@Anoncheg@emacs.ch avatar

Anoncheg

@Anoncheg@emacs.ch

I am looking for any IT job, just write me, I will write to you. I can relocate or work remotely.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

Anoncheg, to random
@Anoncheg@emacs.ch avatar


I have tasks from employers:

  1. Telegram chat-bot (in work) I use open source TDLib
    https://core.telegram.org/tdlib/
  2. Li-on Batteries, prediction of lifetime
  3. Documents recognition service.

I am going to develop my own Telegram and other messangers bots
for smart AI services and earn money by myself.
😶

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar

  • I fixed vertico finally with (expand-file-name) that get rid
    of //.
  • I enhanced selection in Dired. I will publish this as a package
    maybe. It is faster than marks. You just select lines and do staff.

What is conciousness? Autoencoder, what you see is what you get.

fun staff:
grep -Rin ":-(" /usr/share/emacs/29.2/lisp | grep -v smile | wc -l
>> 62
grep -Rin ":-)" /usr/share/emacs/29.2/lisp | grep -v smile | wc -l
>> 30

There is a game very closely looks like legendary Planescepe:Torment
CRPG, it is 2017 "Torment: Tides of Numenera".
I don't have time to play...
😶

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar

My Elisp package for history in Dired was added to MELPA
repository. 🎂

I added alternative implementation of history with tab-line
mode. 🤹 This implementation work perfect, fast and able to work
under root console or any terminal. ⌘

I also solved in Dired:

  • selection by region selection, not by marks ●
  • customization of opening file in other window ●

I am going to write article about configuration of Emacs as a
file manager.
🏄

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar


I answered 15 tricky questions for job at topics of DS, ML,
DL. I used ChatGPT to help me. 😈 All answers should be
correct, I will public my calculations after two week if I
wouldn't get offer. ◕‿◕

I learned VC (Version Control) package for Emacs. ⛏ VC isn't a
mode it is permanently available functionality ⚠, can be used
from file-visiting buffer, Dired buffer, VC Directory buffer.

VC offer very powerful commands: ᕙ( •̀ _ •́ )ᕗ

  • vc-next-action - universal intellectual action ✭
  • vc-annotate - for each line, show the latest revision in which
    it was modified. ⋆
  • vc-region-history - Display the history of changes made in the
    region of file visited by the current buffer. ★

I learned project management for Emacs with EDE and projectile
and I found it useless. Because Dired + Bookmarks just do the
same.
😶

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar

I posted request to MELPA repository for my new package for Dired
filemanger in Emacs. 🐾💜😊

Emacs has is very advanced filemanager (Dired) it is like
MidnightCommander and FarManager, but much better, because it
is customizable as beyond imagination.

Even Vim don't have modern fast navigation, that is why my
package is awesome. 😈

Now I don't need OS Environment, all I need is Emacs, console
and browser. File manager now is in Emacs. ⛅

Thunar was very good, but too simple. ☄️

Maybe my package will became quickly very popular. 💅

😶

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar

I finally 😫 released a major update to my Emacs package
that allow to search with toneless pinyin in Chinese ☯
characters and in normal pinyin text. 😊

It is "isearch submode".

You just press: C-s nu, and it will find for you 女 and nǚ.

By default it looks for both pinyin and Chinese characters,
but you can switch to pinyin with M-s p and to characters
with M-s h keys.

Code uses translation table (named Quail map) from Quail
minor mode (input multilingual text easily).
It is possible to adopt this code to many other
languages. ☆

I spent one month to code this (⚆i⚆) and I am going to use
it for the rest of my life. _/¯(⏿⏿)¯_

video/mp4

Anoncheg, to python
@Anoncheg@emacs.ch avatar


I fixed style and one error with help of Python static analysis tools in my
project for simple file sharing (available on pypi). ⚿
I used Bandit for security checks, pycodestyle and Pylint that
gives me most of warnings. It was very useful Python experience. ⛳

I have been search for new hosting for my blog. 📢 (My wordpress blog was
deleted for unknown reason. 💩)
I found many hosts for blogs that allow html and maybe emacs Org files.

  • sourcehut.org - anternative to github
  • codeberg.org - alternative to github
  • blogs.vivaldi.net/ - allow html and visible in search engines.
  • gitlab.com - overrated, not opensource friendly,
    read https://sfconservancy.org/GiveUpGitHub/ for more.
  • pages.cloudflare.com

I found payed internship in open source and open science. 😱
It is open to applicants around the world.
,000 USD for 3 months. Registration should be opened at Febrary.
I will participate. 💰 For more information read https://www.outreachy.org
😶

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar

Today I have solved simple task for a job to test my algorithmical skills.
It was about an ant which is moving in 4 directions ... well it's disgusting. 💩
I used BFS (Breadth-First Search) to check availability of path on grid.
I solved task partially and used approxamation with linar regression to predict
final result. 🤹

At wendsday I was at "Tinkoff.AI Computer Vision Meetup #2" _/¯(⏿⏿)¯_
where I watched reports about:

  1. Eye tracking 👀
  2. Documets processing with CV and OCR
  3. pose estimation 웃
    Tinkoff is a good company.

Emacs: I rewrited my package for Chinese pinyin search in text.
It takes "pinyin" as input and find pīnyīn and pīn yīn or pīnyin,
it's very usefull for Chinese learners as I am.
I am going to publish on MELPA shortly. 🍨

😶

Anoncheg, to machinelearning
@Anoncheg@emacs.ch avatar

CV task is finished. :) I successfully solved captcha.
I achieved accuracy 88%, when 90% was required by the company.
They didn't accept my work and I was not hired. ☄️
As I assumed, others have been using neural networks, while I used OpenCV only.
I just don't have GPU. ⛔

