@Angad – Hack Club Scrapbook

Sun, Jul 7

I built a simple, <https://github.com/slashtechno/clock|open-source> clock that displays rhymes. Inspired by the Hack Club workshops <https://workshops.hackclub.com/simple_clock/|Simple Clock> and <https://workshops.hackclub.com/json_db/|JsonDB>, along with the <https://www.kickstarter.com/projects/genmon/poem-1-the-ai-poetry-clock|Poem/1>, an “AI rhyming clock”. I built it to learn JavaScript and it’s my first full-stack project. Since it’s a website, it can be viewed in a variety of ways - including as a live wallpaper or potentially, on an E Ink display.

https://scrapbook-into-the-redwoods.s3.amazonaws.com/a1c20bcb-4686-4571-8d2f-2a90ab32f360-image.png

https://imgutil.s3.us-east-2.amazonaws.com/21e6b0144be0cb4e31df1930dd80bf3c5c5df257d4e5078fcf72f1f17df9f622/09a0f06a-9968-4b77-a472-2d4f64a24124.png

Sat, Jun 29

I published a blog post regarding my project Cross Blogger to my Hugo site. angad.me/blog/2024/rewriting-cross-blogger I also made various improvements to the site. I made it so categories are automatically as buttons on /blog and enabled PaperMod’s support for linking GitHub for edit suggestions. As the post is in plain-text/Markdown, the edit history can easily be viewed on GitHub: github.com/slashtechno/angad/commits/main/content/blog/2024/rewriting-cross-blogger.md

https://scrapbook-into-the-redwoods.s3.amazonaws.com/29ce78d6-4874-4f36-8059-6ba95dcd8986-image.png

https://imgutil.s3.us-east-2.amazonaws.com/ac1a5277131dcfc7058adc3e7e5aaddd29f99c0ccf6be6dfb7baf64c035bbe4f/4da08b0a-e1a0-4650-ad0e-3275b5c59164.png

Fri, Jun 28

Nearly two years ago, I began work on Cross Blogger. Cross Blogger was my first real Go project with the first release being rather small with all the code in a main.go file. In the last couple of weeks, I rewrote the entire codebase, adding features like Docker support and proper OAuth. The most significant change, however, was the ability to use it as a headless CMS for static sites (such as Hugo sites) powered by Blogger. Configuration is rather simple and can use most common formats such as JSON, YAML, TOML, etc. The project works by fetching posts from Blogger on a customizable interval and comparing against the last known posts. Any new posts are converted to Markdown and optionally, pushed to a Git repository. Assuming the Git repository is connected to a service like Vercel or is using a workflow, the changes will be deployed automatically. If overwriting is enabled in the configuration file, posts unpublished in Blogger will be deleted from the Hugo site. Multiple destinations can be used too. In practice, that means multiple websites, for example, frontmatter mappings can also be customized to either disable frontmatter options or to make it work with other static site generators. The code is also structured in a manner making it easy to add support for new features. Any feedback would be appreciated. Feel free to make PRs or issues in the repository GitHub repository: github.com/slashtechno/cross-blogger YouTube video (same as attached video): www.youtube.com/watch?v=H3HyQ7h5ogE

https://imgutil.s3.us-east-2.amazonaws.com/ac1a5277131dcfc7058adc3e7e5aaddd29f99c0ccf6be6dfb7baf64c035bbe4f/8b29a300-95ac-4662-8bbc-f38a10bd4f3e.png

Fri, Jun 28

Nearly two years ago, I began work on Cross Blogger. Cross Blogger was my first real Go project with the first release being rather small with all the code in a main.go file. In the last couple of weeks, I rewrote the entire codebase, adding features like Docker support and proper OAuth. The most significant change, however, was the ability to use it as a headless CMS for static sites (such as Hugo sites) powered by Blogger. Configuration is rather simple and can use most common formats such as JSON, YAML, TOML, etc. The project works by fetching posts from Blogger on a customizable interval and comparing against the last known posts. Any new posts are converted to Markdown and optionally, pushed to a Git repository. Assuming the Git repository is connected to a service like Vercel or is using a workflow, the changes will be deployed automatically. If overwriting is enabled in the configuration file, posts unpublished in Blogger will be deleted from the Hugo site. Multiple destinations can be used too. In practice, that means multiple websites, for example, frontmatter mappings can also be customized to either disable frontmatter options or to make it work with other static site generators. The code is also structured in a manner making it easy to add support for new features. Any feedback would be appreciated. Feel free to make PRs or issues in the repository GitHub repository: github.com/slashtechno/cross-blogger YouTube video (same as attached video): www.youtube.com/watch?v=H3HyQ7h5ogE (cross posted from #ship)

