Skip to content

Commit

Permalink
Adds %ai list, %ai help magic commands (jupyterlab#100)
Browse files Browse the repository at this point in the history 
* Initial help command support

* Adds help command

* Matches error message

* Styles model ID in error

* Introduces TextWithMarkdown class

* Removes incorrect comment

* Alphabetizes classes

* List of models

* Adds model list

* Documents %ai list

* Optional parameter

* Updates example notebook

* Documents parameter

* In Markdown, displays a table of models

* Adds column for env vars

* Uses emoji in its own column

* Renames class to TextOrMarkdown

* Removes dictionary of env vars
  • Loading branch information
@JasonWeill
JasonWeill authored Apr 25, 2023
1 parent 19f8b8f commit c803079
Show file tree
Hide file tree
Showing 4 changed files with 430 additions and 4 deletions.
9 changes: 9 additions & 0 deletions docs/source/users/index.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -238,6 +238,15 @@ providers, as defined in [`langchain.llms`](https://langchain.readthedocs.io/en/
- `openai-chat`
- `sagemaker-endpoint`

You can find a list of supported providers and models by running `%ai list`. Some providers
define a list of supported models. If a provider does not define a list of supported models,
consult the vendor's documentation. The [Hugging Face web site](https://huggingface.co/)
includes a list of models, for example.

Optionally, you can pass a provider ID as a parameter to `%ai list` to get all
models provided by one provider. For example, `%ai list openai` will display only models
provided by the `openai` provider.

