avatars¶
Table of contents¶
Resources¶
Our main docs are at https://docs.octopize.io/
Our web application is at https://www.octopize.app
To contact us, reach out at support@octopize.io
Overview¶
The Python client is structured in layered components so you can choose the highest level of abstraction that fits your workflow and drop down only when you need more control.
Core layers¶
Manager (high-level convenience) A thin facade you instantiate first. It authenticates the user, creates workflow
Runnerobjects, and offers shortcuts to recent jobs, results, and health checks. For most day‑to‑day usage you only interact with theManagerplus aRunner.Runner (workflow orchestration) Encapsulates an avatarization “set”: uploading tables, linking them, advising or setting parameters, executing jobs, and gathering results. It maintains in‑memory state (tables, jobs, results URLs) and uses the authenticated API session carried by the Manager.
ApiClient (low-level resource interface) Handles HTTP calls, authentication tokens, retries/timeouts and exposes resource clients (
jobs,datasets,results,users, etc.). Use it directly when you need fine‑grained control beyond whatManager/Runnerprovide.
Supporting pieces¶
Pydantic models – Strongly typed request/response schemas ensuring your inputs are validated before they reach the server.
avatar_yaml configuration – A Python representation of the YAML structure used to describe avatarization parameters (tables, links, privacy / signal settings). The
Runnerbuilds and manipulates aavatar_yaml.Configunder the hood; you can also load from / export to YAML directly.Processors – Pre / post processing helpers that can be referenced in configs to transform data prior to or after avatarization.
Typical usage pattern¶
from avatars import Manager
manager = Manager()
manager.authenticate("user", "pass")
runner = manager.create_runner("demo_set")
runner.add_table("wbcd", "fixtures/wbcd.csv")
runner.advise_parameters() # optional server guidance
runner.set_parameters("wbcd", k=15) # fine‑tune
runner.run() # launch avatarization
runner.shuffled("wbcd") # access result table
Where to go next¶
Start with Installation / Tutorial for a first run.
You can download other tutorials here: https://github.com/octopize/avatar-python/tree/main/notebooks
See Detailed doc for the conceptual overview and the full API surface.
Consult dedicated pages for
Manager,RunnerandApiClientwhen you need deeper control.