b6b4cd0a4c
### Overview Major restructuring from monolithic `arcade-ai` package to modular library architecture with standardized uv-based dependency management.  ### New Package Structure - **`arcade-tdk`** - Lightweight toolkit development kit (core decorators, auth) - **`arcade-core`** - Core execution engine and catalog functionality - **`arcade-serve`** - FastAPI/MCP server components - **`arcade-ai`** - Meta package that includes CLI functionality. Optionally include evals via the `evals` extra. Optionally include all packages via the `all` extra. ### Key Benefits - **Lighter Dependencies**: Toolkits now depend only on `arcade-tdk` (~2 deps) vs full `arcade-ai` (~30+ deps) - **Faster Builds**: uv provides 10-100x faster dependency resolution and installation - **Better Modularity**: Clear separation of concerns, consumers import only what they need - **Standard Tooling**: Eliminates custom poetry scripts, uses standard Python packaging ### Migration Impact - All 20 toolkits converted from poetry → uv with `arcade-tdk` dependencies plus `arcade-ai[evals]` and `arcade-serve` dev dependencies. When developing locally, devs should install toolkits via `make install-local`. - Modern Python 3.10+ type hints throughout - Standardized build system with hatchling backend - Enhanced Makefile with robust toolkit management commands - Removed `arcade dev` CLI command - Reduce the number of files created by `arcade new` and add an option to not generate a tests and evals folder. This foundation enables faster development cycles and cleaner dependency chains for the growing toolkit ecosystem. ### Todo After this PR is merged - [ ] Post-merge workflow(s) (release & publish containers, etc) - [ ] Release order plan. @EricGustin suggests releasing in the following order: 1. `arcade-core` version 0.1.0 2. `arcade-serve` version 0.1.0 and `arcade-tdk` version 0.1.0 3. `arcade-ai` version 2.0.0 4. Patch release for all toolkits (all changes in toolkits are internal refactors) - [ ] [Update docs](https://github.com/ArcadeAI/docs/pull/318) --------- Co-authored-by: Eric Gustin <eric@arcade.dev> Co-authored-by: Eric Gustin <34000337+EricGustin@users.noreply.github.com>
389 lines
12 KiB
Python
389 lines
12 KiB
Python
from arcade_evals import (
|
|
EvalRubric,
|
|
EvalSuite,
|
|
ExpectedToolCall,
|
|
tool_eval,
|
|
)
|
|
from arcade_evals.critic import BinaryCritic
|
|
from arcade_tdk import ToolCatalog
|
|
|
|
import arcade_jira
|
|
from arcade_jira.critics import (
|
|
CaseInsensitiveBinaryCritic,
|
|
CaseInsensitiveListOfStringsBinaryCritic,
|
|
)
|
|
from arcade_jira.tools.issues import (
|
|
add_labels_to_issue,
|
|
create_issue,
|
|
remove_labels_from_issue,
|
|
update_issue,
|
|
)
|
|
|
|
# Evaluation rubric
|
|
rubric = EvalRubric(
|
|
fail_threshold=0.85,
|
|
warn_threshold=0.95,
|
|
)
|
|
|
|
|
|
catalog = ToolCatalog()
|
|
catalog.add_module(arcade_jira)
|
|
|
|
|
|
@tool_eval()
|
|
def create_issue_eval_suite() -> EvalSuite:
|
|
suite = EvalSuite(
|
|
name="Create issue eval suite",
|
|
system_message=(
|
|
"You are an AI assistant with access to Jira tools. "
|
|
"Use them to help the user with their tasks."
|
|
),
|
|
catalog=catalog,
|
|
rubric=rubric,
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Create issue",
|
|
user_message="Create a 'High' priority task for John Doe with the following properties: "
|
|
"title: 'Test issue', "
|
|
"description: 'This is a test issue', "
|
|
"project: 'ENG-123', "
|
|
"issue_type: 'Task', "
|
|
"due on '2025-06-30'. "
|
|
"Label it with Hello and World.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=create_issue,
|
|
args={
|
|
"title": "Test issue",
|
|
"description": "This is a test issue",
|
|
"project": "ENG-123",
|
|
"issue_type": "Task",
|
|
"priority": "High",
|
|
"assignee": "John Doe",
|
|
"due_date": "2025-06-30",
|
|
"labels": ["Hello", "World"],
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="title", weight=1 / 8),
|
|
CaseInsensitiveBinaryCritic(critic_field="description", weight=1 / 8),
|
|
CaseInsensitiveBinaryCritic(critic_field="project", weight=1 / 8),
|
|
CaseInsensitiveBinaryCritic(critic_field="issue_type", weight=1 / 8),
|
|
CaseInsensitiveBinaryCritic(critic_field="priority", weight=1 / 8),
|
|
CaseInsensitiveBinaryCritic(critic_field="assignee", weight=1 / 8),
|
|
BinaryCritic(critic_field="due_date", weight=1 / 8),
|
|
CaseInsensitiveListOfStringsBinaryCritic(critic_field="labels", weight=1 / 8),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Create issue with parent and reporter",
|
|
user_message=(
|
|
"Create a task for John Doe to 'Implement message queue service' "
|
|
"as a child of the issue ENG-321 and reported by Jenifer Bear. "
|
|
"It should be due on 2025-06-30. "
|
|
"Label it with 'Project XYZ'."
