- Extend portscan's check_port to return service banners alongside open port status
- Update scan_ip and full_scan to propagate banner data in open_ports tuples
- Add _DEFAULT_COLUMNS constant and _get_table_columns caching to AsyncDataSet
- Implement _invalidate_column_cache method for schema change tracking
- Add OpenAPIFormatter class with methods for parsing OpenAPI specs
- Implement extract_endpoints to list all API paths with methods and summaries
- Add get_request_body_schema and get_parameters for detailed endpoint introspection
- Include resolve_schema_ref for handling $ref references in schema definitions
- Support async HTTP fetching of specs via aiohttp with configurable timeout
The change shortens the artificial delay in test_upsert_update and test_update methods
within TestAsyncDataSet, lowering the wait from 2 seconds to 1.1 seconds to speed up
test execution while still allowing asynchronous operations to complete.
- Add automatic `updated_at` timestamp to args in `upsert` method
- Add missing `asyncio.sleep(2)` in `test_upsert_update` to ensure timestamp difference
- Implement `asyncSetUp` and `asyncTearDown` for proper test database lifecycle
- Add comprehensive test methods for insert, get, update, and upsert operations
- Import missing type annotations for better code clarity
Implement a complete 3D city exploration environment using Three.js, featuring a sky-blue fogged scene, first-person camera controls with pointer lock, WASD/arrow key movement, and dynamic shadow mapping. The scene includes ambient and directional lighting, a ground plane, and procedurally generated building structures with varied heights and colors.
Implement a new docker_volume_monitor.py module that uses the Docker SDK to inspect running containers, retrieve volume sizes via the inspect API, and automatically stop containers when any mounted volume exceeds a configurable 1GB threshold, with a dry-run flag for safe testing.
- Add core WordleBot class with methods for generating guesses and processing color-coded feedback
- Implement letter frequency analysis and candidate word filtering based on previous guesses
- Include support for hard mode constraints and optimal guess selection strategy
Implement a fully-featured asynchronous Python client for the Open WebUI HTTP API using aiohttp. The client supports Bearer/JWT authentication, model listing, chat completions with optional server-side streaming, Ollama proxy helpers (generate, list tags, embed), and RAG utilities for file upload and knowledge collection management. Includes async context manager for automatic session lifecycle, custom exception hierarchy (OpenWebUIError, APIResponseError), and 100% type-hinted production-ready code.
Implement a new vim_terminal_trainer.py module using curses for real-time character capture, auto-advancing on exact answer match, timer bar countdown, and persistent high-score storage in ~/.vim_terminal_trainer/score.json
Introduce a new regex.md document detailing the author's philosophy and approach to writing custom regex interpreters from scratch, emphasizing unique design over performance or external research.
Implement WordGenerator module in word_generator.py with a predefined syllable list, generate_word method producing configurable-length pseudo-random words by concatenating random syllables, and generate_unique_words method returning a specified count of distinct words via set deduplication, intended for human-friendly memorable identifiers.
Add two new files to the repository: disk_usage.sh provides a comprehensive disk analysis tool that displays overall usage, largest directories, files over 1GB, log file sizes, and large installed packages for both Debian and RPM-based systems. liebranca.py implements a Pinecone-based AI assistant named "liebranca2" with CLI chat interface and file upload functionality, configured to respond as the character Liebranca with specific personality instructions.
Introduce a new markdown file containing an AI-generated "Kiki dream" titled "The Clockwork Aurora", featuring surreal imagery of time rivers, floating islands, a silver-haired girl with a glowing pocket watch, and a labyrinth of frozen moments. The narrative explores themes of fractured time and cosmic clockwork mechanics.
Introduce repos.finddirty.py, a Python utility that scans a specified directory for Git repositories and reports those with uncommitted changes. The script uses pathlib for filesystem traversal, subprocess to execute git status commands, and sys for argument handling. It defines is_git_repo() to check for .git directories and is_dirty() to parse git status output for "tree clean" indicators, returning None for non-repo paths. Licensed under MIT.
Implement tokenizer.c containing is_valid_char helper and main function for reading text files, extracting alphanumeric words with dash/underscore support, and tracking word positions, line numbers, and paragraph counts across the input stream.
Updated the pricing documentation to reflect the new 2026 per-token costs for OpenAI's premium tier models, including GPT-5 Ultra and o3-max, with input/output rates of $150/$600 per million tokens respectively.
Introduce a new markdown file detailing usage-based fees for input and output tokens of OpenAI's highest-tier models, with specific per-token rates for GPT-4o and GPT-4o-2024-11-20, and a step-by-step methodology to convert abstract token pricing into human-understandable time-based costs for enterprise applications.
