`jellycell.cli.commands.checkpoint`¶

jellycell checkpoint — reproducible project snapshots.

Bundles a self-contained .tar.gz of the project (notebooks, data, artifacts, reports, manuscripts, jellycell.toml, and the content-addressed cache) so a reviewer can unpack it and see the last-run outputs without re-executing anything. Restores land in a new sibling directory by default to avoid any chance of clobbering in-flight work.

Scope carefully chosen:

Whitelist which directories go in. Anything else (virtualenvs, pip caches, .git, editor scratch) is never archived.
Include .jellycell/cache/ so artifacts can be rebuilt locally from cache hits even without re-running the notebook.
Skip a small hardcoded set of junk dirs that always appear (__pycache__, .ruff_cache, .mypy_cache, etc.).
No gitignore handling for v1 — use jellycell cache clear first if you want lean checkpoints.

Module Contents¶

Classes¶

`CheckpointSummary`	Metadata embedded in the tarball as `checkpoint.json`.
`CheckpointCreateReport`	JSON schema for `jellycell checkpoint create --json`. §10.1 contract.
`CheckpointListEntry`	One row in `checkpoint list`.
`CheckpointListReport`	JSON schema for `jellycell checkpoint list --json`. §10.1 contract.
`CheckpointRestoreReport`	JSON schema for `jellycell checkpoint restore --json`. §10.1 contract.

Functions¶

`checkpoint_create`	Write `.jellycell/checkpoints/<name>.tar.gz`.
`checkpoint_list`	Show all checkpoints under `.jellycell/checkpoints/`.
`checkpoint_restore`	Extract `.jellycell/checkpoints/<name>.tar.gz` to a target dir.

Data¶

checkpoint_app

API¶

jellycell.cli.commands.checkpoint.checkpoint_app¶: ‘Typer(…)’

class jellycell.cli.commands.checkpoint.CheckpointSummary(/, **data: Any)¶

Bases: pydantic.BaseModel

Metadata embedded in the tarball as checkpoint.json.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

schema_version: int¶: 1

created_at: str¶: None

project_name: str¶: None

message: str | None¶: None

files: int¶: None

size_bytes: int¶: None

class jellycell.cli.commands.checkpoint.CheckpointCreateReport(/, **data: Any)¶

Bases: pydantic.BaseModel

JSON schema for jellycell checkpoint create --json. §10.1 contract.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

schema_version: int¶: 1

project: str¶: None

path: str¶: None

files: int¶: None

size_bytes: int¶: None

message: str | None¶: None

class jellycell.cli.commands.checkpoint.CheckpointListEntry(/, **data: Any)¶

Bases: pydantic.BaseModel

One row in checkpoint list.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

name: str¶: None

path: str¶: None

created_at: str¶: None

message: str | None¶: None

size_bytes: int¶: None

class jellycell.cli.commands.checkpoint.CheckpointListReport(/, **data: Any)¶

Bases: pydantic.BaseModel

JSON schema for jellycell checkpoint list --json. §10.1 contract.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

schema_version: int¶: 1

project: str¶: None

checkpoints: list[jellycell.cli.commands.checkpoint.CheckpointListEntry]¶: ‘Field(…)’

class jellycell.cli.commands.checkpoint.CheckpointRestoreReport(/, **data: Any)¶

Bases: pydantic.BaseModel

JSON schema for jellycell checkpoint restore --json. §10.1 contract.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

schema_version: int¶: 1

project: str¶: None

source: str¶: None

target: str¶: None

jellycell.cli.commands.checkpoint.checkpoint_create(ctx: typer.Context, project: pathlib.Path | None = typer.Argument(None, help='Project root.'), message: str | None = typer.Option(None, '--message', '-m', help='Short label for this checkpoint.'), name: str | None = typer.Option(None, '--name', help='Override the auto-generated checkpoint name.')) → None¶: Write .jellycell/checkpoints/<name>.tar.gz.

jellycell.cli.commands.checkpoint.checkpoint_list(ctx: typer.Context, project: pathlib.Path | None = typer.Argument(None, help='Project root.')) → None¶: Show all checkpoints under .jellycell/checkpoints/.

jellycell.cli.commands.checkpoint.checkpoint_restore(ctx: typer.Context, name: str = typer.Argument(..., help='Checkpoint name (without .tar.gz).'), project: pathlib.Path | None = typer.Option(None, '--project', help='Project root the checkpoint belongs to.'), into: pathlib.Path | None = typer.Option(None, '--into', help='Target directory. Defaults to <project>-restored-<name>/ as a sibling.'), force: bool = typer.Option(False, '--force', help='Allow extraction into an existing, non-empty directory (merges / overwrites).')) → None¶: Extract .jellycell/checkpoints/<name>.tar.gz to a target dir.

jellycell.cli.commands.checkpoint¶

Module Contents¶

Classes¶

Functions¶

Data¶

API¶

`jellycell.cli.commands.checkpoint`¶