv

[shy locust]

Time to learn ansible

[cloud elbow]

[git]

I made a commit, but I have new changes again, what's the best way to update with the same commit message?

[tawdry needle]

it's not a good idea to reuse the same commit message in general

you could write a shell script to get the previous one and then make a new commit with that message, but i don't know why would you want to

[cloud elbow]

oh god, it it just a small update to the last one

I had a typo

I want it to be just one commit

I did not push it yet

[short peak]

https://www.atlassian.com/git/tutorials/atlassian-git-cheatsheet

[heavy knot]

not sure where to ask this or if i can ask this in this discord, but can someone help me with docker? im getting really weird errors and i have no clue what could cause the errors

none of my imports are working when i go in the docker with docker-compose exec (containername) sh and try to run a py file

except the requirements.txt installed imports

[umbral briar]

container is isolated environment, only the modules you installed in Dockerfile are available

[heavy knot]

im trying import files that are in the same folder on the docker container

ill show you 1 sec

[rapid sparrow]

it depends if you are in your own branch, or in public branch

if in public branch... better reverting previous commit. (it will create second commit that handles reverthing previous one) Not exactly fulfilling the goal to keep shorter commit history though. But that's the only best way to cancel commit in public branch

if in your own branch (which you have before it will be merged to some public branch)

just do git reset --soft HEAD~1

it will cancel the last commit while keeping all your changes from it

and commit everything again

[heavy knot]

these files are in the root of the container, for some reason the Flask directory is still there even tho i removed it from my project

Dockerfile  __pycache__  app.py  boot       dev                 etc   kubemanifests.yaml  lib64  mnt  proc              root  sbin     srv  tmp  var
Flask       app          bin     config.py  docker-compose.yml  home  lib                 media  opt  requirements.txt  run   scrapes  sys  usr

then ill need to go to the directory app

Dockerfile  __init__.py  __pycache__  config.py  database.sqlite  forms.py  manager.py  models.py  reddit  routes.py  static  templates

and then i type: python manager.py db --help and its saying the imports arent working, while the imports i use are:

from app import app, db
from flask_migrate import Migrate, MigrateCommand
from flask_script import Manager

and it cant find the module app

[rapid sparrow]

more of a code?

show us at least docker file

[heavy knot]

version: '3.7'

services:

  flask_app:
    container_name: 'flask_app'
    build: '.'
    environment:
      - FLASK_APP=app.py
      - SQLALCHEMY_DATABASE_URI=sqlite:///database.sqlite
    ports:
      - "8001:5000"
    volumes:
      - .:/Flask
    depends_on:
      - redis

  simple_worker:
    image: 'python:3.8-alpine'
    container_name: 'worker_flask'
    build: '.'
    volumes:
      - flaskapp:/var/lib/flask/data
    environment:
      - SQLALCHEMY_DATABASE_URI=sqlite:///database.sqlite
    depends_on:
      - redis
      - flask_app
    command: rqworker --url redis://redis:6379/0

  redis:
    container_name: 'redis_flask'
    image: 'redis:4-alpine'
    ports:
      - '6379:6379'

  db:
    image: mysql:5.7.22
    restart: always
    environment:
      MYSQL_DATABASE: main
      MYSQL_USER: root
      MYSQL_PASSWORD: root
      MYSQL_ROOT_PASSWORD: root
    volumes:
      - .dbdata:/var/lib/mysql
    ports:
      - 33067:3306

volumes:
  flaskapp:

FROM python:3.8-slim

# layer caching for faster builds
COPY requirements.txt /
RUN pip install -r /requirements.txt

#COPY app.py /app.py
COPY . /
WORKDIR /

ENV FLASK_ENV=development

CMD flask run --host=0.0.0.0

#CMD gunicorn --workers $WORKERS \
#  --threads $THREADS \
#  --bind 0.0.0.0:$PORT_APP \
#  --log-level DEBUG \
#  app:app

from app import app, db
from flask_migrate import Migrate, MigrateCommand
from flask_script import Manager

migrate = Migrate(app, db)

manager = Manager(app)
manager.add_command('db', MigrateCommand)

if __name__ == '__main__':
    manager.run()

[rapid sparrow]

a moment

[heavy knot]

okay

this is a better view of the structure

[rapid sparrow]

FROM python:3.8-slim

COPY requirements.txt /
RUN pip install -r /requirements.txt

COPY . /
WORKDIR /

ENV FLASK_ENV=development

CMD flask run --host=0.0.0.0

this should be fixed for sure... for...

FROM python:3.8-slim

ENV PYTHONUNBUFFERED 1
ENV PYTHONDONTWRITEBYTECODE 1

COPY ./requirements.txt ./
RUN pip install -r requirements.txt

# Create directories app_home and static directories
ENV HOME=/app
ENV APP_HOME=/app/web
RUN mkdir $HOME
RUN mkdir $APP_HOME
WORKDIR $APP_HOME

COPY . .

RUN chmod 777 $APP_HOME
ENV FLASK_ENV=development

CMD flask run --host=0.0.0.0

That would work better.

the thing is / address is filled with system things

it is already full

not the best place to run application

[heavy knot]

oohh okay thanks! im gonna change it to that

btw when i try to remove the Flask directory its saying there are still things running in there, do you think what could cause that? afaik there cant be anything running in there and it shouldnt even exist since i removed the directory

[rapid sparrow]

make .dockerignore file btw

and include into it ignoring everything your .gitignore has

venv especially

[heavy knot]

okay thanks ill do that

[rapid sparrow]

docker-compose up --build

will cause rebuild and run

or actually

docker-compose build

😉

[heavy knot]

im gonna try it!

[rapid sparrow]

the last and the most important advice

when you call docker ps

from another terminal, you can see your conainers running right?

[heavy knot]

yeah

[rapid sparrow]

docker exec -it <containerID> <linux command> will run command inside the container

so

docker exec -it <containerID> ls /
or docker exec -it <containerID> ls /app/web

and you will see stuff copied to cintainer

or even...

docker exec -it <containerID> apt install mc

docker exec -it <containerID> mc

and navigate inside the container with Midnight Commander)

[heavy knot]

nicee im gonna try it out

pip install takes a while to do since it includes all anaconda installs also

[heavy knot]

this one isnt working its giving this error:

docker exec -it f7bf3a0f7f7b apt install mc
Reading package lists... Done
Building dependency tree
Reading state information... Done
Package mc is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

E: Package 'mc' has no installation candidate

[rapid sparrow]

apt get update

before trying that

[heavy knot]

im getting this error when trying it:

docker exec -it f7bf3a0f7f7b apt get update
E: Invalid operation get

[rapid sparrow]

ops..

apt-get update

or apt update

[heavy knot]

nice its working!

[rapid sparrow]

as alternative

you could actually connect with ssh to your container)

[heavy knot]

nice ill also try that

btw

do you know why Flask is still there and i cant remove it?

[rapid sparrow]

take down flask_app conainer I guess

docker ps

docker ps -a

[heavy knot]

thanks it worked!

im still getting that weird error btw that the imports arent working:

root@968472cb319f:~/web/app# python manager.py db --help
Traceback (most recent call last):
  File "manager.py", line 1, in <module>
    from app import app, db
ModuleNotFoundError: No module named 'app'

its weird since the manager.py is in the app directory

[tawdry needle]

@rapid sparrow why PYTHONUNBUFFERED?

[rapid sparrow]

my memory says that docker-compose running application swallows your stdout
PYTHONUNBUFFERED helps to return it back

[tawdry needle]

i haven't had any issues with this, i've been using python and docker compose at work extensively for months

maybe it's an old bug or something

[rapid sparrow]

https://stackoverflow.com/questions/51362213/docker-compose-not-printing-stdout-in-python-app
people have issues like that in stack overflow

perhaps at least some docker-compose versions do that

[tawdry needle]

interesting, it could have to do as much with the OS inside the container as with docker-compose

it looks more like they just managed to write a buggy application and it happens to be in docker-compose

why this question was not closed as unclear i have no idea

no minimal example, screenshot of some random output, not even a snippet of a dockerfile

[rapid sparrow]

By default, Python buffers output to sys.stdout.
PYTHONUNBUFFERED: 1 This will cause all output to stdout to be flushed immediately.

perhaps there is a differnece between sys.stdout and stdout

and docker-compose shows only stdout

[heavy knot]

does someone know a fix for this? might be because i dont use full imports like: from app/web/app import app, db, but it isnt an ideal way of doing it

[tawdry needle]

sys.stdout is how python represents stdout

[tawdry needle]

you have ./app/app./py and ./manager.py?

python packages are looked up relative to the current working directory, not relative to the script being executed

[heavy knot]

i have a folder app with an init file inside it and a manager.py file inside it and i do the imports like this in the manager.py file: from app import app, db

but it says that it cant find the module app

[red grove]

Spent the whole day working on ansible and I think I'm finally close to a once command deploy of our ubuntu/django/postgres/nginx stack install

Just need to sort out redis

[tawdry needle]

what is your current working directory when running this? and how are you running it?

if you are in my-stuff/app and you run python ./manager.py then you need to cd up one level. you should be in my-stuff and you should run python -m app.manager

[heavy knot]

my current working directory is the directory where the file is in, i run it in docker, first i cd to the directory then i run: python manager.py db --help

[tawdry needle]

my-app/          <--- WORKDIR should be here
  README.txt
  app/           <--- WORKDIR should not be here
    __init__.py
    app.py
    db.py
    manager.py

and you should run python -m app.manager db --help

[heavy knot]

thanks! im gonna try it

[tawdry needle]

https://docs.python.org/3/tutorial/modules.html#packages
https://realpython.com/python-modules-packages/
@heavy knot some reading for you

[heavy knot]

thanks! ill read it

[heavy knot]

/usr/local/bin/python: No module named app.manager``` im getting this now, i also tried app.manager.py

i already found the error, its working

thanks!

