claudette

Claudette is Claude’s friend

News June 22: Claudette just launched! For a quick intro read the Claudette launch blog post.

NB: If you are reading this in GitHub’s readme, we recommend you instead read the much more nicely formatted documentation format of this tutorial.

Claudette is a wrapper for Anthropic’s Python SDK.

The SDK works well, but it is quite low level – it leaves the developer to do a lot of stuff manually. That’s a lot of extra work and boilerplate! Claudette automates pretty much everything that can be automated, whilst providing full control. Amongst the features provided:

• A Chat class that creates stateful dialogs
• Support for prefill, which tells Claude what to use as the first few words of its response
• Convenient image support
• Simple and convenient support for Claude’s new Tool Use API.

You’ll need to set the ANTHROPIC_API_KEY environment variable to the key provided to you by Anthropic in order to use this library.

Note that this library is the first ever “literate nbdev” project. That means that the actual source code for the library is a rendered Jupyter Notebook which includes callout notes and tips, HTML tables and images, detailed explanations, and teaches how and why the code is written the way it is. Even if you’ve never used the Anthropic Python SDK or Claude API before, you should be able to read the source code. Click Claudette’s Source to read it, or clone the git repo and execute the notebook yourself to see every step of the creation process in action. The tutorial below includes links to API details which will take you to relevant parts of the source. The reason this project is a new kind of literal program is because we take seriously Knuth’s call to action, that we have a “moral commitment” to never write an “illiterate program” – and so we have a commitment to making literate programming and easy and pleasant experience. (For more on this, see this talk from Hamel Husain.)

“Let us change our traditional attitude to the construction of programs: Instead of imagining that our main task is to instruct a computer what to do, let us concentrate rather on explaining to human beings what we want a computer to do.” Donald E. Knuth, Literate Programming (1984)

Install

pip install claudette

Getting started

Anthropic’s Python SDK will automatically be installed with Claudette, if you don’t already have it.

import os
# os.environ['ANTHROPIC_LOG'] = 'debug'

To print every HTTP request and response in full, uncomment the above line.

from claudette import *

Claudette only exports the symbols that are needed to use the library, so you can use import * to import them. Alternatively, just use:

import claudette

…and then add the prefix claudette. to any usages of the module.

Claudette provides models, which is a list of models currently available from the SDK.

models

('claude-3-opus-20240229',
 'claude-3-5-sonnet-20240620',
 'claude-3-haiku-20240307')

For these examples, we’ll use Sonnet 3.5, since it’s awesome!

model = models[1]

Chat

The main interface to Claudette is the Chat class, which provides a stateful interface to Claude:

chat = Chat(model, sp="""You are a helpful and concise assistant.""")
chat("I'm Jeremy")

Hello Jeremy, it’s nice to meet you. How can I assist you today?

• id: msg_01KLrPzN79dXmDhTH2Nv6tjn
• content: [{‘text’: “Hello Jeremy, it’s nice to meet you. How can I assist you today?”, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 19, ‘output_tokens’: 20}

r = chat("What's my name?")
r

Your name is Jeremy, as you just told me.

• id: msg_01VsZr9D5ptbWrggdX28y1J5
• content: [{‘text’: ‘Your name is Jeremy, as you just told me.’, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 47, ‘output_tokens’: 14}

As you see above, displaying the results of a call in a notebook shows just the message contents, with the other details hidden behind a collapsible section. Alternatively you can print the details:

print(r)

Message(id='msg_01VsZr9D5ptbWrggdX28y1J5', content=[TextBlock(text='Your name is Jeremy, as you just told me.', type='text')], model='claude-3-5-sonnet-20240620', role='assistant', stop_reason='end_turn', stop_sequence=None, type='message', usage=In: 47; Out: 14; Total: 61)

Claude supports adding an extra assistant message at the end, which contains the prefill – i.e. the text we want Claude to assume the response starts with. Let’s try it out:

chat("Concisely, what is the meaning of life?",
     prefill='According to Douglas Adams,')

According to Douglas Adams, “42.” More seriously, it’s often considered to be finding personal fulfillment, happiness, and purpose.

• id: msg_01RMBD95RyQjbsNckGYmRMLu
• content: [{‘text’: ‘According to Douglas Adams, “42.” More seriously, it's often considered to be finding personal fulfillment, happiness, and purpose.’, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 81, ‘output_tokens’: 27}

You can add stream=True to stream the results as soon as they arrive (although you will only see the gradual generation if you execute the notebook yourself, of course!)

for o in chat("Concisely, what book was that in?", prefill='It was in', stream=True):
    print(o, end='')

It was in "The Hitchhiker's Guide to the Galaxy" by Douglas Adams.