Here is steps how I made it:🐱

  1. Found two rectangles at hint image for hints with help of cv.findContours
    and collected statistic about max/min position of them.
  2. Fixed angle for main image with help of cv.HoughLinesP
  3. Found train position on main images by HSV colour range.
  4. Found subimages points on first main image with help of cv.SIFT and cv.FlannBasedMatcher.
  5. Calcuate hint position by clustering points with help of cv.kmeans automatic clustering.
  6. Calculate closes train by x and by y coordinates to found hint subimages.
    🥁

    😶
Anoncheg, to emacs
@Anoncheg@emacs.ch avatar

Part1: In Emacs there is a super useful command: C-u C-space set-mark-command
that allow to return to recent place where you was doing anything. 🤘
But I have problem with this command in Org-mode - it doesn't expand headers
tree and text around it header was fold. 💔
The reason for this was in "Sparse Tree" feature and historical approach:
entire document is folded as much as possible.
That is why, as I understand, (org-fold-show-entry) and (outline-show-entry)
hide some headers and text.⛇
To show this text there is command for it:
C-c C-r (org-reveal)
I don't know why this command require '(4) parameter to work properly. ¯_(ツ)_/¯

I wrote hook that will automatically call org-reveal with appropriate
parameter '(4):
(add-hook 'org-mode-hook (lambda ()
(advice-add 'org-fold-show-context :after (lambda (&rest args) (org-reveal '(4)) ))
))

-mode

Anoncheg, to datascience
@Anoncheg@emacs.ch avatar

I finished task for Allmagen company, I hope they will hire me.
It was about predicing probability of event.
I used classic RandomForest ensemble for binary classification with calibration.
It had extremely unbalanced classes and I solved it with oversambling by SMOTE technique.
It is very simple technique, but I did not find current Python implementation and spend some time on fixing old implementation.
I dont want to install huge untrusted library with dependencies for small function.
There is left much space for improvement: I didn't use time siries analysis and didn't use time data for prediction.

I used Org mode instead of Jupyter, my :timout parameter for source block was very helping.

I have been reading article about new applying NN for tabular data: https://research.yandex.com/blog/embeddings-for-numerical-features-in-tabular-deep-learning
It stated that efficint model is to use: embeddings -> periodic encoding -> Liner -L ReLU, where

In Emacs I was able to emulate M-x call: (execute-extended-command nil indent-for-tab-command nil)
it was required for proper indentation.

-mode

Anoncheg, to random
@Anoncheg@emacs.ch avatar

test

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar

I solved error in Info in :
user-error: No such node or anchor: Positions
the cause is unknown but after extractig bz2 file error disapeared.

Here is steps:

  1. cd /usr/share/info/emacs-28/
  2. sudo bzip -d elisp.info.bz2

I tried with Emacs, it is Linux distribution of modified
packages for under GPLv3. It perfectly easy to install from
F-Droid. It works great with Org mode and with Unexpected
app ⌨. But still it less comfortable than a real keyboard. And of
course it suffer from privacy (and security 🔒) as full
of closed source code. ◕‿◕

Here is diagram that I created with Emacs Artist mode for my being recently
writen "redshift" script-app.

Anoncheg, to email
@Anoncheg@emacs.ch avatar

:emacs: I fixed displaying of links in client for
: when hover cursor and copy.
🌐 I learned how to use standalone
obfuscation proxy obfs4 to private all Internet connections as socks5.
🐶 Here is amazing article:
https://hamy.io/post/000d/how-to-hide-obfuscate-any-traffic-using-obfs4/
I learned more about Elisp: it have issues with readability because of
much flexibility and freedom to choose, but almost all programmers
don't leave comments. Self-documenting code should be a code with
human short documentations about main parts, not a that "you
think" is easy to understand. One small good program is as good as a
bunch of products.

Anoncheg, to emacs
@Anoncheg@emacs.ch avatar
  1. fixed emacs: message of found auto save # was hidden by another messages
  2. I learned code extension mechanism called "advice", it can add code before/after/around any Elisp defined function.
  3. I enhance my scripts with image posting to , I use Org mode and call saved bash scripts,
    here is how it looks now:
Anoncheg, to random
@Anoncheg@emacs.ch avatar

test

Anoncheg, to python
@Anoncheg@emacs.ch avatar

I learned how to write tests in pytest - popular library and difference between flat and src layout of project.

Anoncheg, to fediverse
@Anoncheg@emacs.ch avatar

Today:
I was playing with API and create cURL scripts in Bash and skeletons for to create messages here very fast.

Anoncheg, to random
@Anoncheg@emacs.ch avatar

It is post 🎃​. This is going to be the place for daily work report. 💼
Short history:

  • 2018-2019 junior SAP ABAP, learning English and Dutch.
  • 2019-2023 real work: ML, DL, DS, Python :python:.
  • All
  • Subscribed
  • Moderated
  • Favorites
  • megavids
  • thenastyranch
  • magazineikmin
  • cubers
  • InstantRegret
  • cisconetworking
  • Youngstown
  • vwfavf
  • slotface
  • Durango
  • rosin
  • everett
  • kavyap
  • DreamBathrooms
  • provamag3
  • mdbf
  • khanakhh
  • modclub
  • tester
  • ethstaker
  • osvaldo12
  • GTA5RPClips
  • ngwrru68w68
  • Leos
  • anitta
  • tacticalgear
  • normalnudes
  • JUstTest
  • All magazines