[tawdry needle]

👍

[muted cobalt]

Is it possible for us to cache our dependencies in docker build?

[tawdry needle]

i've been looking into this, apparently you can but i haven't tried to do it myself yet https://stackoverflow.com/q/58018300/2954547

[muted cobalt]

I actually just found out how to do it myself, even using Pipenv.

FROM python:3.8-slim

ENV PYTHONDONTWRITEBYTECODE 1
ENV PYTHONFAULTHANDLER 1

# Let service stop gracefully
STOPSIGNAL SIGQUIT

# Copy project files into working directory
WORKDIR /app

RUN apt-get update && apt-get install gcc -y

COPY Pipfile Pipfile
COPY Pipfile.lock Pipfile.lock

RUN pip install pipenv
RUN pipenv install --deploy --system

ADD . /app

# Run the API.
CMD python launch.py runserver --initdb

If you only copy in the pipfiles before installing dependencies it will cache

Since no files in your working dir have been changed

[tawdry needle]

that's not quite the same as caching, but if it works for you then it works

[short peak]

what about having a docker image with the "cached" stuff?

then use that "cached" image as base

[muted cobalt]

Well it does cache the dependencies

[rancid schooner]

Hey @muted cobalt!

It looks like you tried to attach file type(s) that we do not allow (.yml). We currently allow the following file types: .gif, .jpg, .jpeg, .mov, .mp4, .mpg, .png, .mp3, .wav, .ogg, .webm, .webp, .flac, .m4a.

Feel free to ask in #community-meta if you think this is a mistake.

[muted cobalt]

I've basically copied a old version of (python-discord/bot) lint-and-test workflow.
However I'm getting ~120k lines of flake8 errors. Locally there are no flake8 errors...

https://paste.pythondiscord.com/iwegarumok.http

[tawdry needle]

I might have misunderstood, I had in mind caching Pip downloads across full rebuilds

Using build layers is totally reasonable and normal

You could probably even use a multi-stage build where the artifact is the venv that will run the app

[muted cobalt]

Could, but this is already 60 times faster than our last Dockerfile 😂

(not a joke)

[tawdry needle]

Yikes

[muted cobalt]

Just gotta fix this now and I'm good 😄

[tardy ginkgo]

Hey yall, I got a PyInstaller question:
Is there a way to create an executable/.app that runs my python script with command line arguments?
i.e. produce a .app that emulates running my script with added command line arguments. (the user wont input arguments, they will simply run the .app and my script will be run with some predefined arguments)

[muted cobalt]

just code the CLI arguments into the script before you compile it?

[tardy ginkgo]

that's kind of what im trying to do without making redundant files.
its to compile for different build targets. code checks the "target" variable and some features become enabled/disabled. id love to have a way to set that "target" variable in my .spec file that i pass to PyInstaller

or maybe i can ask my question this way:
is there a way for PyInstaller to supply commandline arguments (even an environment variable would do) to the script thats being compiled?
if so, I can have one .spec file per build target and that would make me very happy rn

[rapid sparrow]

You are probably just looking how to use makefiles
Or if more python oriented... one of its cli interface building libraries, I recommend pip install click

Just launch your pyinstaller with command from it

os.system("your pyinstaller command")

[tardy ginkgo]

python oriented makefile, yes!
i have not heard of this module, i'll look into it. thank you!

[rapid sparrow]

python scripts.py build -t linux

For example for your interface

Build target could be having default value

[tardy ginkgo]

python scripts.py build -t linux
i dont think i quite get this. does this solution not use PyInstaller?

[rapid sparrow]

It is just a tool to create console user interface

What would be used inside is completely your choice

[rapid sparrow]

On this command it can activate pyinstaller commands

[tardy ginkgo]

but would pyinstaller create an executable that running it would be the equivalent of running a script on the command line with arguments?
to reiterate, im trying to produce an executable that itself will not take an argument, rather the executable will be the executable of script.py arg1 if you will

so that i can make
target1: pyinstaller executable with arg1 target2: pyinstaller exec with arg2 ...

[rapid sparrow]

Yes, that exactly what you would do

Actually you just wrote simple makefile :)

[tardy ginkgo]

i see. lemme bang my head on it for a bit

i have the makefile, i use make

but i need to compile the same code with just one variable be different at compile time

that i couldnt achieve via pyinstaller

[rapid sparrow]

Click library is the same as, makefile, it just allows you to write makefile in python

With more comfortable features

[tardy ginkgo]

gotchu. thanks!

[muted cobalt]

I've basically copied a old version of (python-discord/bot) lint-and-test workflow.
However I'm getting ~120k lines of flake8 errors. Locally there are no flake8 errors...

https://paste.pythondiscord.com/iwegarumok.http

[rapid sparrow]

may be you forgot to copy flake8 config file which you probably had?

if such thing exists, it was probably in your root folder

[muted cobalt]

it wouldnt have been there

only config val we change is max line length

[rapid sparrow]

actually

[muted cobalt]

and it’s in the tox.ini file

[rapid sparrow]

may be you have broken \n\r new line in your workflow

instead of running
run: "python -m flake8
--format='::error file=%(path)s,line=%(row)d,col=%(col)d::
[flake8] %(code)s: %(text)s'"

it runs:
run: python -m flake8

check what it is launching)

[muted cobalt]

it’s legit copied from their workflow though?

except python -m

[rapid sparrow]

shrugs I just consider all potential choices

[muted cobalt]

yeah

I’ll try later

[rapid sparrow]

you said that

it runs normal locally

but not running well in workflow

perhaps it has problems with some dependencies?

in how they are installed in workflow

actually...

just make git clone of your repository to another folder

and try launching flake8 from zero

perhaps some .gitignore issue

[urban pecan]

just letting you know, i released a 0.2 version of justuse 🙂 now it's also possible to decorate functions/methods/classes etc. matching a name-pattern on import

on the auto-install front i'm a bit stuck with the actual "install" part, downloading and checking works though

oh, and you can pass in optional defaults now, that should be returned if something fails during import

so instead of

try:
  import foo
except ImportError:
foo = None

you can do

foo = use("foo", default=None)

and aspectizing (remote-decorating) works like
foo = use("foo", aspectize={(use.functions, ""): log})

which would decorate all functions in the module with your "log" decorator of choice

use.isfunction*

(an alias of inspect.isfunction)

[inner pollen]

Do you guys know any library which helps me to find a line of code or a log from all the imported modules in a particular module of python. Do you know how to use grep command to do so.

[supple venture]

https://github.com/Piston-CLI/piston-cli/runs/2801084957?check_suite_focus=true I am getting this error on a git workflow, not able to understand what that means

[heavy knot]

your main branch is protected - that means you can not directly push to it

[supple venture]

but I supply it a token with this permissions

[heavy knot]

Maybe this doc will help: https://github.com/stefanzweifel/git-auto-commit-action#push-to-protected-branches

@supple venture To me it seems that you maybe need to allow admins to bypass this rule: "remote: error: Required status check "Lint" is expected. At least 1 approving review is required by reviewers with write access."

[supple venture]

But I haven’t enabled this, so admins have basically no restrictions

[supple venture]

I added myself to allow bypass let’s see if it works

[cloud elbow]

I have two postgres versions

• 9.6 (my data is in this version)
• 13 (empty | new)

how do upgrade the databases from 9.6 to 13 and move it to version 13?

[tawdry needle]

@cloud elbow
https://www.postgresql.org/docs/current/upgrading.html

For major releases of PostgreSQL, the internal data storage format is subject to change, thus complicating upgrades. The traditional method for moving data to a new major version is to dump and reload the database, though this can be slow. A faster method is pg_upgrade.
I recommend reading this entire page carefully

[cloud elbow]

I don't have the bin directory wtf

(using saltstack)

[tawdry needle]

they might be installed somewhere else

it's probably os-dependent

maybe /usr/bin

it should already be on your path

[muted cobalt]

I've basically copied a old version of (python-discord/bot) lint-and-test workflow.
However I'm getting ~120k lines of flake8 errors. Locally there are no flake8 errors...

https://paste.pythondiscord.com/iwegarumok.http

Looking at this it seems like it's definitely an issue with the caching

[tawny temple]

You need to ensure the .cache folder is excluded in your flake8 config

@muted cobalt

[muted cobalt]

Ohhhh

That makes sense

[carmine maple]

hello folks, I have a (private) python repo on github and I'd like to know whether is it possible to set up something like

1. pyinstaller creates the .exe
2. the exe is zipped alongside a .txt file
3. the result .zip is marked as a release
   (please ping me when replying)

[tawny temple]

Yes, you could do it with GitHub actions

[carmine maple]

looking up what that is, thank you

[rapid sparrow]

+1 sounds like definitely possible
https://github.com/marketplace/actions/automatic-releases
It looks like it is related to it

hmm, why is it having a dog with man as picture? 🤔

[carmine maple]

oh this also looks great, thanks!