Tool use

Tool use lets Claude use external tools.

We use docments to make defining Python functions as ergonomic as possible. Each parameter (and the return value) should have a type, and a docments comment with the description of what it is. As an example we’ll write a simple function that adds numbers together, and will tell us when it’s being called:

def sums(
    a:int,  # First thing to sum
    b:int=1 # Second thing to sum
) -> int: # The sum of the inputs
    "Adds a + b."
    print(f"Finding the sum of {a} and {b}")
    return a + b

Sometimes Claude will say something like “according to the sums tool the answer is” – generally we’d rather it just tells the user the answer, so we can use a system prompt to help with this:

sp = "Never mention what tools you use."

We’ll get Claude to add up some long numbers:

a,b = 604542,6458932
pr = f"What is {a}+{b}?"
pr

'What is 604542+6458932?'

To use tools, pass a list of them to Chat, and to force it to always answer using a tool, set tool_choice to that function name:

chat = Chat(model, sp=sp, tools=[sums], tool_choice='sums')

Now when we call that with our prompt, Claude doesn’t return the answer, but instead returns a tool_use message, which means we have to call the named tool with the provided parameters:

r = chat(pr)
r

Finding the sum of 604542 and 6458932

ToolUseBlock(id=‘toolu_018a7fT3Mnoa2XqtAMXqawXU’, input={‘a’: 604542, ‘b’: 6458932}, name=‘sums’, type=‘tool_use’)

• id: msg_011r1ZTjJSLFCJKceMoFSPky
• content: [{‘id’: ‘toolu_018a7fT3Mnoa2XqtAMXqawXU’, ‘input’: {‘a’: 604542, ‘b’: 6458932}, ‘name’: ‘sums’, ‘type’: ‘tool_use’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: tool_use
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 414, ‘output_tokens’: 53}

Claudette handles all that for us – we just call it again, and it all happens automatically:

chat()

Finding the sum of 7063474 and 1

ToolUseBlock(id=‘toolu_01Pc9N6fQSUNnuqrUKR2DkmL’, input={‘a’: 7063474}, name=‘sums’, type=‘tool_use’)

• id: msg_01Ct8URrPt5hf5mjUSndwigS
• content: [{‘id’: ‘toolu_01Pc9N6fQSUNnuqrUKR2DkmL’, ‘input’: {‘a’: 7063474}, ‘name’: ‘sums’, ‘type’: ‘tool_use’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: tool_use
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 502, ‘output_tokens’: 35}

You can see how many tokens have been used at any time by checking the use property. Note that (as of May 2024) tool use in Claude uses a lot of tokens, since it automatically adds a large system prompt.

chat.use

In: 916; Out: 88; Total: 1004

We can do everything needed to use tools in a single step, by using Chat.toolloop. This can even call multiple tools as needed solve a problem. For example, let’s define a tool to handle multiplication:

def mults(
    a:int,  # First thing to multiply
    b:int=1 # Second thing to multiply
) -> int: # The product of the inputs
    "Multiplies a * b."
    print(f"Finding the product of {a} and {b}")
    return a * b

Now with a single call we can calculate (a+b)*2 – by passing show_trace we can see each response from Claude in the process:

chat = Chat(model, sp=sp, tools=[sums,mults])
pr = f'Calculate ({a}+{b})*2'
pr

'Calculate (604542+6458932)*2'

chat.toolloop(pr, trace_func=print)

Finding the sum of 604542 and 6458932
Message(id='msg_016tRhA75RqXTkCMR9WoZvyE', content=[TextBlock(text='Certainly! To calculate (604542+6458932)*2, we\'ll need to use the available tools to perform the addition and multiplication operations. Let\'s break it down step by step:\n\n1. First, we\'ll add 604542 and 6458932 using the "sums" function.\n2. Then, we\'ll multiply the result by 2 using the "mults" function.\n\nLet\'s start with the addition:', type='text'), ToolUseBlock(id='toolu_017aomxeCUA6DxM9U6UQjMu2', input={'a': 604542, 'b': 6458932}, name='sums', type='tool_use')], model='claude-3-5-sonnet-20240620', role='assistant', stop_reason='tool_use', stop_sequence=None, type='message', usage=In: 538; Out: 168; Total: 706)
Finding the product of 7063474 and 2
Message(id='msg_016w97cNedVktfhHUoL2Ey7a', content=[TextBlock(text="Great! The sum of 604542 and 6458932 is 7063474.\n\nNow, let's multiply this result by 2:", type='text'), ToolUseBlock(id='toolu_0115UMmS2GMeej1CafKLKvBB', input={'a': 7063474, 'b': 2}, name='mults', type='tool_use')], model='claude-3-5-sonnet-20240620', role='assistant', stop_reason='tool_use', stop_sequence=None, type='message', usage=In: 721; Out: 106; Total: 827)
Message(id='msg_0139BkRu9b9yKPp8N8NHBMpj', content=[TextBlock(text='Now we have our final result. \n\nThe calculation (604542+6458932)*2 equals 14126948.\n\nTo break it down:\n1. 604542 + 6458932 = 7063474\n2. 7063474 * 2 = 14126948\n\nSo, the final answer to (604542+6458932)*2 is 14126948.', type='text')], model='claude-3-5-sonnet-20240620', role='assistant', stop_reason='end_turn', stop_sequence=None, type='message', usage=In: 841; Out: 95; Total: 936)

