npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

@iyio/convo-openai

v0.9.6

Published

A Convo bot service using OpenAI GTP-3

Downloads

10

Readme

Convo-openai

@iyio/convo-openai has been succeeded by @iyio/convo-lang

https://www.npmjs.com/package/@iyio/convo-lang

Below are the docs for @iyio/convo-lang not this package../../iyio-common/packages/convo-lang/assets

convo-lang (a work in progress)

A conversational prompting and programming language. Convo aims to provide a uniform prompting language that is LLM agnostic and support features such as function calling and tool usage.

As a second reminder, you are viewing the docs for @iyio/convo-lang which is the successor of @iyio/convo-openai. Follow the link below to view the @iyio/convo-lang package

https://www.npmjs.com/package/@iyio/convo-lang

Packages

The Convo-lang ecosystem is contained in the following 3 packages.

  • @iyio/convo-lang - Contains the convo-lang parser, runtime and code to use convo in your application.
  • @iyio/convo-lang-cli - A CLI interface that allows you to execute and parse convo files.
  • @iyio/convo-lang-tools - Contains the convo-lang vscode extension which includes syntax highlighting, in-editor script execution, script parser and other helpful tools for working with convo-lang

Installation

For use in an application install the @iyio/convo-lang package

npm i @iyio/convo-lang

# To use OpenAI as a completion provider you will need to install the following packages as well
npm i @iyio/ai-complete 

For use on the command line install the @iyio/convo-lang-cli package

npm i @iyio/convo-lang-cli @iyio/ai-complete-openai

You will also probably want to install the vscode extension for syntax highlighting and other developer niceties. You can install the vscode extension by searching "convo-lang" in the vscode extension tab.

https://marketplace.visualstudio.com/items?itemName=IYIO.convo-lang-tools

Syntax Example

convo

Using convo-lang in an application

When using convo-lang in an application you will primarily interact with Conversation objects. Conversation objects store the messages of a convo script and allow new messages to be appended and LLMs to respond to messages form the user.

import { Conversation } from '@iyio/convo-lang';
import { initRootScope, EnvParams } from '@iyio/common';
import { aiCompleteConvoModule } from '@iyio/ai-complete';
import { openAiModule } from '@iyio/ai-complete-openai';

// initRootScope is used to configure services and configuration variables
initRootScope(reg=>{

    // register OpenAI configuration variables. These variates could also be stored as environment
    // variables and loaded using reg.addParams(new EnvParams()).
    reg.addParams({
        "openAiApiKey":"YOUR_OPEN_AI_KEY",
        "openAiChatModel":"gpt-4-1106-preview",
        "openAiVisionModel":"gpt-4-vision-preview",
        "openAiAudioModel":"whisper-1",
        "openAiImageModel":"dall-e-3",
    })

    // EnvParams can optionally be used to load configuration variables from process.env
    reg.addParams(new EnvParams());

    // Registers the AiComplete module that is used to relay messages to LLMs
    reg.use(aiCompleteConvoModule);

    // Registers the OpenAI module that will relay messages to OpenAI
    reg.use(openAiModule);

    // aiCompleteLambdaModule can be used to relay messages to a lambda function for use in the browser
    //reg.use(aiCompleteLambdaModule);
})

const main=async ()=>{
    const convo=new Conversation();

    // adding /*convo*/ before a template literal will give you convo syntax highlighting when you have
    // the convo-lang vscode extension installed.

    convo.append(/*convo*/`
        > system
        You are a friendly and very skilled fisherman. Taking a group of tourist on a fishing expedition
        off the coast of Maine.

        > user
        What kind of clothes should I wear?
    `);

    // Calling completeAsync will answer the user's question using the configured LLM
    await convo.completeAsync();


    // The convo property of the Conversation object will be updated with the answer from the LLM
    console.log(convo.convo)

    // You can get a flatted view of the conversation by calling flattenAsync. The flattend version
    // of the conversation contains messages with all templates populated and is suiteable to be 
    // used to render a view of the conversation to the user.
    const flat=await convo.flattenAsync();
    console.log('flat',flat.messages);
}

main();

Using the convo-lang extension

With the convo vscode extension install you can execute convo scripts directly in vscode. Just press (CMD+R) to run a script.

