Simon Willison simonw

## _llm
#compdef llm

# Instructions:
#
# - Have llm working on your setup, of course. Have some models or templates to actually use.
# - Have zsh and zsh-completions configured and working (i.e., installed, fpath configured, tabbing works for other commands, etc)
# - Place this file `_llm` alongside your completions (e.g., ~/.oh-my-zsh/custom/plugins/zsh_completions/src/)
# - Restart your terminal session or `source ~/.zshrc`
# - Try it out by tabbing while `-m` or `-t` and it should suggest whatever models you have available or templates you've configured.

## prompt.txt

You are a security robot. Your job is to observe and report any activity. Specifically, notice the people and what they are doing.

Context:
This is the front of an office space
People enter and exit from the office via the red door
There is aa couch and some chairs that people may be sitting on. There is also a snack area, and an area with equipment.  right of the image.

Return JSON.
{

## OIDC_for_Github.md

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              2 stars
            
          
                sakopov
                / OIDC_for_Github.md
            
            
              Last active
              September 3, 2024 09:53
            
              
                OIDC for Github and AWS
              
          
    Some rough cloudformation to add OIDC support for Github in AWS.

Create new identity provider for Github. (1b511abead59c6ce207077c0bf0e0043b1382612 is the known thumbprint for Github).


Note, you can discover current thumbprint using openssl.
 $ openssl s_client -servername token.actions.githubusercontent.com -showcerts -connect token.actions.githubusercontent.com:443 <    /dev/null 2>/dev/null | sed -ne '/-BEGIN CERTIFICATE-/,/-END CERTIFICATE-/p' | sed "0,/-END CERTIFICATE-/d" > certificate.crt
 $ openssl x509 -in certificate.crt -fingerprint -noout | cut -f2 -d'=' | tr -d ':' | tr '[:upper:]' '[:lower:]'
 1b511abead59c6ce207077c0bf0e0043b1382612


## digger.py
"get embeddings from an LLM database"
# LICENSE https://www.apache.org/licenses/LICENSE-2.0.txt
# Copyright (C) Randall Bohn 2023
# requires: llm>=0.9, sqlite_utils, numpy, umap-learn==0.5.3
from typing import Dict
import llm
import numpy as np
from sqlite_utils import Database
from umap import UMAP


## aicli.py
#!/usr/bin/env python3
import os
from datetime import datetime, timezone
from pathlib import Path

from prompt_toolkit import PromptSession
from prompt_toolkit.history import FileHistory
import openai
from rich.console import Console
from rich.markdown import Markdown

## csv2datasette
#!/usr/bin/env bash
# sudo ln -s "$(pwd)/csv2datasette" /usr/bin/csv2datasette

# csv2datasette is meant to explore CSV data. It is not meant to create a sustainable DB.
# csv2datasette is a bash script which open CSV files directly in Datasette. It offers
# a number of options for reading and exploring CSV files, such as --stats, inspired by WTFCsv.
#
# `--stats` option includes, for each column: the column name, the number of unique values,
# the number of filled rows, the number of missing values, the mininmum value, the maximum value,
# the average, the sum, the shortest string, the longest string, the number of numeric values,

## previewImageUploads.js
// Licensed under a CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
// http://creativecommons.org/publicdomain/zero/1.0/

(function (win,doc) {
    'use strict';
    if (!win.FileReader || !win.addEventListener || !doc.querySelectorAll) {
        // doesn't cut the mustard.
        return;
    }
    doc.querySelectorAll('input[type="file"][accept="image/*"]').forEach( function (fileInput) {

