Add hot-reload capability
adds watchfiles dependency and changes uvicorn server logic
but does not have any breaking changes.
---------
Co-authored-by: Eric Gustin <eric@arcade.dev>
Co-authored-by: sdreyer <sterling@arcade-ai.com>
Co-authored-by: Eric Gustin <34000337+EricGustin@users.noreply.github.com>
This PR fixes two bugs:
1. `serve_default_worker` was not registering toolkits to the worker
(non mcp). This caused workers created during `arcade workerup` and
`arcade serve` to have 0 tools.
2. We were logging the number of files in the toolkit as the number of
tools in the toolkit.
MCP stdio Implementation:
The PR adds support for standard input/output (stdio) as a transport
mechanism for the Message Control Protocol. This is a replacement to the
SSE (Server-Sent Events) transport that was worked on in PR #359 but
will not be merged as it's not deprecated.
This will allow developers to use Arcade tools (written by the dev or
Arcade) in Claude, Cursor, windsurf, etc.
The engine Gateway already supports adding HTTPS streamable (replacement
for SSE) MCP servers as tool servers, and will soon support full gateway
capability in the client API as well.
To use any existing Toolkit just
## Examples
### Quickstart setup with existing toolkits
```bash
pip install arcade-ai
pip install <name of toolkit> # ex. arcade-google
arcade serve --mcp
```
### Run with Claude
Just add the following to the Claude config
```json
{
"mcpServers": {
"arcade": {
"command": "bash",
"args": ["-c", "export ARCADE_API_KEY=arc_xxxx && /path/to/python /path/to/arcade serve --mcp"]
}
}
}
```
### Customizing the Tool Server
Developers can customize their served tools and server furthermore by
importing the worker sdk
```python
import arcade_google # pip install arcade_google
import arcade_search # pip install arcade_search
from arcade.core.catalog import ToolCatalog
from arcade.worker.mcp.stdio import StdioServer
# 2. Create and populate the tool catalog
catalog = ToolCatalog()
catalog.add_module(arcade_google) # Registers all tools in the package
catalog.add_module(arcade_search)
# 3. Main entrypoint
async def main():
# Create the worker with the tool catalog
worker = StdioServer(catalog)
# Run the worker
await worker.run()
if __name__ == "__main__":
import asyncio
asyncio.run(main())
```
Then to run with claude, just run this python file instead of the prebuilt server used in ``arcade serve --mcp``
* `arcade dashboard` opens the Arcade Dashboard in a web browser.
Defaults to `https://api.arcade.dev/dashboard`, but is configurable via
flags.
* `arcade dashboard --local` opens your locally hosted Arcade Dashboard
in a web browser.
* Performs a health check of the engine and will print a warning to the
console if the Engine is not healthy / not running.
----------------------------------------
* Inspiration from https://minikube.sigs.k8s.io/docs/handbook/dashboard/
| Name | Description |
|------------------------------|--------------------------------------------------------------------------------|
| Microsoft.CreateDraftEmail | Compose a new draft email in Outlook |
| Microsoft.UpdateDraftEmail | Update an existing draft email in Outlook
|
| Microsoft.CreateAndSendEmail | Create and immediately send a new email
in Outlook to the specified recipients |
| Microsoft.SendDraftEmail | Send an existing draft email in Outlook |
| Microsoft.ReplyToEmail | Reply only to the sender of an existing email
in Outlook |
| Microsoft.ReplyAllToEmail | Reply to all recipients of an existing
email in Outlook |
| Microsoft.ListEmails | List emails in the user's mailbox across all
folders |
| Microsoft.ListEmailsInFolder | List the user's emails in the specified
folder |
⚠️ Local Worker will not work with Slack toolkit until this is merged
Support [PEP 604](https://peps.python.org/pep-0604/) Union Types for
tool input param types and return types.
https://github.com/ArcadeAI/arcade-ai/pull/345 introduced PEP 604 Union
Types into our toolkits, but the worker did not support this syntax, so
it was failing on start up if a tool used bar syntax in its return type
(Slack).