You can also run snippets of convo scripts that are embedded in other document types. Just highlight the convo code you want to run and open the command pallet and search for "Complete Convo Conversation" and press enter, then the snippet will be opened in a new convo file and completed. This is great for quick prototyping and testing prompts in your application without having to start you full application.

Using the CLI

The convo CLI can be used to execute convo scripts from the command line

# install the convo cli
npm install -g @iyio/convo-cli

# Results will be printed to stdout
convo talky-time.convo

# Results will be written to a new file named something-happened.convo
convo talky-time.convo --out something.convo

# Result will be written to the source input file. This allows you to have a continuous conversation
convo talky-time.convo --out .

There is currently only one way to configure the convo cli and vscode extension. This will be extended soon to include reading configuration files from your current workspace.

CLI and extension configuration

To allow convo to access OpenAI, create a JSON file called ~/.config/convo/convo.json and add the following contents. Remember to replace the API key with your OpenAI api key.

{
    "env":{
        "openAiApiKey":"YOUR_OPEN_AI_KEY",
        "openAiChatModel":"gpt-4-1106-preview",
        "openAiVisionModel":"gpt-4-vision-preview",
        "openAiAudioModel":"whisper-1",
        "openAiImageModel":"dall-e-3",

    }
}

Vision

When used with vision capable LLMs convo can execute vision based prompts. Images can be defined using the Markdown image syntax.

Vision capabilities must be enabled.

const convo=new Conversation({capabilities:['vision']});

Currently vision is enabled by define by the vscode extension and CLI. This may change soon.

(note - the image in the example below is the cover of Abbey Road by The Beatles)

> user
Tell me a joke about this image
![](https://liirnspace-bucksmediabucket70ce2cea-11rljmkq8tolk.s3.amazonaws.com/vaeCiOZeTyi2U1xyiOiNKgr5hwQ7A9TZqv_xqCvAlIoA)


@toolId call_QuwU683taTBW7nsZNMNHPrmm
> call queryImage(
    "query": "Tell me a joke about this image",
    "imageUrls": [
        "https://liirnspace-bucksmediabucket70ce2cea-11rljmkq8tolk.s3.amazonaws.com/vaeCiOZeTyi2U1xyiOiNKgr5hwQ7A9TZqv_xqCvAlIoA"
    ]
)
> result
__return={"result":"Why did The Beatles cross the road? Because they knew it was the only way to get to the other side of their Abbey Road album cover!"}


> assistant
Why did The Beatles cross the road? Because they knew it was the only way to get to the other side of their Abbey Road album cover!

The convo-lang syntax

The convo-lang syntax is designed to be easily readable and follows a chat like message structure. A convo-lang script is a series of messages and those messages can either be a text based message, a function or a set of top-level statements. All messages start with a header which defines the message's type.

Message Header

All message headers start with the (>) character followed by a role, keyword or identifier

Text based messages

Text based message start with a header that defines the message's role. The role can be any identifier with the exception of any top-level keyword identifiers.

// text based message
> user
message content

Top-level statement messages

Top-level statements allow you to define variables and execute statements with calling a function. Top-level statements must use one of the follow keywords as their identifier.

  • do - Used to define variables and execute statements.
  • result - Used to store the result value of a function call. Result messages are typically created by the runtime and automatically added to conversations. In most cases you will not directly create result messages.
  • define - Used to define variables and types. define messages are not allowed to call functions.
  • debug - Used to write debug information. In most cases you will not directly create debug messages.
// top-level statement
> define
someVar=77

Function messages

Function messages start with an identifier followed by a parameter definition and optionally and function body.

// function
> goFast(
    speed: int
    direction: enum("left" "right")
)
// function with a body
> goFast(
    speed: int
    direction: enum("left" "right")
) -> (
    print('I\'m going {{speed}} mile per hour and about to turn {{direction}}')

    // reportStatus would be defined outside of convo and implemented in your language of choice.
    reportStatus(speed direction)

    return(switch(gt(speed 100) "I'm going Fast" "I'm driving like a grandma"))
)

Extern function

To allow convo to communicate with the rest of your application you can register extern functions that cal be called by functions define in convo.

Comments

Comments start with either a pound character or 2 forwards slashes. Comments starting with a pound symbol are captured and available at run time as part of type definitions. Comments starting with 2 forward slashes are non capturing.