https://imgutil.s3.us-east-2.amazonaws.com/ac1a5277131dcfc7058adc3e7e5aaddd29f99c0ccf6be6dfb7baf64c035bbe4f/55b5ea8b-8547-48b7-9ab0-3921fc8a4951.png

Thu, Jun 27

I made a PR to @Michael-U0756DZHJ74’s LLM-powered Discord bot. Made the following changes: • Added support for any OpenAI-compatible API, such as OpenRouter • Changed configuration schema • Use toml.dump instead of manually writing the lines of TOML github.com/Mikeymoo5/discordLLM/pull/1

https://scrapbook-into-the-redwoods.s3.amazonaws.com/09feaea3-ff77-4e11-a45a-0d6af5372ced-image.png

https://imgutil.s3.us-east-2.amazonaws.com/0f3a5a6cdd459647d4fa478a1fe3953292f48df1b4d5253599734cdd73438de6/6198fdc9-5d94-4dcc-8fa2-49c0961adb1a.png

Wed, Jun 19

I rewrote cross-blogger, a program I wrote two years ago and a soon-to-be headless CMS for posting from Blogger (or other user-friendly blogging platforms) to a static site generator, such as Hugo. Currently, it functions rather well, as well as being extremely fast. However, it's not fully automated. I plan on making it automated (and adding Git integration) so publishing a blog post to a static site doesn't include having to deal with deployment. PR showing changes made with the rewrite: github.com/slashtechno/cross-blogger/pull/37

https://imgutil.s3.us-east-2.amazonaws.com/df9eb41b26b2d846db8a55b57d1f92618b672deafd2a17b6466d91779c33fb96/23f3793c-6655-457b-b749-a3a80f52e5a8.png

Mon, Jun 17

Published v0.1.2 of scrape-and-ntfy! github.com/slashtechno/scrape-and-ntfy/releases/tag/v0.1.2 I fixed webhook compatibility so it works with Slack now.

https://scrapbook-into-the-redwoods.s3.amazonaws.com/6c572e35-df1b-4125-bcf7-3ecb7a4cf054-image.png

https://imgutil.s3.us-east-2.amazonaws.com/0c37390113e74d99a86d53170262cbe3627670e85ef088977e5cb6baf6aa5772/13585c3a-6197-470a-8f92-f6cd5e8f9481.png

Tue, Jun 11

I added price/numeric notifier types to Scrape and Ntfy! Below you can see a screenshot of webhook-based notifications detailing price changes. In this case, I was just using a simple HTML file to allow me to edit the prices for testing. I also made some other changes such as not requiring headless mode and a custom browser path to be used together. github.com/slashtechno/scrape-and-ntfy/compare/45fc8cf4544ecf1d783827279a9256e562fa1faf...9a7f01ab99950716938d3c65f4f05843de28e74e

https://scrapbook-into-the-redwoods.s3.amazonaws.com/2530943f-d80e-4ea8-824f-e9880ae3ffc9-image.png

https://imgutil.s3.us-east-2.amazonaws.com/7aeb14f8056ad5aef15135118317f18e9f052b89262296bc75c8c4d8e384c919/ae473c04-2516-460b-9679-d9d03602a1da.png