Additionally, optionals defined with `Union[T, None]` were already
supported, but didn't have any unit tests, so I added them.
Addresses general improvements to all toolkits including changing ruff
from python 3.9 to python 3.10 which is the reason for the removal of
Optional[] among others.
Also, turns out that our `make install` for toolkits wasn't correctly
checking for whether poetry was installed (&> /dev/null syntax isn't
supported by our check-toolkits GitHub action, so we were installing
poetry twice. I replaced with the more portable >/dev/null 2>&1)
Question: Should we also change ruff to py310 for the `arcade/` package
in a later PR?
-------------------
CU-86b4gzyp6
The tool used in the mcp demo isn't a valid ToolDefinition.
Unable to list tool 'Demo.lorem': 1 validation error for ToolDefinition
requirements Input should be a valid dictionary or
instance of ToolRequirements For further information visit
https://errors.pydantic.dev/2.10/v/model_type
Some tools have date-related arguments where LLMs fail to provide the
correct values if the user asks in a relative way (e.g. next Monday,
last month, etc). By including today's date and day of the week in the
`arcade chat` system prompt, we avoid that issue.
Upon Worker startup, when an exception raised in a toolkit at import
time, the worker logged output will not display the cause of the
exception, or even that there was an exception raised in the toolkit.
Instead, the worker logs only a not very useful message:
```
❌ Failed to start Arcade Worker: Could not find tool {tool_name} in module {toolkit_name}.tools.{module_name}
```
This PR improves the logged message by adding the cause of the error.
Arcade tools can rename parameters like so,
```py
@tool()
def func_with_renamed_param(
param1: Annotated[str, "MyRenamedParam", "The first parameter"],
):
pass
```
but there is no check for whether the renamed parameter is a valid
identifier.
Anthropic models, for example, will fail if a renamed parameter is not a
valid identifier (which was the cause for
https://github.com/ArcadeAI/arcade-ai/pull/319).
## PR Description
### 1. Bug Fix
A bug was observed where tools that were using [parameterized
generics](https://docs.python.org/3/library/stdtypes.html#types-genericalias)
(e.g., `dict[str, Any]`) for their input parameters or output, then this
would cause the Worker to fail on startup with the error `issubclass()
arg 1 must be a class` for Python3.13+. This error would not occur for
Python versions less than 3.13.
In Python <3.13, parameterized generics would implicitly be treated as
their underlying type (its origin), so it was possible to treat
`dict[str, Any]` as a class and pass it to `issubclass`. This is not the
case for Python 3.13, so we need to explicitly strip the GenericAlias
(e.g., `dict[str, Any]`) down to its origin (e.g., `dict`), before
checking if it is a subclass of `Enum`.
### 2. Better Error Message
When a tool used an unsupported parameter/output type, then Arcade would
display the following message:
```
Failed to start Arcade Worker:
Type error encountered while adding tool get_website_map from arcade_web.tools.firecrawl.
Reason: issubclass() arg 1 must be a class
```
But now it displays
```
Failed to start Arcade Worker:
Error encountered while adding tool get_website_map from arcade_web.tools.firecrawl.
Reason: Unsupported type: tuple[str, str]
```
SDK support for tool secrets (stored and managed by the engine):
- [x] New `requires_secrets=` option in the `@tool` decorator
- [x] Internal plumbing in the catalog and `ToolContext`
- [x] Full test coverage of all added code
- [x] Bumped minor version (new feature)
This PR can be merged without waiting for Engine changes, because it is
additive only (no breaking changes).
After this is merged, I will open another PR to update existing toolkits
that will benefit from this feature!
## PR Description
Add the ability to mark a tool as deprecated and display the warning in
the user's runtime. This PR also lays the foundation for future work for
emitting other levels of logs (debug, info, etc) that occur during the
tool's execution.
NOTE: Updates to the Arcade Clients (Python and JS) still need to be
done before the deprecation warning is emitted, but this PR needs to be
merged before those updates!
Let's cross our fingers that we'll never need to deprecate
`@tool.deprecated`!
