Templates let you create reusable sandbox images from Dockerfiles. Pre-install languages, tools, and dependencies so your sandboxes start ready to go.
Building a Template
from opencomputer import Template
templates = Template(api_url, api_key)
template = await templates.build('python-ml', """
FROM ubuntu:22.04
RUN apt-get update && apt-get install -y python3 python3-pip
RUN pip3 install numpy pandas scikit-learn
""")
print(template.template_id) # tpl-abc123
print(template.status) # building → ready
await templates.build(name, dockerfile)
Builds a new template from a Dockerfile.
A unique name for the template.
The Dockerfile content to build from.
TemplateInfo
Listing Templates
all_templates = await templates.list()
for t in all_templates:
print(f"{t.name} ({t.tag}) — {t.status}")
await templates.list()
Returns all templates.
Returns: list[TemplateInfo]
Getting a Template
template = await templates.get('python-ml')
print(template)
await templates.get(name)
Gets a template by name.
Returns: TemplateInfo
Deleting a Template
await templates.delete('python-ml')
await templates.delete(name)
Deletes a template by name.
Returns: None
TemplateInfo
TemplateInfo is a dataclass:
| Field | Type | Description |
|---|
template_id | str | Unique template identifier |
name | str | Template name |
tag | str | Image tag |
status | str | Build status: "building", "ready", "error" |
Using a Template
Once built, pass the template name when creating a sandbox:
sandbox = await Sandbox.create(template='python-ml')
result = await sandbox.commands.run(
'python3 -c "import numpy; print(numpy.__version__)"'
)
print(result.stdout) # 1.24.0
await sandbox.kill()