[rapid sparrow]

https://jacobtomlinson.dev/posts/2021/automating-releases-of-python-packages-with-github-actions/

with some googling it is possible to find for python

this one auto publishes to PyPi

nice

[carmine maple]

I did some googling but didn't find what I was searching for

[rapid sparrow]

I am googling like... "github actions <insert my combination of words>"
usually adding "release" word
the last combination I tried "github python auto release actions" and previous one as "Github actions auto making release"

[sharp plank]

so this code is in my_file.py

hkdf = cy.hazmat.primitives.kdf.hkdf.HKDF(
    algorithm=cy.hazmat.primitives.hashes.SHA256(),
    length=32,
    salt=None,
    info=None,
    backend=cy.hazmat.backends.default_backend()
  )```
but get an error: 
```AttributeError: module 'cryptography' has no attribute 'hazmat'```
 but the same code in terminal interpreter works just fine

[primal basin]

im stuck in a terraform init loop; im just learning TF, and trying to set up a remote backend. I deleted a S3 bucket on the console and now my terraform CLI keeps me in this loop about the deleted S3 bucket...

ive tried terraform init -reconfigure and still blabbing about the deleted S3 bucket.. any way I can do a fresh restart?

│ 
│ Reason: Unsetting the previously set backend "s3"``` is the error msg im getting

so then i run terraform init -reconfigure

tehn i get Terraform has been successfully initialized!

but upon terraform apply it shows an error about the deleted S3 bucket

[carmine maple]

I'm having troubles trying to get git to work with vs code
I have a private repo. I generated the ssh key and linked it to my github account. When I clone using git bash, it downloads just fine, but cloning through vs code just hangs forever

[tawny temple]

I'm having trouble with poetry on Windows through the git shell. It's telling me that I need elevated permissions:

poetry run pre-commit install

  OSError

  [WinError 740] The requested operation requires elevation

  at c:\users\mark\scoop\apps\python\3.9.5\lib\subprocess.py:1420 in _execute_child
      1416│             sys.audit("subprocess.Popen", executable, args, cwd, env)
      1417│
      1418│             # Start the process
      1419│             try:
    → 1420│                 hp, ht, pid, tid = _winapi.CreateProcess(executable, args,
      1421│                                          # no special security
      1422│                                          None, None,
      1423│                                          int(not close_fds),
      1424│                                          creationflags,

Works fine through PowerShell. I tried looking up this issue for poetry but didn't find anything. Has anyone on Windows encountered this before?

Perhaps related, when I try to run my pre-commit hooks, one of them fails saying that node.js needs elevated permissions. Again, it works fine through PowerShell.

[shy locust]

Might be something about your environment variables

Confirm that you can clone or fetch through bash terminal, then in that terminal , type code and hit enter to open up vs code with a Child of that bash environment

Then try git fetching or something through vs code

If it works, there you go

[carmine maple]

Thank you! When I tried git clone through the terminal this time, I was prompted to authenticate Github through the browser for git's cred manager, and after doing that, it now works through VS Code as well

[shy locust]

Noice

[supple venture]

this is not the right channel to post projects, you can post them on the reddit, thanks!

[supple venture]

You can, open the git console, to see the errors/commands vscode is running behind the back, it makes debugging it easier

[inner pollen]

Sure

[carmine maple]

I was so close. When I commit through VS code, it appears that all changes commit, but opening the repository on the website shows nothing went through and changed

[shy locust]

Look up creating a bash profile with credentials and how to point your bash terminal at the credential file

My best guess is that the hanging is related to an issue with your http proxy or https proxy setting, and the other half is that you don't have your github username and password saved in your vsc bash environment

If you got a 503 during that first vsc hang you need to set a proxy

[carmine maple]

it now downloads the repo correctly but nothing happens when I commit

[shy locust]

Anyone can clone your repo (since its probably public) without logging in, so you need to pass your username and password to your git request via the profile when attempting to commit

[carmine maple]

it is private, the cloning didn't work earlier but now it does

committing doesn't work though

[tawdry needle]

Did you forget to push to a remote after committing?

[shy locust]

Waitaminit

[carmine maple]

I set it up on another computer and now it somehow works

thank you all who helped! 😄

[supple venture]

Need some help with github permissions, the token has literally every permission it can have.

https://github.com/Piston-CLI/piston-cli/actions/runs/932809078

[urban elbow]

Unexpected input(s) 'token', valid inputs are ['install_url', 'nix_path', 'extra_nix_config'] I don't really know much about github actions, but that seems to be saying that token isn't a valid option, (i.e. it doesn't want a token at all), rather than that the token is invalid.

[supple venture]

Not sure, but I think its the problem on the github runner side

actions/runner#514

[signal cloak]

GitHub

Unwanted warning for undefined input variable in action.yml

[urban elbow]

ah, yeah that's probably right, didn't realise that was a warning not an error. I'm not too sure then

[supple venture]

No worries , I am using a token which I use to push myself (same permissions, different token) and I never got that error, so even I was suprised when I first got it. I tried to mess around with repo perms and token perms, didn't help.

[brisk lake]

any one know if its possible to set env variables in pyproject.toml

eg, i have a flask app and i want to have pyproject.toml recognize the entrypoint without me setting it with export

[tawdry needle]

no, pyproject.toml by default does not have any way to set environment variables in general. you could theoretically write a program that reads env vars from it, but in the case of flask it wouldn't be easy; you'd have to invoke a different script which would then invoke flask, and at that point you might as well just use a Makefile or a shell script

i recommend using direnv https://direnv.net/ to automatically set environment variables in your dev environment

[brisk lake]

neat, will take a look

[hollow summit]

is anyone familiar with microsoft azure

[supple venture]

go ahead and ask!

[hollow summit]

actually i wanted to create alerts for the vm

Create an alert for the server down notification and check whether it is
implemented successfully?

[opal ibex]

is there a way that i can use selenium to select a drop down list and then click one of the items from that list?
https://www.technomarket.bg/search?query=remington this is the site and i want to click the drop down list as shown in the picture, I am typing using the xpath to do this:

elemental= driver.find_element_by_xpath("/html/body/tm-root/main/tm-search/div/tm-product-filter/div/mat-drawer-container/mat-drawer-content/div[2]/div[1]/div[2]/tm-product-filter-facet[2]/div/mat-form-field/div/div[1]/div/mat-select/div/div[2]")
elemental.click()```

(i am copying the full xpath from the page of the site) but i get an error that the element is not found
any idea why?

[hollow summit]

1)Create two users (user01 and user02) in Azure Active Directory.
2)Make sure that both the users can login to their accounts through Azure Portal.

how do i do the second step

[frank stone]

I want to switch my repo from GitHub to GitLab. I already changed git config --global user.name and email and added an origin with the new repo link. How do I give it my gitlab password now?

> git push origin-gitlab main
[e-mail]: Permission denied (publickey,keyboard-interactive).
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

[blazing badger]

has anyone used selenium python to automate angular apps?

[tawdry needle]

@brisk lake https://flask.palletsprojects.com/en/2.0.x/cli/#environment-variables-from-dotenv

[tawny temple]

Do you want to authenticate with an SSH key or with HTTPS?

[short peak]

I have but it's kinda painful in my opinion. Enjoyed more doing the same with cypress. Just my opinion

[frank stone]

Right now I'm trying to do with SSH, since the other stores the password as plaintext

[tawny temple]

Have you given gitlab your ssh key?

You have to do it on their website

[frank stone]

I was able to use gitlab option to import the repo from github, and then cloned from gitlab to a new folder. Now I'm trying to save the password/key so I don't have to type it everytime

I created a pair of keys and added the pub key to gitlab. ssh-add ~/.ssh/id_rsa

[tawny temple]

Does ssh -T [e-mail] work?

Well gitlab.com in this case

[frank stone]

I tried that with my ssh link and it says Could not resolve hostname .. Name or service not known

now I tried ssh -T [e-mail] and it says Welcome

[tawny temple]

Yeah that is what I meant to use

[frank stone]

it still asks for username and password when I type git push

[tawny temple]

Did you use the ssh link for the remote?

type git remote -v

It will list your remotes

[frank stone]

yeah I need to change to ssh

yes now it's working, not asking for anything

thank you

[tawny temple]

You're welcome

[blazing badger]

looked at cypress this wont work on my use case since the business wont spend 😂

[steady canopy]

Does it make sense to dockerize a python script that needs to be portable for multiple users? Right now my setup instructions are just to clone a repo and then set up the virtualenv and install packages etc., but one person so far has had issues with python versions on their machine being all messed up. Granted that's probably a user error on their part, but I'm trying to minimize work for people to have to do in order to be able to run the script. I don't have a lot of creating Docker images experience, so my look at it makes me wonder why it would be worth it since then to run the script (from what I know) they'd have to exec into the container, no? Or is there a better way and the right thing to do is to make an image for the script?

[tawny temple]

I don't know what your target audience is, but if someone can't manage Python versions on their machine properly then trying to use Docker is probably going to be over their heads too.

If you create your image with an entrypoint defined, they would just have to start the container and then stop it when they are done. They don't need to exec into it.

You're trying to minimise work for users, but keep in mind that if you use Docker, then that's just another thing they'd have to install (unless they wanna go the manual route like your original instructions).

