Define and validate schema.org structured data in Python with Pydantic

requirements.txt

pydantic
rdflib
pandas
matplotlib
schemaorg

schemaorg-pydantic.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define and validate schema.org structured data in Python with Pydantic "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pydantic import BaseModel, Schema\n",
    "from pydantic.main import MetaModel\n",
    "from schemaorg import main as schemaorg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Thing(BaseModel):\n",
    "    \"\"\"The most generic type of item.\"\"\"\n",
    "    \n",
    "    # Need to define extra items at the top level \n",
    "    class Config:\n",
    "        title = 'Thing'\n",
    "        schema_extra = {\n",
    "            '$schema': 'https://schema.org',\n",
    "            '$id': 'https://schema.org/Thing',\n",
    "        }\n",
    "    \n",
    "    additionalType: str = Schema(\n",
    "        ...,\n",
    "        title='additionalType',\n",
    "        description=(\n",
    "            \"An additional type for the item, typically \"\n",
    "            \"used for adding more specific types from \"\n",
    "            \"external vocabularies in microdata syntax. \"\n",
    "            \"This is a relationship between something and \"\n",
    "            \"a class that the thing is in. In RDFa syntax, \"\n",
    "            \"it is better to use the native RDFa syntax - \"\n",
    "            \"the 'typeof' attribute - for multiple types. \"\n",
    "            \"Schema.org tools may have only weaker \"\n",
    "            \"understanding of extra types, in particular \"\n",
    "            \"those defined externally.\"\n",
    "        )\n",
    "    )\n",
    "    \n",
    "    alternateName: str = Schema(\n",
    "        ...,\n",
    "        title='alternateName',\n",
    "        description=\"An alias for the item.\"\n",
    "    )\n",
    "        \n",
    "    description: str = Schema(\n",
    "        ...,\n",
    "        description=\"A description of the item.\"\n",
    "    )\n",
    "        \n",
    "    disambiguatingDescription: str = Schema(\n",
    "        ...,\n",
    "        description=\"A sub property of description. A short description of the item used to disambiguate from other, similar items. Information from other properties (in particular, name) may be necessary for the description to be useful for disambiguation.\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I want to enforce the attributes in the `Config` accessor, *and* I think the accessor syntax is a bit ugly. Let's see if we can make it go away."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "class MetaSchema(MetaModel):\n",
    "    \"\"\"Metaclass that checks for three required class attributes:\n",
    "        1. _id: the ID of the event\n",
    "        2. _version: the version of the current schema.\n",
    "        3. _title: the name of the schema.\n",
    "\n",
    "    These attribute are mapped to pydantic.BaseModel's `Config` inner class\n",
    "    for proper schema generation+validation.\n",
    "    \"\"\"\n",
    "    def __new__(cls, name, base, dct):\n",
    "        # Check that required keys are found.\n",
    "        if not all((key in dct for key in ['_id', '_title', '_version', '_schema'])):\n",
    "            raise AttributeError('Required class attributes are missing from the {} class.'.format(name))\n",
    "\n",
    "        # Check that keys are the proper types.\n",
    "        if not all((\n",
    "            type(dct['_id']) in (str, type(None)),\n",
    "            type(dct['_version']) in (float, type(None)),\n",
    "            type(dct['_title']) in (str, type(None)),\n",
    "            type(dct['_schema']) in (str, type(None)),\n",
    "            )):\n",
    "            raise TypeError('Check the class attributes types: \"_id\" must be a string, '\n",
    "                            '\"_version\" must be an integer, and \"_title\" must be a string.')\n",
    "\n",
    "        # Add a Config inner class to this Pydantic model.\n",
    "        class Config:\n",
    "            title = dct['_title']\n",
    "            schema_extra = {\n",
    "                '$id': dct['_id'],\n",
    "                '$schema': dct['_schema'],\n",
    "                'version': dct['_version']\n",
    "            }\n",
    "\n",
    "        dct['Config'] = Config\n",
    "        return super(MetaSchema, cls).__new__(cls, name, base, dct)\n",
    "\n",
    "\n",