### Example
1. Mark your tool as deprecated
```python
from typing import Annotated
from arcade.sdk import tool
@tool.deprecated("Use the 'Math.AddInt' tool instead.") # order of decorators does not matter
@tool
def add(
a: Annotated[int, "The first number"], b: Annotated[int, "The second number"]
) -> Annotated[int, "The sum of the two numbers"]:
"""
Add two numbers together
"""
return a + b
```
2. Call the deprecated tool
```python
from arcadepy import Arcade
client = Arcade()
tool_input = {"a": 9001, "b": 42}
response = client.tools.execute(
tool_name="Math.Add",
input=tool_input,
user_id="me@example.com",
)
print(f"The result of adding {tool_input['a']} and {tool_input['b']} is: {response.output.value}")
```
3. Observe the DeprecationWarning:
```
❯ python examples/call_a_tool_directly.py
/Users/ericgustin/repos/Team/arcade-ai/examples/call_a_tool_directly.py:22: DeprecationWarning: 'Math.Add' is deprecated: Use the `Math.AddInt` tool instead.
response = client.tools.execute(
The result of adding 9001 and 42 is: 9043
```
## PR Description
Changes `pip install 'arcade-ai[fastapi]'` to `pip install arcade-ai`.
In other words, FastAPI is now a required dependecy of arcade-ai.
Additionally, I snuck in some minor cleanup changes.
# PR Description
This PR adds ~~four~~ three improvements to evals.
~~## 1. Add parameterized eval cases~~
~~Adds a new method named `add_parameterized_case`. Just like pytest’s
parameterized tests, eval cases can be parameterized with multiple user
messages. Adds a case to the `EvalSuite` for each user message. All
cases have the same expected tool call(s), params, additional_messages.
This reduces duplicate code and makes it easy to observe how a model
performs based on increasingly more difficult prompts.~~
```python
""" NO LONGER IN THIS PR
user_messages = [
"Call the delete tweet by id tool with the tweet ID '148975632'.",
"Delete the tweet with ID '148975632'.",
"I don't want to have this tweet (148975632) on my account anymore.",
"do the opposite of post for https://x.com/x/status/148975632",
]
suite.add_parameterized_case(
name="Delete a tweet by ID",
user_messages=user_messages,
expected_tool_calls=[
ExpectedToolCall(
func=delete_tweet_by_id,
args={"tweet_id": "148975632"},
)
],
critics=[
BinaryCritic(
critic_field="tweet_id",
weight=1.0,
),
],
)
"""
```
~~PASSED Delete a tweet by ID (user_message 1 of 4) -- Score: 100.00%~~
~~PASSED Delete a tweet by ID (user_message 2 of 4) -- Score: 100.00%~~
~~PASSED Delete a tweet by ID (user_message 3 of 4) -- Score: 100.00%~~
~~FAILED Delete a tweet by ID (user_message 4 of 4) -- Score: 0.00%~~
~~Summary -- Total: 4 -- Passed: 3 -- Failed: 1~~
## 2. Parameters that are not explicitly criticized are assigned a
`NoneCritic`.
A NoneCritic has no effect on the evaluation results and does not
actually evaluate. Parameters that have a NoneCritic will be displayed
as ‘un-criticized’ in the evaluation summary (if `-d` flag is used).
## 3. Add a hardcoded `seed` parameter for evals.
The seed parameter aides in receiving (mostly) consistent outputs -
aiding in reproducibility for evaluations.
## 4. Disallow more than one critic for the same field.
Raises a `ValueError` if more than one critic is assigned to a field.
---------
Co-authored-by: Eric Gustin <eric@arcade-ai.com>
# PR Description
* This PR updates code in `examples/` to be compatible with version
1.0.0
* This PR removes the Spotify examples since the Arcade hosted worker
doesn't currently cataloge the Spotify toolkit. We can reintroduce these
examples when it does.
* This PR performs various renames across the codebase for
`arcade-ai.com` --> `arcade.dev` and `Arcade AI` --> `Arcade`
