`jellycell.cache.manifest`¶

Pydantic schema for cell execution manifests.

One JSON file per cell execution, stored under .jellycell/cache/manifests/. The schema here is a spec §10.1 contract — every field that users or agents can see is versioned via schema_version.

See spec §2.3 for the full manifest shape.

Module Contents¶

Classes¶

`StreamOutput`	Captured stdout/stderr stream.
`DisplayDataOutput`	A `display_data` mime bundle (e.g., images, html).
`ExecuteResultOutput`	The final expression value of a cell.
`ErrorOutput`	An exception raised during cell execution.
`ArtifactRecord`	A file produced by a cell (via `jc.save`, `jc.figure`, `jc.table`).
`Manifest`	A single cell execution’s manifest (spec §2.3).

Data¶

`MANIFEST_SCHEMA_VERSION`
`OutputRecord`

API¶

jellycell.cache.manifest.MANIFEST_SCHEMA_VERSION¶: 1

class jellycell.cache.manifest.StreamOutput(/, **data: Any)¶

Bases: pydantic.BaseModel

Captured stdout/stderr stream.

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.

model_config¶: ‘ConfigDict(…)’

type: Literal[stream]¶: ‘stream’

name: Literal[stdout, stderr]¶: None

blob: str¶: None

class jellycell.cache.manifest.DisplayDataOutput(/, **data: Any)¶

Bases: pydantic.BaseModel

A display_data mime bundle (e.g., images, html).

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.

model_config¶: ‘ConfigDict(…)’

type: Literal[display_data]¶: ‘display_data’

mime: str¶: None

blob: str¶: None

w: int | None¶: None

h: int | None¶: None

class jellycell.cache.manifest.ExecuteResultOutput(/, **data: Any)¶

Bases: pydantic.BaseModel

The final expression value of a cell.

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.

model_config¶: ‘ConfigDict(…)’

type: Literal[execute_result]¶: ‘execute_result’

mime: str¶: None

blob: str¶: None

execution_count: int | None¶: None

class jellycell.cache.manifest.ErrorOutput(/, **data: Any)¶

Bases: pydantic.BaseModel

An exception raised during cell execution.

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.

model_config¶: ‘ConfigDict(…)’

type: Literal[error]¶: ‘error’

ename: str¶: None

evalue: str¶: None

traceback: list[str]¶: None

jellycell.cache.manifest.OutputRecord¶: None

class jellycell.cache.manifest.ArtifactRecord(/, **data: Any)¶

Bases: pydantic.BaseModel

A file produced by a cell (via jc.save, jc.figure, jc.table).

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.

model_config¶: ‘ConfigDict(…)’

path: str¶

None

Path relative to the project root.

sha256: str¶

None

Hex sha256 of the file contents.

size: int¶

None

Size in bytes.

mime: str | None¶

None

MIME type, if known.

caption: str | None¶

None

Human-readable caption for the artifact — used as the figure / table heading in tearsheets and the alt text on image embeds.

Optional, additive field (§10.1 safe). Populated when the producing jc.* call passes caption="..."; otherwise the tearsheet falls back to the cell name.

notes: str | None¶

None

Longer analyst-authored notes shown under the caption in tearsheets. Use for methodology, caveats, data-source reminders. Optional additive.

tags: list[str]¶

‘Field(…)’

Free-form tags (e.g. ["regression", "diagnostic"]) for grouping and filtering. Currently surface-only metadata; future versions may use tags for tearsheet section ordering. Optional additive.

class jellycell.cache.manifest.Manifest(/, **data: Any)¶

Bases: pydantic.BaseModel

A single cell execution’s manifest (spec §2.3).

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.

model_config¶: ‘ConfigDict(…)’

schema_version: int¶

None

§10.1 contract.

cache_key: str¶: None

notebook: str¶: None

cell_id: str¶: None

cell_name: str | None¶: None

source_hash: str¶: None

dep_keys: list[str]¶: ‘Field(…)’

env_hash: str¶: None

executed_at: datetime.datetime¶: None

duration_ms: int¶: None

status: Literal[ok, error, cached]¶: None

outputs: list[jellycell.cache.manifest.OutputRecord]¶: ‘Field(…)’

artifacts: list[jellycell.cache.manifest.ArtifactRecord]¶: ‘Field(…)’

to_json() → str¶: Serialize to canonical JSON with 2-space indentation.

classmethod from_json(text: str) → jellycell.cache.manifest.Manifest¶: Parse a manifest from JSON text.

write(path: pathlib.Path) → None¶: Write to path as JSON with a trailing newline.

classmethod read(path: pathlib.Path) → jellycell.cache.manifest.Manifest¶: Read a manifest from disk.

jellycell.cache.manifest¶

Module Contents¶

Classes¶

Data¶

API¶

`jellycell.cache.manifest`¶