    "class JsonSchema(pydantic.BaseModel, metaclass=MetaSchema):\n",
    "    \"\"\"A pydantic base Model for JSON schemas.\"\"\"\n",
    "    _id: str = None\n",
    "    _version: float = None\n",
    "    _title: str = None\n",
    "    _schema: str = None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Thing(JsonSchema):\n",
    "    \"\"\"The most generic type of item.\"\"\"\n",
    "    # Define top level attributes\n",
    "    _id = 'https://schema.org/Thing'\n",
    "    _version = 3.9\n",
    "    _title = 'Thing'\n",
    "    _schema = 'https://schema.org'\n",
    "\n",
    "    additionalType: str = Schema(\n",
    "        ...,\n",
    "        title='additionalType',\n",
    "        description=(\n",
    "            \"An additional type for the item, typically \"\n",
    "            \"used for adding more specific types from \"\n",
    "            \"external vocabularies in microdata syntax. \"\n",
    "            \"This is a relationship between something and \"\n",
    "            \"a class that the thing is in. In RDFa syntax, \"\n",
    "            \"it is better to use the native RDFa syntax - \"\n",
    "            \"the 'typeof' attribute - for multiple types. \"\n",
    "            \"Schema.org tools may have only weaker \"\n",
    "            \"understanding of extra types, in particular \"\n",
    "            \"those defined externally.\"\n",
    "        )\n",
    "    )\n",
    "    \n",
    "    alternateName: str = Schema(\n",
    "        ...,\n",
    "        title='alternateName',\n",
    "        description=\"An alias for the item.\"\n",
    "    )\n",
    "        \n",
    "    description: str = Schema(\n",
    "        ...,\n",
    "        description=\"A description of the item.\"\n",
    "    )\n",
    "        \n",
    "    disambiguatingDescription: str = Schema(\n",
    "        ...,\n",
    "        description=\"A sub property of description. A short description of the item used to disambiguate from other, similar items. Information from other properties (in particular, name) may be necessary for the description to be useful for disambiguation.\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Thing.schema()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Validate a new object"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What happens when we create an invalid object?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "thing = Thing()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's try a valid object..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "thing = Thing(\n",
    "    alternateName='New Thing',\n",
    "    description='This is a new thing',\n",
    "    disambiguatingDescription='This thing is unique.',\n",
    "    additionalType='No additional type'\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No error was raised."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Autogenerate pydantic objects from schema.org"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "class SchemaOrg(MetaSchema):\n",
    "    \n",
    "    def __new__(cls, name, base, dct):\n",
    "        annotations = {}\n",
    "        \n",
    "        data = schemaorg.Schema(name)\n",
    "        \n",
    "        dct = dict(\n",
    "            _title=name,\n",
    "            _id=data.id,\n",
    "            _version=float(data.version),\n",
    "            _schema=data.base,\n",
    "            __doc__=data.comment,\n",
    "            __annotations__={}\n",
    "        )\n",
    "\n",
    "        # Currently, sets all class variables to type==str for\n",
    "        # demostration purposes.\n",
    "        # Need to develop datatypes for Schema.org objects.\n",
    "        for key, val in data._properties.items():\n",
    "            dct[key] = Schema(\n",
    "                ...,\n",
    "                description=val['comment'],\n",
    "                title=val['label']\n",
    "            )\n",
    "            dct['__annotations__'][key] = str\n",
    "\n",
    "        base = (BaseModel,) + base\n",
    "        \n",
    "        return super(SchemaOrg, cls).__new__(cls, name, base, dct)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Thing(metaclass=SchemaOrg): pass\n",
    "class Event(metaclass=SchemaOrg): pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Thing.schema()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Event.schema()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.7",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}