I suppose you could offer Docker as an alternative.

[steady canopy]

everyone that's using it should already have docker installed

it was my manager that's python versions were messed up, he doesn't really do dev work anymore and didn't feel like fixing it when i thappened I guess

so it just kind of begged the question if it's portable enough I guess

[short peak]

it's open source/free to use - https://github.com/cypress-io/cypress
they offer additional features that you don't necessarily need/want/afford/etc

[tawny temple]

Well okay, you can make a Docker image for it. They'd just need to run one command to start the container.

[steady canopy]

but if the script takes arguments would those arguments be passed as run arguments then?

[tawny temple]

Yes, you can pass them to docker run

Though I think if you pass arguments you have to specify the command to run

[steady canopy]

gotcha.. will keep as a consideration

seems kinda like unnecessary work to me, would prefer to have a jump host we have tooling on rather than relying on personal machines

[frank stone]

When git push is run, which remote does it use by default?

[heavy knot]

Permission denied <repo> to username

I made a new github account and tried to git push origin master it gave me this problem, how to solve?

Please ping me if you know the help.

[tawny temple]

Origin by default, but if you checkout a branch from a remote, then that branch will remember which remote it came from and git push will respect that.

That can also be changed with git branch -u ...

[supple venture]

bump

[nimble mango]

I'm working on a tool that checks your email or (emails) if they are pwned, basically it works by seeing the error code of a website and if it's error code 200 it says a certain thing and if it's error code 404 it says it's not pwned, basically, I'm getting an error 403, and I tried Proxies, a useragent, etc., and it's just giving me a syntax error with the headers for some reason, and an error 403 without the headers, but with the proxies, somene help! 🤞

[tawdry needle]

@nimble mango doesn't troy hunt's site have an api? i suspect that haveibeenpwned forbids scraping and takes measures to prevent it, hence the 403

and circumventing that would be against our server rules to discuss

https://haveibeenpwned.com/API/v2

[nimble mango]

I believe their API costs money to use

And I don't really wanna pay for an api when I can just use the website

but it was gonna be cool to have a loop just checking my emails

[urban pecan]

btw, i haven't really mentioned it (since it works pretty much flawlessly and i didn't need any help with that), but justuse can safely import code from an URL, which i imagine could be nice for deployment

i'd be curious if i could tag a claim of "unhackable infrastructure" on it, even, maybe someone is interested in testing it 🙂

i'd guess that a system under attack would rather collapse than get corrupted

[woven saffron]

Anyone here got experience with creating docs/docstrings for python+Doxygen?

I'm using Doxygen+doxypypy (because I have to) but I am having a hard time using e.g., the "Attributes" tag to define public class attributes

[paper yacht]

Any recommendations for a tool that handles processes?

[wintry iris]

So I decided to use a linter (with VSCode) again and settled with flake8. I've cleaned up my file as much as I could, but can't get rid of error no newline at end of file flake8(W292). The end of my file looks like this.

# TESTING
if __name__ == "__main__":
    app = qtw.QApplication(sys.argv)
    mw = ConversationsList()
    sys.exit(app.exec())
# end of file

Searched it up and found this article, https://www.flake8rules.com/rules/W292.html , that did not really help though. What am I doing wrong?

[tawdry needle]

what do you mean by "processes"? interacting with the operating system (killing processes, etc.)? running other programs from within python? running multiple python processes ("process parallelism")?

@wintry iris press "enter" at the end of your file and try again?

it's possible that vscode is configured to remove trailing newlines

[paper yacht]

yea i mean interacting with the operating system (killing processes, etc.)?

[wintry iris]

it worked, thank you

[tawdry needle]

https://pypi.org/project/psutil/

[finite fulcrum]

Can I get pycharm to ignore in-file annnotations and use stubs for type hints? It recognizes the stubs but the normal annotation overrides them which makes them useless

[bleak pine]

There is to many tchat for help I will put my problem here. Well i got a python question that even Stack O didnt answer ... I create a private chat secured by AES with Tkinter. It went perfectly even on Android with Pydroid on Google play. Here is my problème : I want to close a Android keyboard with python.

[rich anchor]

I want to click on the products in the basket one by one and get their information, but I can enter the first one, but not the others in Selenium.

[wintry iris]

how does one move one folder in github to another folder, preferably though the website UI?

[primal basin]

might be a dumb q but: I am learning terraform and aws, i am fiddling around making a new vpc, subnets, gateways, security groups for my redshift vpc. I eventually want to access redshift via my terminal, what should i be editing ? the security group ingress/egress?

like do i need to add an extra ingress rule to account for me eventually accessing redshift via my terminal?

if so, what would the rule be

[shy locust]

I think it's Port exposure.
Your vpc security groups should allow connections on the ports necessary if you need to access the vpc and redshift

For configuration changes, create start delete list get describe etc you use the awscli or boto3 to pass instructions, and that doesn't need any modifications to the security groups

If you want to do configuration changes "properly", also set up an assume role that your desktop client can assume via Sts and give that role the necessary permissions to redshift (and ec2 and s3 and so on)

You should still be careful about exposing your vpcs to the Internet, so look up how to do that securely

[primal basin]

Really helpful! I will try this out and good info about cli

thank you

[shy locust]

Sure. I do this for work so @me if you need any help

[shy locust]

BTW if you're doing it in python just use botocore and boto3, once you get advanced enough to automate modifications to your consumed aws services, then boto3 requests are much simpler to configure and send automatically

[rapid sparrow]

I use Sphinx for same purposes and loving it
It has this feature named as autodoc

[primal basin]

I’m making the resources on terraform iunno if that’s the same thing you’re explaining. Terraform to build the resources, and then afterwards, copy a large CSV from S3 into redshift

So I’m not really using Python/boto3 unless I’m missing a use case

i am following this with a few of my own steps: https://medium.com/cloud-tidbits/terraforming-and-connecting-to-your-aws-redshift-cluster-16f93ddd41cc

he accesses redshift through the console, i wanna access my redshift instance via my terminal

[light crow]

when you say "access redshift" you just want to run queries and get some data back?

if so, i've used Redshift's data api with decent success: https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/redshift-data.html

this can be done via aws-cli or through boto3.

[primal basin]

yup, query, make tables, load them etc

ohhhh that looks promising ^

so if I have aws-cli properly configured I can just make use this boto3 API and not have to do any security/port configuring?

[shy locust]

Ah, yes, the security port configuring is if you are making requests from outside the aws environment

Since you're making requests using the aws sdk, no need to mess around with that

[indigo wave]

Hey guys, can anyone tell me how to generate an executable that can be run on Linux and Mac OS from Windows using PyInstaller or Nuitka?

[primal basin]

Thanks for the help 👏🏽

[heavy knot]

  scheduler:
    image: 'python:3.8-alpine'
    container_name: 'scheduler'
    build: '.'
    depends_on:
      - db
    command: cd app/reddit && python -m app.urs.mainscrape

ERROR: for scheduler  Cannot start service scheduler: OCI runtime create failed: container_linux.go:367: starting container process caused: exec: "cd": exe
cutable file not found in $PATH: unknown

hey, does someone know i can make this command work?

[rapid sparrow]

which command you are using to launch it

well, anyway.

just add path in your docker file

WORKDIR /app/follder/project

then you will be able removing "cd" command

or actually...

just debug the container with docker exec command

[heavy knot]

  scheduler:
    image: 'python:3.8-alpine'
    container_name: 'scheduler'
    WORKDIR: '/app/reddit'
    build: '.'
    depends_on:
      - db
    command: python -m urs.mainscrape

docker-compose up -d --build
ERROR: The Compose file '.\docker-compose.yml' is invalid because:
Unsupported config option for services.scheduler: 'WORKDIR'

im getting this error, i think it isnt working because im using docker-compose

[heavy knot]

that doesnt work i think because the container doesnt get created

[supple venture]

WORKDIR /app/follder/project is to be put in your Dockerfile

[heavy knot]

i have 2 py files that needs to be ran in the same directory, how would that work? because ill need 2 dockerfiles in the same directory

[supple venture]

You can name your dockerfile anything and pass it in on build parameter

build:
 dockerfile: xx

[heavy knot]

oohh okay thanks! im gonna try it

[heavy knot]

i tried it like this:

services:

  flask_app:
    container_name: 'flask_app'
    build: '.'
      dockerfile: Dockerfile_flask
    environment:
      - FLASK_APP=app.py
      - SQLALCHEMY_DATABASE_URI=sqlite:///database.sqlite
    ports:
      - "8001:5000"
    depends_on:
      - redis
      - db

  scheduler:
    container_name: 'scheduler'
    build: '.'
      dockerfile: Dockerfile_scheduler
    depends_on:
      - db
    command: python -m urs.mainscrape

but im getting this error:

ERROR: yaml.parser.ParserError: while parsing a block mapping
  in ".\docker-compose.yml", line 6, column 5
expected <block end>, but found '<block mapping start>'
  in ".\docker-compose.yml", line 8, column 7

i think it has something to do with the '.' on build, but if i remove it it says it doesnt have an image or build context

[jade summit]

would this be the best channel for discussing IDEs?

[rapid sparrow]

#editors-ides

[supple venture]

I am not sure about that.

[tawdry needle]

@heavy knot this might be a yaml parsing error, try quoting your commands and such

oh yeah this is wrong

build: '.'
  dockerfile: Dockerfile_flask

that's not valid yaml

  build:
    context: '.'
    dockerfile: Dockerfile_flask
``` @heavy knot it should be this