If your model ID is associated with only one provider, you can omit the `provider-id` and
the colon from the first line. For example, because `ai21` is the only provider of the
`j2-jumbo-instruct` model, you can either give the full provider and model,
Expand Down
250 changes: 250 additions & 0 deletions examples/commands.ipynb
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,250 @@
{
"cells": [
{
"cell_type": "markdown",
"id": "ca3ba02c-c965-4ffe-88b3-ca606a9cf288",
"metadata": {},
"source": [
"# AI commands\n",
"\n",
"You can use the `%ai` cell magic to run special commands. If a model has the same name as a command, you will need to refer to the model using the `provider:model` format only; the `model` shortcut will not work."
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "6c9c4016-711c-4592-b543-24b854f5b160",
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"%reload_ext jupyter_ai"
]
},
{
"cell_type": "markdown",
"id": "cd2a9967-8923-43bf-96b0-9567c09775e2",
"metadata": {
"tags": []
},
"source": [
"## Getting help\n",
"\n",
"`%ai help` will show you a list of all supported commands."
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "edd6ddd5-54fb-4750-8ae9-4143a6b83e2e",
"metadata": {
"tags": []
},
"outputs": [
{
"data": {
"text/markdown": [
"| Command | Description |\n",
"| ------- | ----------- |\n",
"| `help` | Display a list of supported commands|\n",
"| `list` | Display a list of models that you can use (optionally, for a single provider)|\n"
],
"text/plain": [
"help - Display a list of supported commands\n",
"list - Display a list of models that you can use (optionally, for a single provider)\n"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"%ai help"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "0f073caa-265d-40d6-b537-d025b8df9f41",
"metadata": {},
"outputs": [
{
"data": {
"text/markdown": [
"Cannot determine model provider from model ID `foo`.\n",
"\n",
"To see a list of models you can use, run `%ai list`.\n",
"\n",
"If you were trying to run a command, run `%ai help` to see a list of commands."
],
"text/plain": [
"Cannot determine model provider from model ID 'foo'.\n",
"\n",
"To see a list of models you can use, run '%ai list'.\n",
"\n",
"If you were trying to run a command, run '%ai help' to see a list of commands."
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"%ai foo"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "bad2d8a8-6141-4247-9af7-7583426c59a6",
"metadata": {},
"outputs": [
{
"data": {
"text/markdown": [
"| Provider | Environment variable | Set? | Models |\n",
"|----------|----------------------|------|--------|\n",
"| `ai21` | `AI21_API_KEY` | <abbr title=\"You have set this environment variable, so you can use this provider's models.\">✅</abbr> | `ai21:j1-large`, `ai21:j1-grande`, `ai21:j1-jumbo`, `ai21:j1-grande-instruct`, `ai21:j2-large`, `ai21:j2-grande`, `ai21:j2-jumbo`, `ai21:j2-grande-instruct`, `ai21:j2-jumbo-instruct` |\n",
"| `anthropic` | `ANTHROPIC_API_KEY` | <abbr title=\"You have set this environment variable, so you can use this provider's models.\">✅</abbr> | `anthropic:claude-v1`, `anthropic:claude-v1.0`, `anthropic:claude-v1.2`, `anthropic:claude-instant-v1`, `anthropic:claude-instant-v1.0` |\n",
"| `cohere` | `COHERE_API_KEY` | <abbr title=\"You have set this environment variable, so you can use this provider's models.\">✅</abbr> | `cohere:medium`, `cohere:xlarge` |\n",
"| `huggingface_hub` | `HUGGINGFACEHUB_API_TOKEN` | <abbr title=\"You have set this environment variable, so you can use this provider's models.\">✅</abbr> | This provider does not define a list of models. |\n",
"| `openai` | `OPENAI_API_KEY` | <abbr title=\"You have set this environment variable, so you can use this provider's models.\">✅</abbr> | `openai:text-davinci-003`, `openai:text-davinci-002`, `openai:text-curie-001`, `openai:text-babbage-001`, `openai:text-ada-001`, `openai:davinci`, `openai:curie`, `openai:babbage`, `openai:ada` |\n",
"| `openai-chat` | `OPENAI_API_KEY` | <abbr title=\"You have set this environment variable, so you can use this provider's models.\">✅</abbr> | `openai-chat:gpt-4`, `openai-chat:gpt-4-0314`, `openai-chat:gpt-4-32k`, `openai-chat:gpt-4-32k-0314`, `openai-chat:gpt-3.5-turbo`, `openai-chat:gpt-3.5-turbo-0301` |\n",
"| `sagemaker-endpoint` | Not applicable. | <abbr title=\"Not applicable\">N/A</abbr> | This provider does not define a list of models. |\n"
],
"text/plain": [
"ai21\n",
"Requires environment variable AI21_API_KEY (set)\n",
"* ai21:j1-large\n",
"* ai21:j1-grande\n",
"* ai21:j1-jumbo\n",
"* ai21:j1-grande-instruct\n",
"* ai21:j2-large\n",
"* ai21:j2-grande\n",
"* ai21:j2-jumbo\n",
"* ai21:j2-grande-instruct\n",
"* ai21:j2-jumbo-instruct\n",
"\n",
"anthropic\n",
"Requires environment variable ANTHROPIC_API_KEY (set)\n",
"* anthropic:claude-v1\n",
"* anthropic:claude-v1.0\n",
"* anthropic:claude-v1.2\n",
"* anthropic:claude-instant-v1\n",
"* anthropic:claude-instant-v1.0\n",
"\n",
"cohere\n",
"Requires environment variable COHERE_API_KEY (set)\n",
"* cohere:medium\n",
"* cohere:xlarge\n",
"\n",
"huggingface_hub\n",
"Requires environment variable HUGGINGFACEHUB_API_TOKEN (set)\n",
"* This provider does not define a list of models.\n",
"\n",
"openai\n",
"Requires environment variable OPENAI_API_KEY (set)\n",
"* openai:text-davinci-003\n",
"* openai:text-davinci-002\n",
"* openai:text-curie-001\n",
"* openai:text-babbage-001\n",
"* openai:text-ada-001\n",
"* openai:davinci\n",
"* openai:curie\n",
"* openai:babbage\n",
"* openai:ada\n",
"\n",
"openai-chat\n",
"Requires environment variable OPENAI_API_KEY (set)\n",
"* openai-chat:gpt-4\n",
"* openai-chat:gpt-4-0314\n",
"* openai-chat:gpt-4-32k\n",
"* openai-chat:gpt-4-32k-0314\n",
"* openai-chat:gpt-3.5-turbo\n",
"* openai-chat:gpt-3.5-turbo-0301\n",
"\n",
"sagemaker-endpoint\n",
"* This provider does not define a list of models.\n",
"\n"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"%ai list"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "4d84fcac-7348-4c02-9ec3-34b300ec8459",
"metadata": {},
"outputs": [
{
"data": {
"text/markdown": [
"| Provider | Environment variable | Set? | Models |\n",
"|----------|----------------------|------|--------|\n",
"| `openai` | `OPENAI_API_KEY` | <abbr title=\"You have set this environment variable, so you can use this provider's models.\">✅</abbr> | `openai:text-davinci-003`, `openai:text-davinci-002`, `openai:text-curie-001`, `openai:text-babbage-001`, `openai:text-ada-001`, `openai:davinci`, `openai:curie`, `openai:babbage`, `openai:ada` |\n"
],
"text/plain": [
"openai\n",
"Requires environment variable OPENAI_API_KEY (set)\n",
"* openai:text-davinci-003\n",
"* openai:text-davinci-002\n",
"* openai:text-curie-001\n",
"* openai:text-babbage-001\n",
"* openai:text-ada-001\n",
"* openai:davinci\n",
"* openai:curie\n",
"* openai:babbage\n",
"* openai:ada\n",
"\n"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"%ai list openai"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "b56ff0e3-42c2-4927-affd-be6a089dfa43",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.16"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Loading

0 comments on commit c803079

Please sign in to comment.