Skip to content

agentic_instructions.md

Latest commit

 

History

History
79 lines (69 loc) · 3.99 KB

agentic_instructions.md

File metadata and controls

79 lines (69 loc) · 3.99 KB
Raw

agentic_instructions.md — python_projects

Purpose

Collection of Python scripts and tools for system automation, AWS administration, Datadog API interaction, AI/LLM integration, data processing (pandas/xlsx), Vivaldi browser control, and desktop utilities on Arch Linux.

Technology

  • Language: Python 3
  • Libraries: requests, pandas, flask, langchain, aiohttp, openpyxl, xlsxwriter, subprocess
  • External tools: dmenu, fzf, gum, notify-send, wezterm, aws CLI, kubectl
  • APIs: Datadog, OpenAI, Grok, AWS IAM/STS

Contents

  • helpers.py — Helper module: dmenu wrapper function, notify-send wrapper with criticality levels.
  • claude_sessions.py / claude_sessions_floater.py / claude_delete_session.py — Claude Code session management utilities.
  • fzf.py — fzf integration utility.
  • kube.py — Kubernetes utility script.
  • json_to_xlsx.py / xlsx_to_json.py — JSON/Excel conversion utilities.
  • pandas_tests.py / working_pandas_fs_integration.py — Pandas data processing experiments.
  • rag.py — RAG (Retrieval-Augmented Generation) context file selector.
  • audio.py — Audio utility.
  • prac.py / pyd.py / pyoperators.py — Python practice/learning scripts.
  • rotateimage.py — Image rotation utility.
  • test_workbook.py — Excel workbook testing.
  • newScript.py / path.py / responsetest.py — Miscellaneous utilities.

Subdirectories

  • ai/ — AI/LLM clients (Ollama local, OpenAI via langchain).
  • ansible/ — Ansible Tower project creation via awxkit API.
  • asus/ — ASUS router management via aiohttp/asusrouter library.
  • aws/ — AWS IAM user listing and STS caller identity scripts.
  • ddog/ — Datadog API clients for dashboards and usage stats.
  • grok/ — Grok AI chat client with persistent conversation threads.
  • hooks/ — Flask webhook test server.
  • json_scripting/ — JSON processing with pandas (kubectl, AWS output).
  • pandas_scripts/ — Pandas/Excel data processing and Datadog integration list management.
  • rag/ — RAG file selector with fzf multi-select and user prompt collection.
  • rust_python/ — Rust-Python interop via maturin (PyO3).
  • scripts/ — Desktop automation: Vivaldi tab tiler, process killer, password manager, keyboard brightness, wallpaper fetcher, etc.

Key Functions

  • helpers.py::dmenu(color, prompt) -> str — Launches dmenu with custom styling, returns selection.
  • helpers.py::notify_send(msg, criticality) -> str|None — Sends desktop notification via notify-send.
  • scripts/tabTiler.py — Vivaldi tab tiling via Chrome DevTools Protocol (CDP).
  • scripts/killer.py — Process killer with fzf/dmenu selection.
  • scripts/wallhaven.py — Wallpaper fetcher from Wallhaven API.
  • aws/awsListUsers.py::list_users(userlist) -> List[str] — Lists AWS IAM users, returns usernames.
  • ddog/ddog_api.py::dashboards(), usage() — Datadog API queries.

Data Types

  • Pandas DataFrames for tabular data processing.
  • JSON-serialized conversation threads (grok).
  • Excel workbooks via xlsxwriter/openpyxl.

Logging

  • print() for stdout output.
  • subprocess.Popen for shell integration.
  • notify-send for desktop notifications.

CRUD Entry Points

  • Run: python3 <script>.py [args] or ./script.py (with shebang)
  • Dependencies: pip install (requests, pandas, flask, langchain, etc.)

Style Guide

  • #!/bin/env python3 shebang
  • Type annotations on function signatures and variables
  • subprocess.Popen with communicate() for shell commands
  • match/case for dispatch (Python 3.10+)
  • sys.argv for CLI arguments, argparse for complex CLIs
  • Helper modules imported locally
def dmenu(color: str, prompt: str) -> str:
    dmenu_command: List[str] = ["dmenu", "-m", "0", "-fn", "VictorMono:size=20",
                                "-nf", "green", "-nb", "black",
                                "-nf", color, "-sb", "black", "-p", prompt]
    dmenu = sp.Popen(dmenu_command, stdin=sp.PIPE, stdout=sp.PIPE, text=True)
    output, err = dmenu.communicate()
    return output.strip()