[short peak]

I don't think quotes are needed. https://docs.docker.com/compose/compose-file/compose-file-v3/#context

[tawdry needle]

they aren't, yaml lets you omit them

but yaml also interprets things like true as a boolean, so it's good practice to defensively quote yaml imo

[short peak]

I see

[heavy knot]

i have an error

[supple venture]

oh god, yes

[heavy knot]

t = subprocess.run(["dir"])

it it give

  File "C:\Users\kushal\Documents\com.py", line 3, in <module>
    t= subprocess.run(["dir"])
  File "C:\Users\kushal\AppData\Local\Programs\Python\Python39\lib\subprocess.py", line 505, in run
    with Popen(*popenargs, **kwargs) as process:
  File "C:\Users\kushal\AppData\Local\Programs\Python\Python39\lib\subprocess.py", line 951, in __init__
    self._execute_child(args, executable, preexec_fn, close_fds,
  File "C:\Users\kushal\AppData\Local\Programs\Python\Python39\lib\subprocess.py", line 1420, in _execute_child
    hp, ht, pid, tid = _winapi.CreateProcess(executable, args,
FileNotFoundError: [WinError 2] The system cannot find the file specified
[Finished in 2.1s]

[fair hinge]

@heavy knot

subprocess.run('dir')

[heavy knot]

  File "C:\Users\kushal\Documents\com.py", line 4, in <module>
    t = subprocess.run("dir")
  File "C:\Users\kushal\AppData\Local\Programs\Python\Python39\lib\subprocess.py", line 505, in run
    with Popen(*popenargs, **kwargs) as process:
  File "C:\Users\kushal\AppData\Local\Programs\Python\Python39\lib\subprocess.py", line 951, in __init__
    self._execute_child(args, executable, preexec_fn, close_fds,
  File "C:\Users\kushal\AppData\Local\Programs\Python\Python39\lib\subprocess.py", line 1420, in _execute_child
    hp, ht, pid, tid = _winapi.CreateProcess(executable, args,
FileNotFoundError: [WinError 2] The system cannot find the file specified
[Finished in 1.8s]```

no it also give error

I need invite tracker

I have 30%

Need join message cfg

And joinmessage cfg

And leave too

[tawdry needle]

@heavy knot dir might not be a binary, it might be a shell built-in. try subprocess.run('dir', shell=True)

[heavy knot]

it running

thank @tawdry needle

[tulip meteor]

Hi I have a Project repository which has 2 projects frond-end and back-end

Projcet

How can I deploy on Heroku?

backend is a flask api, front end is a reactjs app

[astral tangle]

anyone here deploying to Azure? I'm confused as to how my requirements.txt libs are supposed to be deployed

either by the azure devops build pipeline zipping them up, or the azure web deploy deploying them?!

[short peak]

https://devcenter.heroku.com/articles/getting-started-with-python

[inner hare]

Got a somewhat non-Python question for you guys:

I am working on a semi-private project that I would like to remain visible to the open Internet.
However, I would also like to make sure that forks, pulls and clones can only be done by a group of authenticated users. This is to ensure that users may access the repository but can only make meaningful use of it if they agree to a set of Terms and Conditions.
Has anyone ever tried something like this? Any thoughts on how to do this in GitHub or GitLab?

[urban pecan]

you can't feasibly restrict people from illegally cloning a public repo, but you can do so by license

any law-abiding person needs to adhere to the given license, so you could simply state your conditions in there

you can't prevent people from violating it though, which lawyers are for

(even if one clicks on "i agree to the license conditions" it could be a lie)

[tawdry needle]

maybe gitlab has support for this. alternatively, you can keep the git repository private but post source tarballs for every release

[urban pecan]

@tawdry needle but that doesn't hinder someone with access to copy it locally and then fork etc. whatever they like..

[fast kayak]

Aye, has anyone managed to make sense of Tox's platform conditionals? I'd like to have a single testenv execute one of two commands based on the OS, but I'm quite lost.

[stuck citrus]

hello, do you know any good library to modify pdf and docx documents

[floral smelt]

Hello, I wanted to know if anybody tried a nvidia container with a different version of CUDA

[shy locust]

I need to practise navigating around Linux because I'm taking over my company's machine image release pipeline shortly, what would you guys recommend for practise on my own time

I'm on windows
Docker with Linux or Linux partition

[rapid sparrow]

Install for yourself ubuntu 20.04 desktop, it is beginner friendly

Quite comfortable for devops stuff

[snow sage]

Hey, could someone give me an explanation on what git rebase does?

I've tried reading docs but I'm still a little confused

[finite fulcrum]

It lets you change the base commit, for example if your main branch has commits a b c and you have a feature branch that's based on a, then rebasing would let you change the base commit to c (you get all the changes from main without a merge) and then it copies the existing commits on top of it.
Or with an interactive rebase you can manipulate the history manually, you can change commit messages, re-order, squash them etc.

[snow sage]

Oh I see, thank you!

[shy locust]

Cheers man

[mortal cosmos]

Hello All, I thought I would share another tool you can use to visualize your data, let me know what you think
https://github.com/codemation/easycharts

[heavy knot]

hey

[burnt ibex]

Everwhere on the Nginx site, it takes about free trials

Is there not a free tier of some sort?

[short peak]

what?

[tawdry needle]

Nginx is free and open source, but there is a paid enterprise version with extra features

[burnt ibex]

Got it

Where are the docs for the free part?

[supple venture]

https://www.nginx.com/resources/wiki/start/index.html

[burnt ibex]

Ty

[warm pollen]

@heavy knot what is the full traceback? It seems like something is trying to resolve a cache file at ~/.cache.json

alright, the error is that the home folder doesn't exist

Yeah

You can either create it first and use another location

!d pathlib.Path.mkdir

[rancid schooner]

pathlib.Path.mkdir

Path.mkdir(mode=0o777, parents=False, exist_ok=False)```
Create a new directory at this given path. If *mode* is given, it is combined with the process’ `umask` value to determine the file mode and access flags. If the path already exists, [`FileExistsError`](https://docs.python.org/3/library/exceptions.html#FileExistsError "FileExistsError") is raised.

If *parents* is true, any missing parents of this path are created as needed; they are created with the default permissions without taking *mode* into account (mimicking the POSIX `mkdir -p` command).

If *parents* is false (the default), a missing parent raises [`FileNotFoundError`](https://docs.python.org/3/library/exceptions.html#FileNotFoundError "FileNotFoundError").

If *exist\_ok* is false (the default), [`FileExistsError`](https://docs.python.org/3/library/exceptions.html#FileExistsError "FileExistsError") is raised if the target directory already exists.

[warm pollen]

Yup

Don't forget to set parent=True

Interesting

How is FOURIER_CACHE defined?

[warm pollen]

lol I'm here now

You need to do FOURIER_DIR.mkdir(parents=True, exist_ok=True) @heavy knot

[wooden ibex]

Why are you putting everything /?

that's not best practice

[calm tapir]

Anyone know what this means? I am able to ssh just fine through terminal. But in VSCode I keep running into this.

[eager cedar]

bash command not found, do you have /bin/bash on your machine? 🙂

[calm tapir]

On my local machine or the remote machine?

@eager cedar on my local machine I'm using zsh but the remote machine I'm connecting to uses bash

[eager cedar]

I don't use vscode, so I have 0 clue about the implementation and use of that feature

but seems to me like one should check if which bash on both machines I guess.

[calm tapir]

@eager cedar What editor do you use to connect to a remote machine?

[eager cedar]

why would I use an editor for that?

