#Python

BERT for search and OCR: MLM mechanics, WordPiece, and encoder successors

Why Google added BERT to search in 2019, how MLM training really works (15% mask, 80/10/10, WordPiece), and where encoder-only models still beat LLMs — rerank, classification, and OCR correction.

Sugar YMP-01 thermal printer on M1 Mac: halftone + 1D 49 F0 nn density command

Tested on M1 Max: Floyd-Steinberg halftone + BLE pacing + a vendor-specific density command `1D 49 F0 nn` to print sharp photos on the Sugar YMP-01 thermal mini printer from Python.

HP Sprocket 200 over RFCOMM on M1 Mac: 11× faster than BLE in 5.38s for 140KB

Tested on M1 Max: switching a Python BLE client to RFCOMM (SPP Ch.2) cuts transfer from ~60s to 5.38s for a 140KB JPEG. Covers PyObjC quirks, macOS Bluetooth entitlement, and an isolation experiment confirming olie.xdev's 'possibly not required' steps.

Orbis local test: psf/requests and expressjs/express as 3D dependency graphs

Tested Orbis in Docker: psf/requests (35 modules) makes a clean 3D graph, expressjs/express (98 modules) collapses into one giant sphere. Notes on what tree-sitter actually captures for code-review prep.

uv for Claude SDK Python agents: pyproject.toml + uv.lock as the stable entry

uv 0.9.21 as the entry for small Claude SDK Python experiments: uv init, uv add, uv run, uv.lock keep agent projects reproducible across machines and Codex/Claude Code sessions. Operational notes, not a benchmark of the DEV article's uv 0.11.11.

Human-LLM text segmentation on M1 Max: WCP works, raw log-likelihood doesn't

Tested arXiv:2605.03723 on M1 Max + Qwen3-8B-Base: WCP runs in pure Python, but raw log-likelihood floods boundaries even on human-only text.

Running Apache Airflow 3.2.0 with PostgreSQL and Executing Your First DAG

Steps to set up Apache Airflow 3.2.0 with a PostgreSQL metadata database, run your first DAG, and configuration notes for moving beyond standalone mode.

Wiring Up a Multimodal Japanese Local RAG with FastAPI, Chroma, Open WebUI, and Ollama on M1 Max

Hands-on log of building the DEV article's PDF RAG on M1 Max 64GB, extending it with images via CLIP, and pushing through Japanese with bge-m3 + Qwen3.6 35B. Documents the modality gap, the dual inference server crash, and LLM-jp 4-8B's empty chat template silently dropping the system role.

Reading an Article on Building a Local PDF RAG with FastAPI, llama.cpp, Chroma, and Open WebUI

Notes on a DEV Community article that wires up FastAPI as an OpenAI-compatible RAG API layer with llama.cpp, Chroma, and Open WebUI, plus where the architecture fits and what to watch for.

Async HLS Video Extraction Pipeline and Pinterest's Internal MCP Ecosystem

Async pipeline design for extracting HLS video streams into MP4, plus technical details from Pinterest Engineering Blog's enterprise MCP ecosystem.

Investigating PC Control for the Nikon COOLPIX A1000

An experiment log investigating USB/WiFi/BLE connection specs of the Nikon COOLPIX A1000 and attempting PC connection via OSS tools. PTP/IP protocol analysis and gphoto2 testing results.

The EML Operator: Expressing All Elementary Functions with Just exp(x) - ln(y)

A paper claims that a single binary operator eml(x, y) = exp(x) - ln(y) combined with the constant 1 can express all elementary functions — arithmetic, trig, logarithms, even pi. I read the paper and tested it in 5 languages.