Tags

Tags start with the (@) character followed by a name and an optional value. Tags allow you to store metadata in the syntax tree of convo that is available at runtime.


// This is a non documenting comment and will not be available at run time

# This is a documenting comment and will be available at run time and will be used as the 
# description for the function below
> meFunc() -> (
    print('👋')
)

Types

Convo support simple duck typing and allows for user defined types. Custom types are created using the struct function. Each labeled parameter passed to the struct function will define a property. At runtime types are validated using the Zod. User defined types must start with an uppercase letter.

Car = struct(
    color: string
    engine: struct(
        hp: number
        torque: number
        fuel: enum( "gas" "diesel" "jp-7" )
    )
    owners: array(struct(
        name: string
        // address is optional
        address?: string
    ))
)

Variable scoping

Variables are either scoped to the function they are defined in or shared. Shared variables are visible to all functions and top level statements. Top level statements can only define shared variables. To define a shared variable in a function tag an assignment statement with the @shared tag.

JSON support

JSON style values can be used in place of the map and array functions.

// obj1 and obj2 are the same

obj1 = map(
    go: "fast"
    turn: "left"
    times: 1000
)

obj2 = {
    "go": "fast",
    "turn": "left",
    "times" 1000
}

Markdown support - (coming soon)

Text based messages in convo support a subset of the markdown syntax and the markdown structure is available at compile time.

Keywords

  • string - String type object
  • number - Number type object
  • int - Integer type object
  • time - Time type object. The time type is represented as an integer timestamp
  • void - Void type object.
  • boolean - Boolean type object.
  • any - Any time object
  • true - True constant
  • false - False constant
  • null - Null constant
  • undefined - Undefined constant
  • convo - A reference to the current conversation. This is equivalent to window in Javascript.
  • __args - A reference to the parameters passed the the current function as any object.
  • __return - A reference to the last return value of a function called by a call message
  • __error - A reference to the last error

System Functions

pipe( ...values: any )

Pipes the value of each argument received to the argument to its right.

struct( ...properties: any )

Defines a user defined type

Car = struct(
    color: string
    engine: struct(
        hp: number
        torque: number
        fuel: enum( "gas" "diesel" "jp-7" )
    )
    owners: array(struct(
        name: string
        // address is optional
        address?: string
    ))
)

enum( ...values: any )

Defines an enumeration

Size = enum( "sm" "md" "lg" )

is( ...value:any type:any )

Checks if all of the parameters left of the last parameter are of the type of the last parameter

num = 7

// true
is(num number)

// false
is(num string)

str = "lo"

// true
is(str string)

// false
is(str number)

// false
is(str num number)

// true
is(str num any)

Person = struct(
    name: string
    age: number
)

user1 = map(
    name: "Jeff"
    age: 22
)

user2 = map(
    name: "Max"
    age: 12
)

// true
is(user1 Person)

// true
is(user1 user2 Person)

// false
is(user1 user2 num Person)

map( ...properties: any )

Creates an object

// meObj has 2 properties, name and age
meObj = map(
    name: "Jeff"
    age: 22
)

array( ...values: any )

Creates an array

meAry = array( 1 2 3 "a" "b" "c" )

jsonMap( ...properties: any )

Used internally to implement JSON object syntax support. At compile time JSON objects are converted to standard convo function calls.

jsonStyle = {
    "go": "fast",
    "turn": "left",
    "times" 1000
}

convoStyle = obj1 = jsonMap(
    go: "fast"
    turn: "left"
    times: 1000
)

jsonMap( ...properties: any )

Used internally to implement JSON array syntax support. At compile time JSON arrays are converted to standard convo function calls.

jsonStyle = [ 1, 2, 3, "a", "b", "c" ]

convoStyle = array( 1 2 3 "a" "b" "c" )

and( ...values: any )

Returns true if all given arguments are truthy.


// true
and( 1 )

// false
and( 0 )

// true
and( 1 2 )

// false
and( 0 1 )

// true
and( eq(1 1) eq(2 2) )

// false
and( eq(1 1) eq(2 1) )

and( ...values: any )

Returns true if any given arguments are truthy.


// true
or( 1 )

// false
or( 0 )

// true
or( 1 2 )

// true
or( 0 1 )

// true
or( eq(1 1) eq(2 2) )

// true
or( eq(1 1) eq(2 1) )

// false
or( eq(1 3) eq(2 1) )

not( ...values: any )

Returns true if all given arguments are falsy.