Now we have our final result.

The calculation (604542+6458932)*2 equals 14126948.

To break it down: 1. 604542 + 6458932 = 7063474 2. 7063474 * 2 = 14126948

So, the final answer to (604542+6458932)*2 is 14126948.

• id: msg_0139BkRu9b9yKPp8N8NHBMpj
• content: [{‘text’: ’Now we have our final result. calculation (604542+6458932)2 equals 14126948.break it down:. 604542 + 6458932 = 7063474. 7063474 2 = 14126948, the final answer to (604542+6458932)*2 is 14126948.’, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 841, ‘output_tokens’: 95}

Images

Claude can handle image data as well. As everyone knows, when testing image APIs you have to use a cute puppy.

fn = Path('samples/puppy.jpg')
display.Image(filename=fn, width=200)

We create a Chat object as before:

chat = Chat(model)

Claudette expects images as a list of bytes, so we read in the file:

img = fn.read_bytes()

Prompts to Claudette can be lists, containing text, images, or both, eg:

chat([img, "In brief, what color flowers are in this image?"])

The flowers in this image are purple. They appear to be small, daisy-like flowers, possibly asters or some type of purple wildflower, blooming in the background behind the adorable puppy in the foreground.

• id: msg_01EsC5ZMpo5chJRzcjkceLu9
• content: [{‘text’: ‘The flowers in this image are purple. They appear to be small, daisy-like flowers, possibly asters or some type of purple wildflower, blooming in the background behind the adorable puppy in the foreground.’, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 110, ‘output_tokens’: 51}

The image is included as input tokens.

chat.use

In: 110; Out: 51; Total: 161

Alternatively, Claudette supports creating a multi-stage chat with separate image and text prompts. For instance, you can pass just the image as the initial prompt (in which case Claude will make some general comments about what it sees), and then follow up with questions in additional prompts:

chat = Chat(model)
chat(img)

This image shows an adorable puppy lying in the grass. The puppy appears to be a Cavalier King Charles Spaniel or a similar breed, with distinctive white and reddish-brown fur coloring. Its face is predominantly white with large, expressive dark eyes and a small black nose.

The puppy is resting on a grassy surface, giving the impression of being outdoors. In the background, you can see purple flowers, which look like asters or michaelmas daisies, adding a lovely splash of color to the scene. There’s also what appears to be a wooden structure or fence behind the flowers, suggesting a garden setting.

The composition of the image creates a charming, pastoral scene that highlights the puppy’s cuteness against the natural backdrop. The contrast between the soft fur of the puppy and the delicate purple flowers makes for a visually appealing and heartwarming picture.

• id: msg_013TMWFVwfmYh2qzUtUqen68
• content: [{‘text’: “This image shows an adorable puppy lying in the grass. The puppy appears to be a Cavalier King Charles Spaniel or a similar breed, with distinctive white and reddish-brown fur coloring. Its face is predominantly white with large, expressive dark eyes and a small black nose.puppy is resting on a grassy surface, giving the impression of being outdoors. In the background, you can see purple flowers, which look like asters or michaelmas daisies, adding a lovely splash of color to the scene. There’s also what appears to be a wooden structure or fence behind the flowers, suggesting a garden setting.composition of the image creates a charming, pastoral scene that highlights the puppy’s cuteness against the natural backdrop. The contrast between the soft fur of the puppy and the delicate purple flowers makes for a visually appealing and heartwarming picture.”, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 98, ‘output_tokens’: 197}

chat('What direction is the puppy facing?')

