JSON to Python Types

json2pyi infers a type schema from a sample JSON file and generates Python type definitions (dataclass, Pydantic BaseModel or PEP-589 TypedDict) accordingly. It runs in browser, requiring no installation.

🌐 Available online 🔗: https://json2pyi.pages.dev

Example

Input:

{
    "page": {
        "id": "kctbh9vrtdwd",
        "name": "GitHub",
        "url": "https://www.githubstatus.com",
        "time_zone": "Etc/UTC",
        "updated_at": "2020-12-03T08:11:21.385Z"
    },
    "components": [
        {
            "id": "8l4ygp009s5s",
            "name": "Git Operations",
            "status": "operational",
            "created_at": "2017-01-31T20:05:05.370Z",
            "updated_at": "2020-10-29T22:51:43.831Z",
            "position": 1,
            "description": "Performance of git clones, pulls, pushes, and associated operations",
            "showcase": false,
            "start_date": null,
            "group_id": null,
            "page_id": "kctbh9vrtdwd",
            "group": false,
            "only_show_if_degraded": false
        },
        /* ... */
    ],
    "incidents": [],
    "scheduled_maintenances": [],
    "status": {
        "indicator": "none",
        "description": "All Systems Operational"
    }
}

Output:

@dataclass
class Page:
    id = str
    name = str
    url = str
    time_zone = str
    updated_at = str

@dataclass
class Components:
    id = str
    name = str
    status = str
    created_at = str
    updated_at = str
    position = int
    description = Optional[str]
    showcase = bool
    start_date = None
    group_id = None
    page_id = str
    group = bool
    only_show_if_degraded = bool

@dataclass
class Status:
    indicator = str
    description = str

@dataclass
class RootObject:
    page = Page
    components = List[Components]
    incidents = List[Any]
    scheduled_maintenances = List[Any]
    status = Status

Or:

from typing import TypedDict, Optional, List

from datetime import datetime


IncidentUpdate = TypedDict("IncidentUpdate", {"body": str, "created_at": datetime, "display_at": datetime, "id": str, "incident_id": str, "status": str, "updated_at": datetime})

IncidentOrScheduledMaintenance = TypedDict("IncidentOrScheduledMaintenance", {"created_at": datetime, "id": str, "impact": str, "incident_updates": List[IncidentUpdate], "monitoring_at": None, "name": str, "page_id": str, "resolved_at": None, "shortlink": str, "status": str, "updated_at": datetime, "scheduled_for": Optional[datetime], "scheduled_until": Optional[datetime]})

Component = TypedDict("Component", {"created_at": datetime, "description": None, "id": str, "name": str, "page_id": str, "position": int, "status": str, "updated_at": datetime})

Status = TypedDict("Status", {"description": str, "indicator": str})

Page = TypedDict("Page", {"id": str, "name": str, "url": str, "updated_at": datetime})

RootObject = TypedDict("UnnammedType3C2BC8", {"page": Page, "status": Status, "components": List[Component], "incidents": List[IncidentOrScheduledMaintenance], "scheduled_maintenances": List[IncidentOrScheduledMaintenance]})

TODO

• Detect tuple (array)
• Detect UUID / datetime
• Detect Enum
• Merge data types with similar structure and common name prefix/suffix
• Detect recursive type definition (e.g. tree)
• Include imports of non-primitive types
• Generate type alias for complex Union
• Improve the logic of determining whether a Union ix complex or not
• Generate TypedDict
• Refactor to unify TypedDict and dataclass generation Seperated intendedly for clear code structure.
• Compile to WASM and provide a Web-based app
• Allow to tweak more options on Web app (partially blocked by vhiribarren/raytracer-rust#8)
• Avoid merging data types with totally different structures in a union
• Avoid unnecessary heap allocation by reducing one-time usage of Vec
• Allow specifying the order of generated data types
• Support more input types, such as JSON Schema
• Support more target languages
• Add usage instructions

Credits

The project is inspired by:

• https://app.quicktype.io/?l=ts
• https://github.com/thautwarm/schema-provider.py
• https://jvilk.com/MakeTypes/
• https://github.com/koxudaxi/datamodel-code-generator/