arcade-mcp/toolkits/microsoft/evals/outlook_mail/eval_read.py
Sam Partee b6b4cd0a4c
🏗️ Restructure: Multi-Package Architecture + uv Migration (#412)
### Overview
Major restructuring from monolithic `arcade-ai` package to modular
library architecture with standardized uv-based dependency management.

![arcade-ai Monorepo
(2)](https://github.com/user-attachments/assets/25f102b0-bb87-4a04-9701-d227d05664b1)

### 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>
2025-06-11 16:48:17 -07:00

210 lines
7 KiB
Python

from datetime import timedelta
from arcade_evals import (
BinaryCritic,
DatetimeCritic,
EvalRubric,
EvalSuite,
ExpectedToolCall,
tool_eval,
)
from arcade_tdk import ToolCatalog
from arcade_microsoft.outlook_mail import (
list_emails,
list_emails_by_property,
list_emails_in_folder,
)
from arcade_microsoft.outlook_mail.enums import WellKnownFolderNames
from evals.outlook_mail.additional_messages import (
list_emails_with_pagination_token_additional_messages,
)
# Evaluation rubric
rubric = EvalRubric(
fail_threshold=0.9,
warn_threshold=0.95,
)
catalog = ToolCatalog()
catalog.add_tool(list_emails, "Microsoft")
catalog.add_tool(list_emails_in_folder, "Microsoft")
catalog.add_tool(list_emails_by_property, "Microsoft")
@tool_eval()
def outlook_mail_read_eval_suite() -> EvalSuite:
"""Create an evaluation suite for Outlook Mail tools."""
suite = EvalSuite(
name="Outlook Mail Tools Evaluation",
system_message=("You are an AI that has access to tools to send, read, and write emails."),
catalog=catalog,
rubric=rubric,
)
suite.add_case(
name="List emails in mailbox",
user_message="get my five most recent emails",
expected_tool_calls=[
ExpectedToolCall(
func=list_emails,
args={"limit": 5},
)
],
critics=[
BinaryCritic(critic_field="limit", weight=1.0),
],
)
suite.add_case(
name="List emails in mailbox with pagination token",
user_message="get the next 3",
expected_tool_calls=[
ExpectedToolCall(
func=list_emails,
args={
"limit": 3,
"pagination_token": "https://graph.microsoft.com/v1.0/me/messages?%24count=true&%24orderby=receivedDateTime+DESC&%24select=bccRecipients%2cbody%2cccRecipients%2cconversationId%2cconversationIndex%2cflag%2cfrom%2chasAttachments%2cimportance%2cisDraft%2cisRead%2creceivedDateTime%2creplyTo%2csubject%2ctoRecipients%2cwebLink&%24top=1&%24skip=1",
},
)
],
critics=[
BinaryCritic(critic_field="limit", weight=0.2),
BinaryCritic(critic_field="pagination_token", weight=0.8),
],
additional_messages=list_emails_with_pagination_token_additional_messages,
)
suite.add_case(
name="List emails in well-known folder",
user_message="summarize my inbox",
expected_tool_calls=[
ExpectedToolCall(
func=list_emails_in_folder,
args={
"well_known_folder_name": WellKnownFolderNames.INBOX,
"folder_id": None,
},
)
],
critics=[
BinaryCritic(critic_field="well_known_folder_name", weight=0.5),
BinaryCritic(critic_field="folder_id", weight=0.5),
],
)
suite.add_case(
name="List emails in folder by id",
user_message="get 5 from folder AQMkADAwATM0MDAAMi04Y2Y1LTQ3MTEALTAwAi0wMAoALgAAAyXxSd3UxTpCkDpGouEg0JMBAFuxokOLZRtDncM4", # noqa: E501
expected_tool_calls=[
ExpectedToolCall(
func=list_emails_in_folder,
args={
"well_known_folder_name": None,
"folder_id": "AQMkADAwATM0MDAAMi04Y2Y1LTQ3MTEALTAwAi0wMAoALgAAAyXxSd3UxTpCkDpGouEg0JMBAFuxokOLZRtDncM4", # noqa: E501
"limit": 5,
},
)
],
critics=[
BinaryCritic(critic_field="well_known_folder_name", weight=0.4),
BinaryCritic(critic_field="folder_id", weight=0.4),
BinaryCritic(critic_field="limit", weight=0.2),
],
)
return suite
@tool_eval()
def outlook_mail_list_emails_by_property_eval_suite() -> EvalSuite:
"""Create an evaluation suite for Outlook Mail tools."""
suite = EvalSuite(
name="Outlook Mail Tools Evaluation",
system_message=("You are an AI that has access to tools to send, read, and write emails."),
catalog=catalog,
rubric=rubric,
)
suite.add_case(
name="List emails by subject",
user_message="get all emails that talk about The Green Bottle",
expected_tool_calls=[
ExpectedToolCall(
func=list_emails_by_property,
args={
"property": "subject",
"operator": "contains",
"value": "The Green Bottle",
},
),
],
critics=[
BinaryCritic(critic_field="property", weight=1 / 3),
BinaryCritic(critic_field="operator", weight=1 / 3),
BinaryCritic(critic_field="value", weight=1 / 3),
],
)
suite.extend_case(
name="List emails by thread",
user_message="get all emails in my thread 1k2jh324h92f24krjb34mtb43kj4bk3tmn34b3k4nnm3tb34mntb34mntb3m4bt3mn4bt3mn4btmnb34tmnb3t4mnb==34tkjh", # noqa: E501
expected_tool_calls=[
ExpectedToolCall(
func=list_emails_by_property,
args={
"property": "conversationId",
"operator": "eq",
"value": "1k2jh324h92f24krjb34mtb43kj4bk3tmn34b3k4nnm3tb34mntb34mntb3m4bt3mn4bt3mn4btmnb34tmnb3t4mnb==34tkjh", # noqa: E501
},
),
],
critics=[
BinaryCritic(critic_field="property", weight=1 / 3),
BinaryCritic(critic_field="operator", weight=1 / 3),
BinaryCritic(critic_field="value", weight=1 / 3),
],
)
suite.extend_case(
name="List emails by date",
user_message="Today is May 1st 2025. Get all emails that are a year old or older",
expected_tool_calls=[
ExpectedToolCall(
func=list_emails_by_property,
args={
"property": "receivedDateTime",
"operator": "le",
"value": "2024-05-01T00:00:00Z",
},
),
],
critics=[
BinaryCritic(critic_field="property", weight=1 / 3),
BinaryCritic(critic_field="operator", weight=1 / 3),
DatetimeCritic(critic_field="value", weight=1 / 3, tolerance=timedelta(days=1)),
],
)
suite.extend_case(
name="List emails by sender",
user_message="get all of my correspondence with the folks over at arcade.dev",
expected_tool_calls=[
ExpectedToolCall(
func=list_emails_by_property,
args={
"property": "sender/emailAddress/address",
"operator": "contains",
"value": "arcade.dev",
},
),
],
critics=[
BinaryCritic(critic_field="property", weight=1 / 3),
BinaryCritic(critic_field="operator", weight=1 / 3),
BinaryCritic(critic_field="value", weight=1 / 3),
],
)
return suite