Input Parsing

PICLE does more than split on spaces.

App.parse_command() has special handling for quoted strings, JSON-like chunks, JSON fields, and scalar conversion.

Example Models

from typing import Any, List, Union
from pydantic import BaseModel, Field, Json, StrictInt, StrictStr


class RunCommand(BaseModel):
    command: StrictStr = Field(None, description="Command text")

    @staticmethod
    def run(**kwargs):
        return kwargs


class PayloadCommand(BaseModel):
    data: Json[Any] = Field(None, description="JSON payload")
    label: Any = Field(None, description="Label")

    @staticmethod
    def run(**kwargs):
        return kwargs


class LimitsCommand(BaseModel):
    retries: StrictInt = Field(None, description="Retry count")
    name: StrictStr = Field(None, description="Job name")
    threshold: Any = Field(None, description="Threshold")

    @staticmethod
    def run(**kwargs):
        return kwargs


class BatchCommand(BaseModel):
    items: Union[StrictStr, List[StrictStr]] = Field(
        None, description="Collect multiple items"
    )
    tag: StrictStr = Field(None, description="Batch tag")

    @staticmethod
    def run(**kwargs):
        return kwargs


class Root(BaseModel):
    run_command: RunCommand = Field(None, alias="run-command")
    payload: PayloadCommand = Field(None)
    limits: LimitsCommand = Field(None)
    batch: BatchCommand = Field(None)

Quoted Values

Single quotes and double quotes let you pass multi-word values as a single field value.

picle#run-command command 'show version | match "Juniper: "'
{'command': 'show version | match "Juniper: "'}

picle#run-command command "hello world"
{'command': 'hello world'}

Double-quoted single-word values work too:

picle#run-command command "hello"
{'command': 'hello'}

JSON-like Chunks

If a value starts with { or [, PICLE keeps collecting tokens until it sees a token that ends with the matching closing bracket.

That makes it possible to pass JSON-like text through the shell.

picle#payload data [1, 2, 3] label demo
{'data': '[1, 2, 3]', 'label': 'demo'}

picle#payload data {"person":{"name":"John","age":30}} label demo
{'data': '{"person":{"name":"John","age":30}}', 'label': 'demo'}

Json[...] Fields Preserve Raw Input

Json[...] fields are special-cased in the parser.

PICLE keeps the raw text instead of converting values like true, false, or null before Pydantic sees them.

picle#payload data true label demo
{'data': 'true', 'label': 'demo'}

picle#payload data false label demo
{'data': 'false', 'label': 'demo'}

picle#payload data null label demo
{'data': 'null', 'label': 'demo'}

Automatic Scalar Conversion

For non-str fields, PICLE tries a lightweight conversion pass before validation.

• True, False, and None become Python values.
• Digit-only values become int.
• Values containing . are tried as float.

picle#limits retries 3 threshold 3.14 name nightly
{'retries': 3, 'threshold': 3.14, 'name': 'nightly'}

String fields keep their string representation:

picle#limits name 123
{'name': '123'}

Repeating A Field

If your field accepts list-like input, you can repeat the field name to collect multiple values.

picle#batch items val1 items val2 items val3 tag nightly
{'items': ['val1', 'val2', 'val3'], 'tag': 'nightly'}

Use this pattern when you want explicit, readable shell grammar for list input.