@plastichub/osr-ai

OSR-AI

AI service prompter

Installation

1. npm i -g @plastichub/osr-ai
2. create a file with your OpenAI key, in C:\\Users\\TeleTubbie\\.config\\osr\\config.json, with the following content

{
    "openai": {
        "see":"https://platform.openai.com/api-keys"
        "key": "some katbotisch key"
    }
}

Examples

Describe the content of a file

osr-ai chatgpt prompt --source=tests/chatgpt/pipe.txt --query="whats in here, add a link to their homepage, check it and add a summary of their activities, as markdown table"

Or

cat tests/chatgpt/pipe.txt | osr-ai chatgpt prompt  --query="whats in here, add a link to their homepage, check it and add a summary of their activities, as markdown table"

whereby pipe.txt contains

abb
bigtree
cead
cidepa
kuga
motovario
nagami
siemens
teknic

Output

Company	Homepage	Summary
abb	Homepage	ABB is a leading global technology company. They focus on electrification, robotics, and automation solutions.
bigtree	Homepage	Bigtree is a software development company specializing in web and mobile applications.
cead	Homepage	CEAD is a company that specializes in 3D printing and additive manufacturing solutions.
cidepa	Homepage	Cidepa is a manufacturer of industrial machinery and equipment.
kuga	Homepage	Kuga is an automotive company that designs and manufactures electric vehicles.
motovario	Homepage	Motovario is a global manufacturer of power transmission and motion control solutions.
nagami	Homepage	Nagami is a design studio specializing in computational design and digital fabrication.
siemens	Homepage	Siemens is a multinational conglomerate that provides a wide range of products and services in various industries.
teknic	Homepage	Teknic is a manufacturer of high-performance motion control products.

---

Example, transform code

osr-ai chatgpt prompt --source="./tests/chatgpt/constants.ts" --query="rewrite this enum as object literal,add a function to retrieve a coordinate by name" --model=gpt-4 --debug

whereby ./tests/chatgpt/constants.ts contains

export enum SpainCities {
    Madrid = "40.4168,-3.7038,12",
    Barcelona = "41.3851,2.1734,12",
    Valencia = "39.4699,-0.3763,12",
    Seville = "37.3891,-5.9845,12",
    Malaga = "36.7213,-4.4214,12",
    Bilbao = "43.2630,-2.9340,12",
    Alicante = "38.3452,-0.4810,12",
    Zaragoza = "41.6488,-0.8891,12",
    Palma = "39.5696,2.6502,12",
    Murcia = "37.9922,-1.1307,12",
    Granada = "37.1773,-3.5986,12",
    Cordoba = "37.8882,-4.7794,12",
    Valladolid = "41.6520,-4.7286,12",
    Vigo = "42.2406,-8.7207,12",
    Gijon = "43.5322,-5.6611,12",
    Lleida = "41.6176,0.6200,12",
    Tarragona = "41.1189,1.2445,12",
    Cadiz = "36.5297,-6.2924,12",
    Almeria = "36.8340,-2.4637,12",
    Oviedo = "43.3614,-5.8593,12",
    Badajoz = "38.8794,-6.9706,12",
    Santander = "43.4623,-3.8090,12",
    Logrono = "42.4627,-2.4441,12",
    Salamanca = "40.9701,-5.6635,12",
    Pamplona = "42.8125,-1.6458,12",
}

Output

export const BarcelonaTowns = {
    Barcelona: "41.3850639,2.1734035",
    Badalona: "41.453966,2.24487",
    "L'HospitaletDeLlobregat": "41.3618046,2.0977132",
    Terrassa: "41.5654824,2.0109316",
    Sabadell: "41.5441063,2.1068342",
    Mataro: "41.540977,2.4440367",
    SantaColomaDeGramenet: "41.4679772,2.2081475",
    CornellaDeLlobregat: "41.3666298,2.0971758",
    SantCugatDelValles: "41.4888889,2.1019444",
    Castelldefels: "41.2755556,1.995",
    CerdanyolaDelValles: "41.4913889,2.1169444",
    Granollers: "41.6072222,2.2686111",
    VilanovaIGeltru: "41.2155556,1.7383333",
    Viladecans: "41.3163889,2.0052778",
    ElPratDeLlobregat: "41.2971628,2.0700942",
    MolletDelValles: "41.5272964,2.1870736",
    Gava: "41.3166667,2.0666667",
    SantBoiDeLlobregat: "41.3669444,2.0569444",
}

Example, transcribe an audio file, to text

osr-ai chatgpt transcribe --debug --source="./tests/chatgpt/real.mp3" --dst="./tests/chatgpt/real.txt"

Todos :)

ChatGPT

• [x] model select
• [ ] loaders
• [ ] language system defaults
• [ ] output templates
• [ ] auto-functions, eg "create file , ... "
• [ ] support Urls, for source, and stdin

OSR Todos

• [ ] Loader (mime/glob) -> storage (mem, mongo,...) -> query -> cache
• [ ] session cache (chat history replay)

References

OpenAI

• Docs

LLM Tooling

• langchain - loaders

Misc

• Voice control for VScode

Images

• ml-mgie (modifier)