Thu, Jun 6

I refactored my Mastodon bot and significantly improved the help command. I'm waiting for my account application to be approved so I can have a production version of the bot that's online 24/7, then I'll #ship. angad.me/blog/2024/introducing-pystodon github.com/slashtechno/pystodon

https://scrapbook-into-the-redwoods.s3.amazonaws.com/21b4273e-7045-46e8-b232-5bffa0a7bf1d-image.png

https://scrapbook-into-the-redwoods.s3.amazonaws.com/3874539b-8482-47d3-b322-8a6ee46fa7a6-image.png

https://imgutil.s3.us-east-2.amazonaws.com/e41f1da6afd0a7de6e5919f7e1fd2d3429c418be774cad5be5d8c81ec6e80c9e/5876b7c7-c6b4-4e5b-af3b-e0b02cc17113.png

Wed, Jun 5

I added time-based reminder support to my Mastodon bot, rathercurious-mastodon. I also structured the project in a way that it can be used as a framework for building other bots. With this update, using a command such as @rathercurious@botsin.space> #remindme in 10m will schedule a reminder that is posted in 10 minutes. To reduce clutter, you can mention this bot in a DM (even in an existing thread) and it will reply to you in a DM. Otherwise, it will just match the visibility of your command. I'll probably end up refactoring the code a bit before running it 24/7. This was my first project that uses a proper database. Surprisingly, I was able to add this feature that I've been meaning to for a while in one day. github.com/slashtechno/rathercurious-mastodon/pull/1

https://scrapbook-into-the-redwoods.s3.amazonaws.com/eb74deb4-3de3-41ae-b97e-34c582e798bd-image.png

https://imgutil.s3.us-east-2.amazonaws.com/394f543ff390ac409e8122d5a948180c44f857cc7789f77520febaa020e628cd/3b1c6fce-8105-4376-8c24-e712fdb3f90f.png

Mon, Jun 3

I added phidata support LLMail! github.com/slashtechno/llmail/pull/4 I released this as part of v0.3.0 I also started using Ollama as I was having trouble with OpenRouter. I tested various Ollama models to find which one was best and so far, Mistral seems to be great.

https://scrapbook-into-the-redwoods.s3.amazonaws.com/00e5442b-d043-4e33-9572-ae44d47e58d7-image.png

https://imgutil.s3.us-east-2.amazonaws.com/e8e3e1f1028f529aabf611e74728a45d5c2c5237bed19a31afb1935c355b4465/56597055-b4b8-4537-8116-77b55a5f6b25.png

Sun, Jun 2

Continued adding support for Phidata tools to my project LLMail. I was having trouble implementing DuckDuckGo and Exa.ai (results were outdated) I was able to add the page scraping function, however. Changing the model to adrienbrault/nous-hermes2theta-llama3-8b:q8_0 and using Ollama instead of OpenRouter also seemed to help.

https://scrapbook-into-the-redwoods.s3.amazonaws.com/8ffd1f31-8a54-483c-a07f-6b4e844c8a1e-image.png

https://imgutil.s3.us-east-2.amazonaws.com/b08d535eebd88ef23998c5e0c3b7de6c3ac20123ddefaa822d824bfa910a2177/1811d61a-ae0d-422f-9d28-e407a8b7bc69.png

Sat, Jun 1

I recently got the domain angad.me and "deprecated" my old site. I realized today that I should add canonical tags for SEO. However, the new, minimal theme I use doesn't support them. Thus, using code from the full theme I use (PaperMod), I added support. github.com/slashtechno/slashtechno.github.io/compare/e3345874eb697933c76c8cb3b56262f6b969d264...7d030b878897c0a3cb872f77b69241c9bd2c23e1e

https://imgutil.s3.us-east-2.amazonaws.com/e41f1da6afd0a7de6e5919f7e1fd2d3429c418be774cad5be5d8c81ec6e80c9e/c942cd06-28c7-47fb-a151-45c73389ecb5.png

Sat, Jun 1

Worked on integrating phidata into LLMail I couldn't get searching implemented with Exa or DuckDuckGo though.

https://scrapbook-into-the-redwoods.s3.amazonaws.com/f2773f6d-b6d1-4e28-8481-d49f8428b9ab-image.png

https://imgutil.s3.us-east-2.amazonaws.com/b08d535eebd88ef23998c5e0c3b7de6c3ac20123ddefaa822d824bfa910a2177/bbccd009-6400-4313-967a-f1d36af55b76.png