Trac RPC API client

Modern, Pythonic and type-safe Trac RPC API client.

Note

This package was developed primarily to satisfy my personal migration needs and as such covers only the read interface of the API!

Please feel free to contact me for complex migration projects from Trac and/or custom development needs. High quality contributions that follow the library structure and donations are also much appreciated.

Installation

$ pip install trac-rpc

Dependencies

• Pydantic V2
• httpx

Usage

>>> from trac_rpc.client import ApiClient, HttpClient

>>> api_client = ApiClient(
    rpc_url="http://127.0.0.1:8000/login/rpc",
    http_client=HttpClient(auth=("admin", "admin")),
)

>>> api_client.get_api_version()
TracApiVersion(epoch=1, major=2, minor=0)

>>> result = api_client.get_ticket_last_field_change(1, "status", "closed")
>>> result.timestamp
datetime.datetime(2025, 2, 27, 13, 37, 11, 171873, tzinfo=TzInfo(UTC))

Important

Trac APIs (e.g. query_tickets) do not return IDs and/or objects sorted in alphanumeric order!

They can either be returned in a custom order (such as priority levels) or grouped by certain fields (such as ticket IDs). If you want to iterate on the objects in a particular order, always remember to sort them appropriately after you call the API.

Customizing models

Changing default string type

Some models like TracMilestone can be parameterized with a string type such as TracStrippedStr instead of str to strip leading and trailing whitespace:

from trac_rpc.client import ApiClient
from trac_rpc.models import TracMilestone
from trac_rpc.validators import TracStrippedStr

api_client = ApiClient(rpc_url="http://127.0.0.1:8000/rpc")
milestone = api_client.get_milestone(" milestone2  ", TracMilestone[TracStrippedStr])

This is especially useful for migrations to detect unnoticed whitespace problems. However, this can potentially cause calls to the Trac RPC API with string object names to fail, so it is not enabled by default.

Adjusting TracTicket model

Each Trac installation is configured in its own way. The default TracTicket model can be easily customized by subclassing:

from trac_rpc.models import TracTicket as TracTicketBase
from trac_rpc.validators import TracSpaceOrCommaSeparated

class TracTicket(TracTicketBase):
    keywords: TracSpaceOrCommaSeparated[str]
    cc: TracSpaceOrCommaSeparated[str]

Usually Trac fields are separated by spaces (TracSpaceSeparated), but alternative types are provided for convenience (TracCommaSeparated, TracSpaceOrCommaSeparated).

Complex validation to ensure data consistency is possible with the validate_in_set validator:

from functools import partial
from typing import Annotated

from pydantic import AfterValidator
from trac_rpc.models import TracTicket as TracTicketBase
from trac_rpc.validators import validate_in_set, TracOptionalField

VERSIONS = {"1.0", "2.0"}


class TracTicket(TracTicketBase):
    validate_version = partial(validate_in_set, allowed=VERSIONS, optional=True)
    ...
    version: Annotated[TracOptionalField[str], AfterValidator(validate_version)]

Development

Setting up test Trac server

See GitHub Actions workflow for integration tests.

Releases

To release a new version and publish it to PyPI:

• Bump version with hatch and commit
  • hatch version minor or hatch version patch
• Create GitHub release (and tag)