// false
or( true )

// true
or( false )

// false
or( 1 )

// true
or( 0 )

// false
or( 1 2 )

// false
or( 0 1 )

// true
or( 0 false )

// false
or( eq(1 1))

// true
or( eq(1 2) )

Control flow functions

if( condition:any ), elif( condition: any ), then( ...statements )

If condition is truthy then the statement directly after the if statement will be executed otherwise the statement directly after if is skipped


age = 36

if( gte( age 21 ) ) then (
    print( "You can buy beer in the US" )
) elif (lt( age 16 )) then(
    print( "You're not even close" )
) else (
    print( '{{sub(21 age)}} years until you can buy beer in the US' )
)

while( condition:any )

While condition is truthy then the statement directly after the while statement will be executed otherwise the statement directly after if is skipped and the while loop will exit.


lap = 0

while( lt( lap 500 ) ) do (
    print("go fast")
    print("turn left")

    // increment by 1
    inc(lap)
)

foreach( iterator:any )

Executes the next statement for each item returned by the passed in iterator.

total = 0
foreach( num=in(array(1 2 3 4 )) ) do (
    total = add( num total )
)

// 10
print(total)

in( value: array(any) )

Iterates of the values of an array

break( ...values: any )

Breaks out of loops either not arguments are passed or if any of the passed arguments are truthy

lap = 0

while( true ) do (
    print("go fast")
    print("turn left")

    // increment by 1
    inc(lap)

    if( eq( lap 500 ) ) then (
        break()
    )
)

do( ...statements: any)

Executes all given statements and returns the value of the last statement. Do is commonly used with loop statements, but it can also be useful in other situations on its own such as doing inline calculations. (note) The do keyword is also used to define top level statement when do is used as a message name.

n = 0
while( lt( n 10 ) ) do (
    // increment by 1
    inc(lap)
)

// 22
print( add( 5 do(
    sum = mul(n 2)
    sum = sub( sum 3 )
)) )

return( value:any )

Returns a value from the current function


> customMath(
    a: number
    b: number
) -> (
    return mul( add(a b) b )
)

> do

value = customMath(a:4 b:3)

// 21
print(value)

eq( ...values:any )

Returns true if all given values are equal. Object equality is checked by by reference. Values must be of the same type to be equal. ( a === b )

// true
eq( 1 1 )

// false
eq( 1 "1" )

gt( a:number b:number)

Returns true if a is grater then b. ( a > b )

gte( a:number b:number)

Returns true if a is grater then or equal to b. ( a >= b )

lt( a:number b:number)

Returns true if a is less then b. ( a < b )

lte( a:number b:number)

Returns true if a is less then or equal to b. ( a <= b )

add( ...values:any )

Adds all arguments together and returns the result. Strings are concatenated. (a + b )

sub( a:number b:number )

Subtracts a from b and returns the result. (a - b )

sub( a:number b:number )

Multiplies a and b and returns the result. (a * b )

div( a:number b:number )

Divides a and b and returns the result. (a / b )

pow( a:number b:number )

Raises a by b and returns the result. Math.pow(a, b )

inc( *a:number byValue?:number )

Increments the value of the given variable by 1 or the value of the second argument. ( a++ ) or ( a+= byValue )

dec( *a:number byValue?:number )

Decrements the value of the given variable by 1 or the value of the second argument. ( a-- ) or ( a-= byValue )

print( ...values:any )

Prints all values to stdout

switch( value:any ...valueOrCase:any ), case( value:any ), test( condition:any ), default()

Switch can be used as either and switch statement or a ternary. When the switch function has exactly 3 arguments and non of the is a case or default statement then switch acts as a ternary.