[calm tapir]

@eager cedar It's tough to code in vi or a command line editor

[eager cedar]

Do you absolutely need to code on the remote machine?

[calm tapir]

@eager cedar Yeah unfortunately, I can't really replicate the environment right now because the remote machine is a raspberry pi.

[eager cedar]

I use vim so I don't have that problem, but one should only have to code on a remote machine in very rare occasions

alright, I would just use a container technology if you have environment differences.

What kind of project are you working on? Maybe I could be more specific in my suggestions then.

@calm tapir

[calm tapir]

@eager cedar Yeah might just create a docker image tbh. It's basically a robot that runs on a raspberry pi. And I need to run a Python script on it. I could just write the code locally and keep copying and pasting it into the raspberry pi for debugging but that is a painful process each time.

[eager cedar]

you could also just rsync automatically on save.

[calm tapir]

@eager cedar Yeah that is definitely an option. I was also thinking of using git but I don't want silly commits on my account.

[heavy knot]

I need help with selenium

[regal venture]

Private repos are free on github if that’s an option

[supple venture]

Sup, need some help I am working on a project with starlette, a fairly simple one, have it private sourced on github right now as it is for school, I am using a similar setup like pixels, but I get this error: https://paste.pythondiscord.com/yadatiliso.rb

Dockerfile

docker-compose.yml

[rapid sparrow]

table admins does not exist

migrate db, or something like that

or create db and tables 😉

[supple venture]

I do that alembic upgrade head in CMD

[hexed flare]

docker compose create volumes as root
cannot make changes to the mounted files made from shell in docker container as a normal user

[undone moss]

Hey! I have my code and wanna add that to my git hub. I created new repository (default) and following instructions got this

I need copy this file to git folder 🤔

[undone moss]

Solved!

[tawdry needle]

change the USER in the dockerfile maybe? or is it because docker itself was run with sudo?

[eager cedar]

You can make a user with the same UID as the host user (default is 1000)

RUN addgroup --gid 1000 test_group
RUN adduser --uid 1000 \
            --gid 1000 \
            --home /home/test_user \
            --quiet \
            --disabled-password \
            --ingroup test_group \
            test_user

[tawdry needle]

@eager cedar docker doesn't "translate" UIDs at all?

[eager cedar]

Define translate in this context

but think about it this way, the container doesn't know anything about your hosts users, so how would it even begin to "translate" UID's to something that doesn't conflict with UID's on the host?

Not sure if that's what you are reffering to :-)

[tawdry needle]

suppose my UID in the container is 1234 and the docker daemon is running on the host as 1111. if a program inside the container writes a file to a volume, will the UID of the file be 1234 or 1111 or something else?

[eager cedar]

1234

I am pretty certain, but let me check

seems like it's just owned by root

I tried to make a docker compose example.

FROM ubuntu:latest

RUN addgroup --gid 1000 test_group
RUN adduser --uid 1000 \
            --gid 1000 \
            --home /home/test_user \
            --quiet \
            --disabled-password \
            test_user
USER test_user

CMD ["tail", "-f", "/dev/null"]

docker build -t meme .

docker-compose.yml

version: "3.8"

services:
  ubuntu:
    image: meme
    container_name: ubuntu
    command: tail -f /dev/null
    volumes:
      - ./kek:/tmp/kek

@tawdry needle

Produces:
drwxr-xr-x 2 root root 40 Jun 21 20:55 kek

because I ran the the docker command as root I presume.

[tawdry needle]

ok that's what i was wondering

[hoary wharf]

Does anyone here have experience with migrating a private pypi to Artifactory?

[gleaming knot]

can anyone recommend a dependency compatibility checker resolving transitive dependencies?

[heavy knot]

Can someone help me with a tool in py

please

[urban pecan]

i have a small issue with use() i'd like to discuss.

use(use.Path(..)) can load from arbitrary relative or absolute paths, which works well because it changes the CWD for each import (and back)

it also works well in jupyter, because you can cd around, but

F:\Dropbox (Privat)\code\justuse> python .\src\use\test4.py with a relative use(use.Path()) inside will invariably fail because the envocation doesn't change the CWD

so now my predicament - do i need to look at the sys.argv to determine whether or not i need to chdir or is there some cleverer way to make the right call?

using the CWD internally is very elegant and ensures that also relative imports just work, so abandoning this approach would make things a lot more complicated overall

nvm, i think i found a solution.. __import__("__main__").__file__ should do the trick

[tawdry needle]

why do you need to change CWD?

maybe instead you can temporarily add stuff to sys.path, or maybe there's a way around sys.path entirely using the "loader" mechanism

[urban pecan]

@tawdry needle sys.path is a hack, i noticed somewhere in my deep dive into the import machinery a comment that basically said "we really should use cwd everywhere"

[tawdry needle]

Even for site packages?

[urban pecan]

and yeah, if CWD is changed correctly, it all works out perfectly

not sure about site packages, i haven't really gotten into auto-install that far yet

but yeah, if the cwd is updated recursively and reset correctly, even relative imports work in concert with use(use.Path())

at first i didn't even expect to have it both hand in hand

but now i kind of take it for granted 🙂

[burnt ibex]

I have a git branch with 7 commits, and I'd like to merge it all into the same commit (sort of like squashing a PR, but on the same branch). How would you do such a thing?

[tawny temple]

Use an interactive rebase and select squash or fixup for the commits.

git rebase -i HEAD~7

[supple venture]

Then you can squash them by adding s instead of the pick in all the commits you wanna squash down

Then it will open the commit file in your configured editor

[burnt ibex]

Thank you both so much!

[supple venture]

Np, and yeah remember to force force after commiting, git push origin my_branch --force

otherwise git would mess up your commit history

[shy locust]

Wow good to know

[rapid sparrow]

I am usually doing in this situation git push --force-with-lease, what would be the difference?

while remaining on the secondary branch where I squash the stuff

[tawny temple]

If your branch is set to track the remote then you can omit the remote and branch name from the push command

And force with lease is a variation of force that aborts if the remote has commits you don't have locally

[tawdry needle]

i.e. --force is serious danger zone material and --force-with-lease is not

[supple venture]

I didn't know about --force-with-lease, thanks!

[tawny temple]

It can still be dangerous if you're keeping your local repo up to date with git fetch (but not actually pulling those changes).

[shy locust]

What's the best way to keep your local branch up to date with the master (say, after merging a pr)

I've been doing fetch master then rebasing my dev branch

[tawny temple]

That is how I do it.

You could achieve the same thing with git pull too I suppose

[shy locust]

Git pull origin master?

[tawny temple]

git pull --rebase origin master unless you have git configured to rebase on pull by default.

[shy locust]

I had the oddest thing where one of my interns who swears she has been doing it was still 1.5k commits and several months behind the remote master somehow

Ahhhh okay. Probably doing that without having auto rebase configured

[tawny temple]

A feature branch living long enough for 1.5k commits on main is quite concerning.

[shy locust]

Right

Inspected everyone's local env and practices after that

[tawny temple]

I mean in the sense that she should be aiming to merge her changes more frequently into the main branch. But I've not worked on gigantic projects so what do I know. Maybe some stuff legitimately has to sit that long in a separate branch.

[shy locust]

Yup

[tawdry needle]

the only downside with rebasing off master (as opposed to merging master) is that conflict resolution can be extremely gnarly if your branch refactors existing code

resolving merge conflicts is one thing, resolving rebase conflicts is a whole different animal

[supple venture]

I local branch, here is it a feature branch or the main branch. I personally never do any changes to main branch so I just call git fetch upstream; git reset --hard upstream/main. And for feature branches what mark said,or the update branch button the github PR

[shy locust]

Ah yeah we don't either, everything has to be done via pr.
Gotcha on the rest, thanks guys 😊

[shy locust]

Too right yeah. If it happens I usually give up trying to resolve it, move whatever change I have out into a seperate folder, wipe my branch clean and just paste stuff back in like a barbarian

bar bar bar ME AM MAKE CODE GUD

I should put this on a bumper sticker

I, Caveman

[supple venture]

Lol, pycharm makes it simpler tho, not easy just simpler

[urban pecan]

the next big thing i'm going to tackle in justuse will be auto-installation, i guess.. then it also should be possible to install and use different versions of the same package in the same codebase 🙂

any input on how that could be achieved?

downloading should work, but i'm not sure yet how packages are actually installed

[urban pecan]

i am thinking of a json file that keeps record of all the packages that are installed somewhere on the system with their versions and hashes, does something like that already exist?

{name: [version: "path": str(Path), hash: str(sha256)]} something like that

basically the pypi metadata, just locally with a pointer to the path

that would make it possible to share packages where possible and have diverging dependencies at the same time

hmm.. then i could lightly monkey-patch the import machinery to check the dependency registry on attempted import and chdir to that path

making messing with sys.path obsolete

and people could hack the registry if they moved packages etc. while at the same time integrity would be ensured via hash

[urban pecan]

any suggestions where that registry file could be put as a default?