## README.md

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              12 stars
            
          
                palewire
                / README.md
            
            
              Last active
              May 24, 2023 18:09
            
              
                How to deploy a Prefect agent to Google Kubernetes Engine
              
          
    How to deploy a Prefect agent to Google Kubernetes Engine

This post contains code and commands you can use to deploy Prefect agents to Google Cloud’s Google Kubernetes Engine. The agents stand ready to execute workflows triggered by Prefect projects. One agent can run tasks from multiple projects.
The example here demonstrates how to create a single agent with minimal customization. It is configured with a Dockerfile, which installs necessary dependencies, and a k8s.cfg file, which connects the system to a Prefect account.
Agents are deployed via the gcloud command-line utility and its kubectl extension. Proper permissions within a project on Google Cloud are required.
Getting started


## single-file-web-component.html
<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8">
  <title>Single File Web Component</title>
</head>
<body>
  <template id=single-file>
    <style>
      h1 {

## app.py
from starlette.applications import Starlette
from starlette.responses import StreamingResponse
from starlette.requests import Request
from starlette.routing import Route
from pathlib import Path
from typing import IO, Generator

"""
Stream a file, in this case an mp4 video, supporting range-requests using starlette
	#compdef llm

	# Instructions:
	#
	# - Have llm working on your setup, of course. Have some models or templates to actually use.
	# - Have zsh and zsh-completions configured and working (i.e., installed, fpath configured, tabbing works for other commands, etc)
	# - Place this file `_llm` alongside your completions (e.g., ~/.oh-my-zsh/custom/plugins/zsh_completions/src/)
	# - Restart your terminal session or `source ~/.zshrc`
	# - Try it out by tabbing while `-m` or `-t` and it should suggest whatever models you have available or templates you've configured.

	You are a security robot. Your job is to observe and report any activity. Specifically, notice the people and what they are doing.

	Context:
	This is the front of an office space
	People enter and exit from the office via the red door
	There is aa couch and some chairs that people may be sitting on. There is also a snack area, and an area with equipment. right of the image.

	Return JSON.
	{
	"get embeddings from an LLM database"
	# LICENSE https://www.apache.org/licenses/LICENSE-2.0.txt
	# Copyright (C) Randall Bohn 2023
	# requires: llm>=0.9, sqlite_utils, numpy, umap-learn==0.5.3
	from typing import Dict
	import llm
	import numpy as np
	from sqlite_utils import Database
	from umap import UMAP
	#!/usr/bin/env python3
	import os
	from datetime import datetime, timezone
	from pathlib import Path

	from prompt_toolkit import PromptSession
	from prompt_toolkit.history import FileHistory
	import openai
	from rich.console import Console
	from rich.markdown import Markdown
	#!/usr/bin/env bash
	# sudo ln -s "$(pwd)/csv2datasette" /usr/bin/csv2datasette

	# csv2datasette is meant to explore CSV data. It is not meant to create a sustainable DB.
	# csv2datasette is a bash script which open CSV files directly in Datasette. It offers
	# a number of options for reading and exploring CSV files, such as --stats, inspired by WTFCsv.
	#
	# `--stats` option includes, for each column: the column name, the number of unique values,
	# the number of filled rows, the number of missing values, the mininmum value, the maximum value,
	# the average, the sum, the shortest string, the longest string, the number of numeric values,
	// Licensed under a CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
	// http://creativecommons.org/publicdomain/zero/1.0/

	(function (win,doc) {
	'use strict';
	if (!win.FileReader \|\| !win.addEventListener \|\| !doc.querySelectorAll) {
	// doesn't cut the mustard.
	return;
	}
	doc.querySelectorAll('input[type="file"][accept="image/*"]').forEach( function (fileInput) {
	<!DOCTYPE html>
	<html lang="en">
	<head>
	<meta charset="UTF-8">
	<title>Single File Web Component</title>
	</head>
	<body>
	<template id=single-file>
	<style>
	h1 {
	from starlette.applications import Starlette
	from starlette.responses import StreamingResponse
	from starlette.requests import Request
	from starlette.routing import Route
	from pathlib import Path
	from typing import IO, Generator

	"""
	Stream a file, in this case an mp4 video, supporting range-requests using starlette