The puppy in the image is facing towards the camera, looking directly at the viewer. Its head is slightly tilted, giving it an endearing and curious expression. The puppy’s body is angled slightly to its left (our right), with its front paws visible and resting on the grass. This positioning allows us to see the full face of the puppy, showcasing its adorable features and the distinctive color pattern of its fur.

• id: msg_015W9aQZt9JsjExJKcN72cDo
• content: [{‘text’: “The puppy in the image is facing towards the camera, looking directly at the viewer. Its head is slightly tilted, giving it an endearing and curious expression. The puppy’s body is angled slightly to its left (our right), with its front paws visible and resting on the grass. This positioning allows us to see the full face of the puppy, showcasing its adorable features and the distinctive color pattern of its fur.”, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 306, ‘output_tokens’: 95}

chat('What color is it?')

The puppy in the image has a combination of two main colors:

1. White: The majority of its face, including the muzzle area and a portion of its body that’s visible, is white.

2. Reddish-brown (often called “chestnut” or “ruby” in this breed): This color appears on the ears and extends to patches on the body, though the full extent isn’t visible in this image.

This color combination is typical for Cavalier King Charles Spaniels, particularly the Blenheim variety. The contrast between the white and the reddish-brown fur creates a striking and adorable appearance, highlighting the puppy’s facial features and giving it a distinctive look.

• id: msg_0146KQcPjSoXFGLjqjpzwU3Q
• content: [{‘text’: ‘The puppy in the image has a combination of two main colors:. White: The majority of its face, including the muzzle area and a portion of its body that's visible, is white.. Reddish-brown (often called “chestnut” or “ruby” in this breed): This color appears on the ears and extends to patches on the body, though the full extent isn't visible in this image.color combination is typical for Cavalier King Charles Spaniels, particularly the Blenheim variety. The contrast between the white and the reddish-brown fur creates a striking and adorable appearance, highlighting the puppy's facial features and giving it a distinctive look.’, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 409, ‘output_tokens’: 157}

Note that the image is passed in again for every input in the dialog, so that number of input tokens increases quickly with this kind of chat.

chat.use

In: 813; Out: 449; Total: 1262

Other model providers

You can also use 3rd party providers of Anthropic models, as shown here.

Amazon Bedrock

These are the models available through Bedrock:

models_aws

('anthropic.claude-3-haiku-20240307-v1:0',
 'anthropic.claude-3-sonnet-20240229-v1:0',
 'anthropic.claude-3-opus-20240229-v1:0',
 'anthropic.claude-3-5-sonnet-20240620-v1:0')

To use them, call AnthropicBedrock with your access details, and pass that to Client:

from anthropic import AnthropicBedrock

ab = AnthropicBedrock(
    aws_access_key=os.environ['AWS_ACCESS_KEY'],
    aws_secret_key=os.environ['AWS_SECRET_KEY'],
)
client = Client(models_aws[-1], ab)

Now create your Chat object passing this client to the cli parameter – and from then on, everything is identical to the previous examples.

chat = Chat(cli=client)
chat("I'm Jeremy")

Hello Jeremy! It’s nice to meet you. How can I assist you today? Is there anything specific you’d like to talk about or any questions you have?

• id: msg_bdrk_011MuMcLUba22CUttpxWuHJC
• content: [{‘text’: “Hello Jeremy! It’s nice to meet you. How can I assist you today? Is there anything specific you’d like to talk about or any questions you have?”, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 10, ‘output_tokens’: 36}

Google Vertex

These are the models available through Vertex:

models_goog

('claude-3-haiku@20240307',
 'claude-3-sonnet@20240229',
 'claude-3-opus@20240229',
 'claude-3-5-sonnet@20240620')

To use them, call AnthropicVertex with your access details, and pass that to Client:

from anthropic import AnthropicVertex
import google.auth

project_id = google.auth.default()[1]
gv = AnthropicVertex(project_id=project_id, region="us-east5")
client = Client(models_goog[-1], gv)

chat = Chat(cli=client)
chat("I'm Jeremy")

Hello Jeremy! It’s nice to meet you. How can I assist you today? Is there anything specific you’d like to talk about or any questions you have?

• id: msg_vrtx_015zVvSZxfmNQ8FRZjb9zPdJ
• content: [{‘text’: “Hello Jeremy! It’s nice to meet you. How can I assist you today? Is there anything specific you’d like to talk about or any questions you have?”, ‘type’: ‘text’}]
• model: claude-3-5-sonnet-20240620
• role: assistant
• stop_reason: end_turn
• stop_sequence: None
• type: message
• usage: {‘input_tokens’: 10, ‘output_tokens’: 36}