// can be 0 to 9
value = rand(9)

// can be 20 to 29
value2 = add(20 rand(9))


switch(

    // Sets the current match value of the switch. The match value of a switch statement can be
    // changed further down the switch
    value

    case(0) print("Lowest")

    case(1) do(
        print("Value is 1")
    )

    case(2) do(
        print("Number two")
    )

    case(3) do(
        print("Tree or three")
    )

    // change the value to a value in ary
    value2

    case(20) do(
        print("2 zero")
    )

    test(lt(value2 28)) do(
        print("less than 28")
    )

    default() print("Fall back to default")

)


// values matched by switches are returned and the value can be assigned to a variable
str = "two"
value = switch(
    str

    case("one") 1
    case("two") 2
    case("three") 3
)

// 2
print(value)


// switches can also be used as a ternary

// yes
print( switch(true "yes" "no") )

// no
print( switch(false "yes" "no") )

sleep( milliseconds:number )

Suspends execution for the given number of milliseconds

rand( max?:int )

Returns a random number. Is the max parameters is passed then a random whole number with a maximum of max will be returned otherwise a random number from 0 to 1 will be returned.

Examples

Calling Shell Scripts

// (Caution) this script is capable to running shell command on you machine.
// (Note) Before commands are ran you will be prompted to allow access to run the command and
//        be given a preview of the command

> define
computerType="MacBook pro"

# Runs a command in a bash shell
> runCommand(
    # the command to run
    command:string
) -> (
    __cwd="/Users/scott/docs/iyio-common/packages"
    print('running...' command)
    return(exec(command))
)

> system
You are a uniux systems expert. Always use the runCommand function when responding.

The user has a {{computerType}}

> user
Create a new file called bobs-my-uncle.txt and add the text "I like fish frogs" to it

Zoo Builder

// Top level statements defining shared variables and custom user types. Top level statements
// start with (> define),  (> do) or  (> result)
> define

location = "Cincinnati, OH"

Animal = struct(

    species: string

    # Weight in pounds
    weight: int

    type: enum( "carnivore" "omnivore" "herbivore")

    # Price in USD to buy the animal
    price: number

)



// This is a function that can be called by an LLM

# Builds a new exhibit featuring a collection of animals
> buildExhibit(

    # The featured animals
    animals: array(Animal)

    # Price in USD to build the exhibit excluding the animals
    price: number

    # The size of the exhibit in square feet
    size: number

    # A short description of the exhibit
    description: string
) -> (

    print('Building new exhibit with {{animals.length}} animals')

    totalPrice = price

    foreach( a=in(animals) ) do (
        // Uncomment the line below after you provide an extern function for buyAnimal
        //buyAnimal(a)
        inc(totalPrice a.price)
    )

    return(map(
        animalsBought: animals.length
        totalPrice: totalPrice
        success: true

    ))
)

// This function is not visible to LLMs and will be used to call an extern function
// defined in Javascript
> local buyAnimal(animal:Animal)


// This is a system prompt that tells the LLM who they are
> system
You are a funny and helpful zoo keeper helping the user build a new zoo. The zoo is located
in {{location}}.



// This is a user prompt that the LLM will respond to
> user
The Bengals are going to be sponsoring the zoo this year. Lets build an exhibit they will like.
Response using the buildExhibit function


// This is a function call return by the LLM
> call buildExhibit(
    "animals": [
        {
            "species": "Bengal Tiger",
            "weight": 420,
            "type": "carnivore",
            "price": 20000
        },
        {
            "species": "White Bengal Tiger",
            "weight": 400,
            "type": "carnivore",
            "price": 25000
        },
        {
            "species": "Bengal Fox",
            "weight": 20,
            "type": "omnivore",
            "price": 5000
        }
    ],
    "price": 1500000,
    "size": 50000,
    "description": "A mesmerizing tribute to the wild beauty of Bengal, the 'Bengals' Kingdom' features stunning Bengal Tigers, including the rare and captivating White Bengal Tiger, and the cunning Bengal Fox. The Bengals' Kingdom spans 50,000 square feet, providing plenty of prowling room for the resident animals. Bengals' fans and animal lovers alike will enjoy this enchanting walk on the wild side!"
)
> result
// Value return by the call to buildExhibit
__return={"animalsBought":3,"totalPrice":1550000,"success":true}


// Response from the LLM based on the return value of buildExhibit
> assistant
Great news! We've successfully built the 'Bengals' Kingdom'!

Here's a sneak peek:

- It's home to the magnificent Bengal Tiger, the rare White Bengal Tiger, and the wily Bengal Fox.
- The exhibit is an enormous 50,000 square feet! That's gonna be a lot of tiger prowling action!
- And the total cost, including our new furry residents? A cool $1,550,000. Yes, tigers are a bit on the pricier side, but hey, they're worth every penny.

Would you like to add anything else to our new zoo?