[tawdry needle]

$XDG_DATA_HOME/python-justuse maybe

or $XDG_CACHE_HOME if it's not "bad" to delete the data

[wooden ibex]

This seems ultra crazy

[urban pecan]

@wooden ibex how so?

[wooden ibex]

Mixing imports like that, when you are monkey patching core parts of compiler, you might want to reconsider your approach

[urban pecan]

sys.path is a hack that simply got out of hand, having a proper registry of installed packages is much less crazy, imo

and i said "light monkeypatch", it would only need to wrap the imports to chdir to where it needs to be

since the import machinery works much better if the current working directory is properly set

it wouldn't require changing the whole machinery

just a nudge to look in the right place

@tawdry needle okay, i'll start with that 🙂

@wooden ibex https://github.com/amogorkon/justuse/blob/main/README.md you'll find much more crazy stuff there 😉

but honestly, i find https://macropy3.readthedocs.io/en/latest/index.html much more crazy than any of that

[wooden ibex]

Ok?

Someone going bonkers isn’t a license for others to go bonkers

[urban pecan]

heh

besides, it's an officially supported mechanism called import hooks https://www.python.org/dev/peps/pep-0302/#id28

meta hooks is probably the thing i'll need to look into, since it's called at the beginning of the import process

[wooden ibex]

runs away screaming

[urban pecan]

lures Rabbit in with fresh carrots

[wooden ibex]

Nope

[urban pecan]

fresh meat?

[wooden ibex]

Nope

[urban pecan]

😦

[wooden ibex]

I don’t want anything near this, this isn’t DevOps, this is crazy dev

[urban pecan]

no more need for virtualenvs?

please? 🙂

[wooden ibex]

No

[gleaming knot]

can anyone recommend a licence compatibility checker resolving transitive dependencies?

[heady gull]

hey guys,
I don't know if this is the right channel, but I want help with matplotlib. If anyone knows how to change the colors of the numbers please help me thx 😄

[gleaming knot]

https://stackoverflow.com/a/4762002

[heady gull]

thx imma check out

thanks so much! it worked! @gleaming knot

[forest goblet]

personally, I don't see anything crazy about taking control of the import process long enough to ensure you get a specific module (version) loaded while not disrupting resolution for the rest of the session

[heady gull]

?

[floral smelt]

For Docker and docker-compose users, when you lauch a command with docker-compose, do you have errors saying x file could not be found despite having the directory configured?

Ex. when you launch docker exec on a Django app, with ls

You have the manage.py file in it

But when you use command: python manage.py runserver 0.0.0.0:8000 it says manage.py not found. Why is that?

[urban pecan]

@wooden ibex btw, i'm not even sure i need to fiddle with meta_path at all, no monkey patching required

less crazy dev, more devops 😉

pip insists that only ever one version of a package is installed, so i can just put the other versions in an extra dir and let pip do its stuff without interfering

if someone wants to have different versions of the same package side by side in the same code, they'd go through use(), which will just handle those version-pinned, well defined versions with no guessing

so, if someone doesn't care about the exact version and wants to go the classic way, they can do that still with no change of behaviour

[tawdry needle]

can you show the dockerfile and docker compose file?

[floral smelt]

FROM python:3.9.5-slim

ENV PYTHONUNBUFFERED=1

ENV PYTHONDONTWRITEBYTECODE=1

WORKDIR /app

RUN apt-get update &&\
    apt-get install -y binutils libproj-dev gdal-bin

COPY requirements.txt .

RUN pip install -r requirements.txt --no-cache-dir

COPY . .

and docker-compose:

version: '3.8'

services:
  db:
    image: postgis/postgis
    container_name: db-postgis
    env_file:
      - ./db/.env
    networks:
      - backend-network
    ports:
      - 5432:5432
    volumes:
      - db-data:/var/lib/postgresql/data

  backend:
    restart: on-failure
    env_file: ./backend/.env
    build: ./backend
    expose:
      - 8000
    ports:
      - 8000:8000
    volumes:
      - ./backend:/backend
    command:
      - ./startapp.sh
    container_name: django-backend
    depends_on:
      - db
    networks:
      - backend-network
      - frontend-network

  frontend:
    build:
      context: ./frontend
      dockerfile: ./Dockerfile
      args:
        environment: development
        status: stable
    expose:
      - 3000
    ports:
      - 3000:3000
    container_name: react-frontend
    env_file: ./frontend/.env.development
    volumes:
      - ./frontend:/frontend
    networks:
      - frontend-network

networks:
  frontend-network: {}
  backend-network: {}

volumes:
  db-data: {}

The Dockerfile is for the Django app

[tawdry needle]

just to be safe, i always use absolute paths on the RHS of COPY

COPY . /app

and what is in startapp.sh?

and one more question: what's the structure of files in .?

[floral smelt]

#!/bin/bash

# Wait for db
echo "Waiting for db.."
if [ "$DATABASE" = "postgis" ]
then
    echo "Waiting for postgis..."

    while ! nc -z $SQL_HOST $SQL_PORT; do
      sleep 0.1
    done

    echo "Postgis started"
fi

# Runs the server
echo "Running Django server"
python manage.py runserver 0.0.0.0:8000

[tawdry needle]

and you're saying that you can do docker-compose run backend ls /app and you see manage.py there, right?

[floral smelt]

./backend
├───api
│   └───migrations
├───control
│   └───.mypy_cache
│       └───3.9
│           ├───collections
│           ├───control
│           ├───importlib
│           ├───os
│           └───_typeshed
└───users
    └───migrations

[floral smelt]

Yeah i can use docer exec backend ls and it shows the contents

Including manage.py

I have a migrate.sh script but I can't trigger it using ./migrate.sh

[tawdry needle]

and backend is the CWD?

or . is the CWD and backend is a package?

[floral smelt]

./backend is the working directory where Django is located

.
├── Dockerfile
├── Dockerfile.prod
├── api
│   ├── __init__.py
│   ├── admin.py
│   ├── apps.py
│   ├── fields.py
│   ├── filters.py
│   ├── migrations
│   │   └── __init__.py
│   ├── models.py
│   ├── serial_fields.py
│   ├── serializers.py
│   ├── tests.py
│   ├── urls.py
│   └── views.py
├── control
│   ├── __init__.py
│   ├── asgi.py
│   ├── settings.py
│   ├── urls.py
│   └── wsgi.py
├── manage.py
├── migrate.sh
├── requirements.txt
├── startapp.prod.sh
├── startapp.sh
└── users
    ├── __init__.py
    ├── admin.py
    ├── apps.py
    ├── forms.py
    ├── managers.py
    ├── migrations
    │   └── __init__.py
    ├── models.py
    ├── tests.py
    ├── urls.py
    └── views.py

[tawdry needle]

ok, i see

...i have no idea

lol, sorry

weird

you'd expect that to work, try deleting all the containers and images then starting over from scratch 🤷‍♂️ and try using absolute paths with COPY just in case

[floral smelt]

I have no idea if it's because the control folder has a different name from backend

Normally it shouldn't be the case

Yeah it shouldn't as this repo https://github.com/testdrivenio/django-docker-traefik names its project differently

I tested them and it works

[rapid sparrow]

ask question, do not ask to ask

[urban pecan]

okay, RFC: let's say, there is ~/justuse-python with a config.ini, registry.json and usage.log and a /packages subfolder, which is used as a default, but can be arbitrarily set in the config. all installed (whether by pip or auto-installed or otherwise) packages are registered in the registry to look up during runtime. if you use() a package with a specified version, it is looked up, and if not available, downloads the zip from pypi, imports it and writes the compiled code to disk, discarding the unzipped, uncompiled sourcecode next opportunity. this should allow classical imports and pip to work in parallel with use() without interferance

could there be any problems with this setup?

[topaz aspen]

how would you control the requirement jupyter labextension install jupyterlab-plotly@4.14.3 ?

I can't add this to my requirements.txt file, so I'm not sure where that should be placed.

(ping me if you respond please)

[urban pecan]

which file permissions should i give the registry etc? i guess 777 is out of question, but what should be preferred?

