2026-07-08 00:18:29
Recently, I was working on migrating our Elasticsearch cluster to serverless. We have an index where some field stores dense vectors used in KNN search.
2026-07-04 00:23:25
In Elasticsearch, when you index document to an index, if you do not specify the mapping for the field, Elastic will use dynamic mapping to infer the data type for your field.
2026-05-14 02:14:50
In our work, we are using renovate to update the dependency packages for Python projects.
Previously when we are using good old pip and requirements.txt for package management,
renovate works fine since it only needs to update the package version in requirements.txt.
We switch to uv this year as our project management tool.
Since uv uses pyproject.toml and uv.lock (lockfile) to ensure reproducible project setup,
we have an issue that when renovate bot creates a PR for the package updates:
it only changes pyproject.toml without updating uv.lock.
2026-05-12 01:16:27
To make python plugins work in Neovim, we need to set up the python provider properly. Usually, this means setting up the correct path to Python executable and also installing the pynvim – Neovim client for Python.
2026-04-17 17:45:00
Notes about how to use ripgrep to search hidden files and configuration related to Neovim.
2026-04-15 00:11:35
We have a Python project where there are several developers, and we have a mature CI pipeline to do code linting, formatting check for python, yaml, json, and testing etc. The problem with CI pipeline is that you can only get feedback after you commit your code and push to the remote repo. The feedback is not instant and may be delayed for several minutes. If there are some linting issues, you then need to fix the issue and make another commit. This is also making the PR history messy.
