jw_cook · HackerTrans

TL;DR:

    gh config set telemetry disabled

jw_cook

3 mesi fadiscuss

iNaturalist would agree with you; they explicitly say[1] it's not meant to be the primary source for your photos. Users generally fall into a couple broad camps:

1. Mostly use the mobile app, and take photos and upload observations directly from there. Local photo collection either isn't a priority or is backed up by their phone's cloud sync.

2. Mostly use inaturalist.org via a desktop browser, with either a standalone digital camera or mobile photos synced to desktop. Local filesystem (hopefully plus backups) is the source of truth.

I have been working on a desktop application[2] with a long-term goal of full bidirectional sync, and a secondary goal of offline usage. The current feature set is fairly modest and read-only, though, focusing on organizing local photos using data from iNat.

[1] https://www.inaturalist.org/pages/about

[2] https://github.com/pyinat/naturtag

jw_cook

3 mesi fadiscuss

Points are rendered server-side, backed by Elasticsearch, and served as PNG tiles for each zoom level. Individual markers are only rendered for small sets. Some of the relevant source code:

https://github.com/inaturalist/inaturalist/blob/main/app/ass...

https://github.com/inaturalist/inaturalist/blob/main/app/ass...

https://github.com/inaturalist/inaturalist/blob/main/app/ass...

jw_cook

3 mesi fadiscuss

Both iNaturalist and Observation.org publish observation data to GBIF:

https://www.gbif.org/dataset/50c9509d-22c7-4a22-a47d-8c48425...

https://www.gbif.org/dataset/8a863029-f435-446a-821e-275f4f6...

jw_cook

3 mesi fadiscuss

IMO it's best-in-class. The next best thing might be google's speciesnet: https://github.com/google/cameratrapai

jw_cook

3 mesi fadiscuss

It is a gem. There are all kinds of fun location/organism-specific tools you can put together with the public read-only data, and owlsnearme is a good example of that. I just used it to check my area and learned there are snowy owls nearby, which is new to me!

The iNat API certainly has some quirks and shortcomings, but in terms of usability it's uncommonly good compared to most biodiversity platforms. I maintain the python API client[1], which is used for data visualizations, doing useful things with your own observation data (which is how I got into it), Jupyter notebooks, Discord bots, and some research/education workflows.

[1] https://github.com/pyinat/pyinaturalist

jw_cook

3 mesi fadiscuss

I've wanted to do something similar, but unfortunately their CV model isn't public and can't be used through their API.

jw_cook

4 mesi fadiscuss

It's not framed exactly the same way, but I think this article is relevant: https://siddhantkhare.com/writing/ai-fatigue-is-real

Discussion: https://news.ycombinator.com/item?id=46934404

jw_cook

8 mesi fadiscuss

Yeah, I hear you. I almost started using a purely text-based todo workflow for those same reasons, but it was hard to give up some web UI features, like easily switching between list and kanban-style views.

My use case looks roughly like this: for a given project (as in hobby/DIY/learning, not professional work), I typically have general planning/reference notes in a markdown file synced across my devices via Nextcloud. Separately, for some individual tasks I might have comments about the initial problem, stuff I researched along the way, and the solution I ended up with. Or just thinking out loud, like you mentioned. Sometimes I'll take the effort to edit that info into my main project doc, but for the way I think, it's sometimes more convenient for me to have that kind of info associated with a specific task. When referring to it later, though, it's really handy to be able to use ripgrep (or other search tools) to search everything at once.

To clarify, though, Vikunja doesn't have a built-in feature that exports all task info including comments, just a REST API. It did take a little work to pull all that info together using multiple endpoints (in this case: projects, tasks, views, comments, labels). Here's a small tool I made for that, although it's fairly specific to my own workflow: https://github.com/JWCook/scripts/tree/main/vikunja-export

jw_cook

8 mesi fadiscuss

The linuxserver.io image for Nextcloud requires considerably less babysitting for upgrades: https://docs.linuxserver.io/images/docker-nextcloud

As long as you only upgrade one major version at a time, it doesn't require putting the server in maintenance mode or using the occ cli.

jw_cook

8 mesi fadiscuss

"You're absolutely right!"

jw_cook

8 mesi fadiscuss

Check if your router has an option to add custom DNS entries. If you're using OpenWRT, for example, it's already running dnsmasq, which can do split DNS relatively easily: https://blog.entek.org.uk/notes/2021/01/05/split-dns-with-dn...

