Skip to content

All Articles

Fully automated bluesky commenting system on mkdocs

Background

This post was basically continuation of the previous post and outlined the steps taken to achieve the following flow, however there was an announcement recently which stops bluesky search as means of automation so I had to modify the flow:

create a blog postpublish the blog posttrigger creation of bluesky post automaticallycomments thread activated on blog postyesPost deployed?

The modified flow that I applied, is as shown below and I will update this post in due course to provide documentation on code changes but code can also see the code here

Bluesky for comments on mkdocs blog

Background

I have recently migrated this blog from Jekyll to Mkdocs using material theme. I was also planning to move away from Disqus commenting system given all the known issues and just when I had finished the migration to Mkdocs, there was this post from Emily Liu - which showcases how easily the bluesky replies to a post can be added in comments section of a blog page. Followed closely by this, was a post from Cory Zue who created NPM packaging.

I initially was looking to use that on my blog and while searching for that package on jsdelivr, I came across the bluesky-comments-tag: a package created by Matt Kane. I understood the code behind this one much better and that was my starting point. There were two problems with the solution thus far that I wanted to solve.

Problem Description

For the comments to appear on the blog, the bluesky post url needs to be added to the frontmatter. Now that is a cyclic process because you can't post the url of your blog post on bluesky until you have published it.

This, I have solved to an acceptable flow as shown below and the "How?" is explained in this post.

Logseq Customisations for Project Management Template

Background

While the overall planning of project timeline gets lot of attention in the world of software, the most important aspect of project management in my experience is maintaining and tracking Issues and Actions. This in normal project management practice is carried out through the use of an Actions Log and an Issues Log. In addition any medium complexity project invariably will have internal and external dependencies / constraints, risks which I tend to track on Constraints Log and Risks Log. Finally, every project has decisions that I track on a Decisions Log.

Additionally, any opportunities that I identify during the course of project that are not in the scope of my project I capture those too in an Opportunities Log

Now all these logs have fairly standard fields so I created and started using an excel template and began calling it CARDIO Log short for log of Constraints, Actions, Risks, Decisions, Issues, Opportunities for a given project. This has worked well for me over the past 15 years or so but there are times where in order to maintain it meant cross linking an action to an issue or a risk and so on and more often than not it would become easier to just track actions in one of the other logs and that would make it a bit chaotic.

That problem, however, is what I thought, can be resolved using Logseq especially after starting with the template by the Logseq community user Luhman and starting with his template and explanation provided here.

Git Basics

Basics of Git

Recently, I was asked to help people understand what is this whole business with GIT and I ran a few sessions on the topic specifically keeping in mind that my audience consisted of tech enthusiasts with minimal tech background. The content below is how I explained and as it went down well, I figured it may be helpful to someone else too; hence this post. 😄

JupyterLite on Github Enterprise with Panel enabled

Recently I found myself in a scenario where I had access to Github Enterprise but was limited to a windows machine. Now I wanted to host JupyterLite on Github Enterprise so it can be used by some other people on my team. Challenge was the "Actions" on this instance of Github enterprise were disabled and so I had to build Jupyterlite locally and push it on Github. I figured while at it, I might as well enable the "Panel" and few other dependencies. The steps that worked for me are documented here. This is ofcourse assuming that Python is already installed on the device.

The final hosted JupyterLite can be seen in action here.

The steps below created a Jupyterlite instance that allows use of Panel using the %pip install panel magic command, and if used on Chrome or its derivatives it also allows exploring local file system.

Tech for Diabetics

Background

As a diabetic there are a number of things we are not able to control but one thing we can do is keep tabs on our data and while in past it would have meant meticulously noting down BG readings in a diary, today it is much much easier with all the apps and connected services. In this article, I aim to go through the list of technologies that I am aware of and use to keep on top of my diabetic data right from my phone.