[tawdry needle]

you can't really have pip managing non-python dependencies. if your project requires things other than packages from pypi, you'll need some other way to specify

[topaz aspen]

yeah I know pip can't which is why i asked that question 😬
sorry if it was unclear

[tawdry needle]

conda is one option, but even then jupyterlab itself might be in a different environment from the kernel (although in this case it probably shouldn't be)

another is just a build/setup script

[urban pecan]

hey @tawdry needle any ideas on my suggested auto-install system?

let's say you use("numpy", version="3.0.1", auto_install=True, hash_value="98f98df.."), it would download the package, install it into ~/justuse-python/packages/numpy-3.0.1.. - out of the way of pip and classical imports with the location with other metadata put into ~/justuse-python/registry.json for quick lookup during runtime

[grand gale]

Soooooo tl;dr of what i wanna do:
I wanna automate some parts of gitflow so when I push to master/a pr is merged i want those things to happen:
get the version from the last commit(s)
ie when the last commit is Merge from "release/v2.3.0" the version is v2.3.0
or when the last one isnt a merge it gets the last merge and then increments based on the feat and fix commits
use that version to create a tag
use that version and the last commits to create a gh release with the changelog generated from the commits
make a sentry release

the sentry release is easy i just need some way to provide it that info

the rest isnt so much

[urban pecan]

that way, use() will enable fully self-contained code, with no venv or external tools needed for dependency management

but it will actually make it easier for external tools to analyse and make updates, for instance an IDE could provide those

while not getting in the way of pip or import, if someone prefers those

[deep estuary]

okay so

I'd use a python script to determine if the commit matches

if it does, output a string, and then you can check that using run conditionals

and then you can also use that to output v2.3.0 or whatever

[grand gale]

yeah thought so as well, but i havent found a way yet to set variables during a run

[deep estuary]

then you can use one of these https://github.com/softprops/action-gh-release to do the actual release

uhhhhh

so

https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#jobsjob_idoutputs

from python, you need to print a string like

::set-output name=is_release::true
::set-output name=release_tag::v2.3.0

then github will pick those up, and you can use job outputs

you can have a release job and only run it if is_release is true using https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#jobsjob_idif

then inside that job you can use release_tag for sentry / gh release / etc

[grand gale]

yeah the i know the if syntax, the confusing shit is the setting vars

anyways tysm

[deep estuary]

i uhhh

hm I might have a thing internally somewhere where I use set output from a py script

https://gist.github.com/jb3/0ed9a4569006751b1bd7be28076e3607 @grand gale might have a little bit of use as reference

but you can also just set the manifests from python

[grand gale]

@deep estuary so smt like this should work?```yaml
name: Lint Kubernetes manifests

on:
push:
branches:
- master
- ltr/v*

jobs:
getvars:
name: Gets the neccecary vars for the following actions
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- name: Get the release version and changelog
run: |
version=$(python $(curl the script))
echo "version=$version" >> $GITHUB_ENV
echo "changelog=$(python $(curl the script))" >> $GITHUB_ENV
git tag -a $version
git push --tags

sentry_release:
needs: getvars
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: getsentry/action-release@v1.0.0
env:
SENTRY_AUTH_TOKEN: ${{ secrets.SENTRY_AUTH_TOKEN }}
SENTRY_ORG: sentry_org
SENTRY_PROJECT: project_name
with:
environment: production
version: project_name@${{ env.version }}

gh_release:
needs: getvars
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: softprops/action-gh-release@v1
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
with:
body: ${{ secrets.GITHUB_TOKEN }}
draft: true
tag_name: ${{ env.version }}```

[deep estuary]

I don't think it'll be in env, will it? Also you need to ensure there is a tag before pushing a tag

[grand gale]

what dont you think will be in env if you mean the version and shit theres this:
https://docs.github.com/en/actions/reference/workflow-commands-for-github-actions#setting-an-environment-variable
the tags are pushed in getvars but im overworking that

[grand gale]

name: Release Versions on Several Platforms

on:
  push:
    branches:
      - master
      - ltr/v*

jobs:
  release:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout Code
        uses: actions/checkout@v2
      - name: Get the Release Version and Changelog
        run: |
          echo "release_version=test" >> $GITHUB_ENV
          echo "changelog=test" >> $GITHUB_ENV
      - name: Push the Release to Sentry
        uses: getsentry/action-release@v1.0.0
        env:
          SENTRY_AUTH_TOKEN: ${{ secrets.SENTRY_AUTH_TOKEN }}
          SENTRY_ORG: sentry_org
          SENTRY_PROJECT: project_name
        with:
          environment: production
          version: project_name@${{ env.release_version }}
      - name: Create a Tag with the Version
        uses: actions/github-script@v3
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}          
          script: |
            await github.git.createRef({
              owner: context.repo.owner,
              repo: context.repo.repo,
              ref: "refs/tags/${{ env.release_version }}",
              sha: context.sha
            })
      - name: Create a Release Draft
        uses: softprops/action-gh-release@v1
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        with:
          body: ${{ secrets.GITHUB_TOKEN }}
          draft: true
          tag_name: ${{ env.release_version }}```

so this should work

the env shit is tested and works, the other shit not yet

[grand gale]

fucking yeah that shit works (with some edits )

[tacit flare]

Hey all - playing around with GitLab, what's the most appropriate way to roll out source code changes to my server? GL CI/CD action that SSH's to my server and pulls the most recent changes from the master/main branch?

[warm pollen]

That's a perfectly valid solution if you don't have any (too) fancy setup with Kubernetes or similar

[tacit flare]

Perfect, thanks @warm pollen

[warm pollen]

No problem!

Also don't forget to generate a new individual ssh key that only has access to the required resources, you wouldn't want someone to gain a root access on your server in case of a leak

[velvet spire]

how do you get it connected to your server?

i assume your server is publicily accessible?

[tacit flare]

Good point @velvet spire , it’s not publicly available.

[topaz aspen]

Any reason you've not mentioned docker?

[tawdry needle]

no reason, depends on your intended usage i guess

[alpine pike]

@deep estuary

[topaz aspen]

Yeah, I don't think I have much intuition for one over the other.

When searching stuff like conda Vs docker a lot of stuff comes up (on threads or whatever) simplifying conda to a requirements alternative, but conda goes further than pip doesn't it?

I guess idk where the limit of conda is, and whether docker "takes over" at that point. I'm also not sure if one should prefer conda over docker if it's possible to use it, or whether it's more subjective

[west seal]

Hi guys how are you doin??I don't know if it's the right channel but am having some problem installing mysqlclient==1.3.10 via pip3 on ubuntu 20.04 (with python3.4.3),any ideas ?Thanks

[gritty inlet]

I would like to know how could I automate the video uploading process of a local video on my pc, to Youtube channel specific playlist (based on the name of the video file being uploaded).
Example: videoSam.mp4 goes to Youtube playlist "Sam", while videoFrodo.mp4 goes to Youtube Playlist "Frodo".
Is that possible?

[grand gale]

@deep estuary if you by any chance are interested in automatic release action heres the one i use atm:```yaml
name: Release Versions on Several Platforms

on:
push:
branches:
- master

jobs:
getversion:
name: Get Version
runs-on: ubuntu-latest
outputs:
version: ${{ steps.version.outputs.version }}
steps:
- name: Checkout Code
uses: actions/checkout@v2
with:
ref: ${{ github.head_ref }}
fetch-depth: 0

  - name: Git Version
    id: version
    uses: codacy/git-version@2.2.0
    with:
      prefix: v
      dev-branch: develop

sentry:
name: Create Sentry Release
needs: getversion
runs-on: ubuntu-latest
steps:
- name: Checkout Code
uses: actions/checkout@v2
- name: Push the Release to Sentry
uses: getsentry/action-release@v1.0.0
env:
SENTRY_AUTH_TOKEN: ${{ secrets.SENTRY_AUTH_TOKEN }}
SENTRY_ORG: org-name
SENTRY_PROJECT: project-name
with:
environment: production
version: project-name@${{ needs.getversion.outputs.version }}

settag:
name: Create Git Tag from Version
needs: getversion
runs-on: ubuntu-latest
steps:
- name: Create the Tag
uses: actions/github-script@v3
with:
github-token: ${{ secrets.GITHUB_TOKEN }}
script: |
await github.git.createRef({
owner: context.repo.owner,
repo: context.repo.repo,
ref: "refs/tags/${{ needs.getversion.outputs.version }}",
sha: context.sha
})

releasedraft:
name: Create a Github Release Draft with the Changelog
needs:
- getversion
- settag
runs-on: ubuntu-latest
steps:
- name: Checkout Code
uses: actions/checkout@v2
with:
ref: ${{ github.head_ref }}
fetch-depth: 0
- name: Build Changelog
id: changelog
uses: mikepenz/release-changelog-builder-action@v1
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
- name: Create Draft
uses: softprops/action-gh-release@v1
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
with:
body: ${{ steps.changelog.outputs.changelog }}
draft: true
tag_name: ${{ needs.getversion.outputs.version }}```

[deep estuary]

nice one

the one CI/CD flow where I could use this uses Jenkins

[grand gale]

rip

[pastel meadow]

can someone - define for me - tools and devops

sorry in drunk

mmmmmmm - how about a custom interface to talk to chips on a breadboard via python

[grand gale]

@deep estuary help, im getting way too much into ci/cd

[deep estuary]

lol

[tawdry needle]

docker lets you encapsulate an entire OS, while conda "only" deals with software dependencies. docker is probably better in general for isolating the dependencies of standalone applications, and much better for deploying software. there's nothing stopping you from using conda inside the docker container!

[noble path]

should i use IoT platforms or just access a website's api itself?

[urban pecan]

if a user wants to auto-install something via justuse and they try use("stuff", auto_install=True), it should raise an exception and give proper advice, but also info on the package they are attempting to install in order to avoid malicious packages. what is important info a user should get at this point to decide on whether or not they are trying to install a malicious package?

i'm thinking number of downloads, what else?