|
|
),
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=create_issue,
|
|
args={
|
|
"title": "Implement message queue service",
|
|
"parent_issue": "ENG-321",
|
|
"issue_type": "Task",
|
|
"assignee": "John Doe",
|
|
"reporter": "Jenifer Bear",
|
|
"due_date": "2025-06-30",
|
|
"labels": ["Project XYZ"],
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="title", weight=1 / 7),
|
|
CaseInsensitiveBinaryCritic(critic_field="parent_issue", weight=1 / 7),
|
|
CaseInsensitiveBinaryCritic(critic_field="issue_type", weight=1 / 7),
|
|
CaseInsensitiveBinaryCritic(critic_field="assignee", weight=1 / 7),
|
|
CaseInsensitiveBinaryCritic(critic_field="reporter", weight=1 / 7),
|
|
BinaryCritic(critic_field="due_date", weight=1 / 7),
|
|
CaseInsensitiveListOfStringsBinaryCritic(critic_field="labels", weight=1 / 7),
|
|
],
|
|
)
|
|
|
|
return suite
|
|
|
|
|
|
@tool_eval()
|
|
def labels_eval_suite() -> EvalSuite:
|
|
suite = EvalSuite(
|
|
name="Labels eval suite",
|
|
system_message=(
|
|
"You are an AI assistant with access to Jira tools. "
|
|
"Use them to help the user with their tasks."
|
|
),
|
|
catalog=catalog,
|
|
rubric=rubric,
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Add labels",
|
|
user_message="Add the labels 'Hello' and 'World' to the issue ENG-123.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=add_labels_to_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"labels": ["Hello", "World"],
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveListOfStringsBinaryCritic(critic_field="labels", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Add labels without notifying watchers",
|
|
user_message=(
|
|
"Add the labels 'Hello' and 'World' to the issue ENG-123. Do not notify watchers."
|
|
),
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=add_labels_to_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"labels": ["Hello", "World"],
|
|
"notify_watchers": False,
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=1 / 3),
|
|
CaseInsensitiveListOfStringsBinaryCritic(critic_field="labels", weight=1 / 3),
|
|
BinaryCritic(critic_field="notify_watchers", weight=1 / 3),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Remove labels",
|
|
user_message="Remove the labels 'Hello' and 'World' from the issue ENG-123.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=remove_labels_from_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"labels": ["Hello", "World"],
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveListOfStringsBinaryCritic(critic_field="labels", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Remove labels without notifying watchers",
|
|
user_message=(
|
|
"Remove the labels 'Hello' and 'World' from the issue ENG-123. Do not notify watchers."
|
|
),
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=remove_labels_from_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"labels": ["Hello", "World"],
|
|
"notify_watchers": False,
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=1 / 3),
|
|
CaseInsensitiveListOfStringsBinaryCritic(critic_field="labels", weight=1 / 3),
|
|
BinaryCritic(critic_field="notify_watchers", weight=1 / 3),
|
|
],
|
|
)
|
|
|
|
return suite
|
|
|
|
|
|
@tool_eval()
|
|
def update_issue_eval_suite() -> EvalSuite:
|
|
suite = EvalSuite(
|
|
name="Update issue eval suite",
|
|
system_message=(
|
|
"You are an AI assistant with access to Jira tools. "
|
|
"Use them to help the user with their tasks."
|
|
),
|
|
catalog=catalog,
|
|
rubric=rubric,
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Update issue with new assignee",
|
|
user_message="Change the assignee of the ENG-123 issue to John Doe.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"assignee": "John Doe",
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveBinaryCritic(critic_field="assignee", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Update issue with new priority",
|
|
user_message="Set the priority of the ENG-123 issue to high.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"priority": "High",
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveBinaryCritic(critic_field="priority", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Update issue with new due date",
|
|
user_message="Set the due date of the ENG-123 issue to 2025-06-30.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"due_date": "2025-06-30",
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveBinaryCritic(critic_field="due_date", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Update issue with new labels",
|
|
user_message="Change the labels in the ENG-123 issue to 'Hello' and 'World'.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"labels": ["Hello", "World"],
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveListOfStringsBinaryCritic(critic_field="labels", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Update issue with new title and description",
|
|
user_message=(
|
|
"Change the title and description of the ENG-123 issue to 'Test issue' "
|
|
"and 'This is a test issue'."
|
|
),
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"title": "Test issue",
|
|
"description": "This is a test issue",
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=1 / 3),
|
|
CaseInsensitiveBinaryCritic(critic_field="title", weight=1 / 3),
|
|
CaseInsensitiveBinaryCritic(critic_field="description", weight=1 / 3),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Clear due date",
|
|
user_message="Clear the due date of the issue ENG-123.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"due_date": "",
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveBinaryCritic(critic_field="due_date", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Remove assignee",
|
|
user_message="Remove the assignee from the issue ENG-123.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"assignee": "",
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=0.5),
|
|
CaseInsensitiveBinaryCritic(critic_field="assignee", weight=0.5),
|
|
],
|
|
)
|
|
|
|
suite.add_case(
|
|
name="Remove assignee",
|
|
user_message="Remove the assignee from the issue ENG-123 without notifying anyone.",
|
|
expected_tool_calls=[
|
|
ExpectedToolCall(
|
|
func=update_issue,
|
|
args={
|
|
"issue": "ENG-123",
|
|
"assignee": "",
|
|
"notify_watchers": False,
|
|
},
|
|
),
|
|
],
|
|
rubric=rubric,
|
|
critics=[
|
|
CaseInsensitiveBinaryCritic(critic_field="issue", weight=1 / 3),
|
|
CaseInsensitiveBinaryCritic(critic_field="assignee", weight=1 / 3),
|
|
BinaryCritic(critic_field="notify_watchers", weight=1 / 3),
|
|
],
|
|
)
|
|
|
|
return suite
|