If not, and you don't want to set up dnsmasq just for Nextcloud over LAN, then DNS-based adblock software like AdGuard Home would be a good option (as in, it would give you more benefit for the amount of time/effort required). With AdGuard, you just add a line under Filters -> DNS rewrites. PiHole can do this as well (it's been awhile since I've used it, but I believe there's a Local DNS settings page).

Otherwise, if you only have a small handful of devices, you could add an entry to /etc/hosts (or equivalent) on each device. Not pretty, but it works.

jw_cook

8 mesi fadiscuss

The article mentions Vikunja as an alternative to Nextcloud Tasks, and I can give it a solid recommendation as well. I wanted a self-hosted task management app with some lightweight features for organizing tasks into projects, ideally with a kanban view, but without a full-blown PM feature set. I tried just about every task management app out there, and Vikunja was the only one that ticked all the boxes for me.

Some specific things I like about it:

  * Basic todo app features are compatible with CalDAV clients like tasks.org
  * Several ways of organizing tasks: subtasks, tags, projects, subprojects, and custom filters
  * list, table, and kanban views
  * A reasonably clean and performant frontend that isn't cluttered with stuff I don't need (i.e., not Jira)

And some other things that weren't hard requirements, but have been useful for me:

  * A REST API, which I use to export task summaries and comments to markdown files (to make them searchable along with my other plaintext notes)
  * A 3rd party CLI tool: https://gitlab.com/ce72/vja
  * OIDC integration (currently using it with Keycloak)
  * Easily deployable with docker compose

jw_cook

9 mesi fadiscuss

This was a fun read, and well written. Thanks for sharing! Adding/improving support for some niche piece of hardware sounds like an ideal way to get started with kernel development, and something I'd like to try myself sometime.

jw_cook

3 anni fadiscuss

For me the main reason to switch to AdGuard was that it can easily run on OpenWRT (and PiHole can't). It's really convenient to run DNS adblock on the same device as your router.

jw_cook

3 anni fadiscuss

ESPHome looks really cool!

I've been looking for a more convenient way to configure some ESP32-based input devices (similar to macropads). I was interested in QMK, but it doens't support ESP32. So far I've been using MicroPython / CircuitPython, which I generally like, but on multiple occasions I've thought "I wish I could just put this in a config file."

The matrix keypad and key collector components look similar to what I was looking for. Can the key collector be used with other multiplexing methods like shift registers?

jw_cook

3 anni fadiscuss

It works with all docutils and Sphinx roles, and almost all directives, including extensions.

A notable exception is autodoc (automodule, autoclass, etc.), and any other directives that generate more rST. The current workaround is to use eval-rst:

https://myst-parser.readthedocs.io/en/latest/syntax/code_and...

Some more discussion about that in these issues:

https://github.com/executablebooks/MyST-Parser/issues/163

https://github.com/sphinx-doc/sphinx/issues/8018

jw_cook

3 anni fadiscuss

I think this is also one of my least favorite syntax choices in rST. I second the suggestions to use Sphinx with Markdown via MyST.

jw_cook

3 anni fadiscuss

I can't recommend this enough! It's such a quality of life improvement to get the powerful dynamic documentation features of rST and Sphinx (and its many extensions), but in the more pleasant and familiar syntax of Markdown. I use MyST + Sphinx for all my project docs now.

This podcast episode is worth a listen for anyone interested in these tools and where they're headed: https://talkpython.fm/episodes/show/354/sphinx-myst-and-pyth...