Test behavior of PyNWB/HDMF with an attribute spec that has a non-text or non-scalar default value.

spec_non_string_value.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a1d8169c-c17f-48b4-93df-3f0c1f8f24af",
   "metadata": {},
   "source": [
    "## Question: Can you make an NWB/HDMF spec with an attribute that has a non-text or non-scalar default value?\n",
    "Answer: Yes, but the default value will be read by PyNWB/HDMF as one of the default Python types str, int, float, and if it is non-scalar, it will be read as a list or list of lists. On write, a dtype precision can be imposed. As in normal PyNWB/HDMF, consistency of dtypes within a list or list of lists (as opposed to a numpy array) is not checked."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "afad9520-af79-44c9-8975-81c597cf672e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pynwb.spec import NWBGroupSpec, NWBAttributeSpec\n",
    "from pynwb.spec import NWBNamespaceBuilder, export_spec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "d53d11d8-15c0-40b3-8002-5099aa9234eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "ns_builder = NWBNamespaceBuilder(\n",
    "    doc=\"test\",\n",
    "    name=\"ndx-spec-non-string-value\",\n",
    "    version=\"0.1.0\",\n",
    "    author=\"Ryan Ly\",\n",
    "    contact=\"rly@lbl.gov\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "fc70c0fd-f2f4-4561-898a-f1f575f2ae82",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_spec1 = NWBGroupSpec(\n",
    "    neurodata_type_def='TestDefaultValueIntScalar',\n",
    "    neurodata_type_inc=\"NWBDataInterface\",\n",
    "    doc=\"test\",\n",
    "    attributes = [\n",
    "        NWBAttributeSpec(\n",
    "            name=\"test_attribute\",\n",
    "            dtype=\"int\",\n",
    "            doc=\"test\",\n",
    "            default_value=10,\n",
    "        ),\n",
    "    ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f2bbc4b0-da4e-4d1c-a5b4-b05a7ac5eabc",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_spec2 = NWBGroupSpec(\n",
    "    neurodata_type_def='TestDefaultValueInt2D',\n",
    "    neurodata_type_inc=\"NWBDataInterface\",\n",
    "    doc=\"test\",\n",
    "    attributes = [\n",
    "        NWBAttributeSpec(\n",
    "            name=\"test_attribute\",\n",
    "            dtype=\"int64\",\n",
    "            doc=\"test\",\n",
    "            shape=[None, None],\n",
    "            default_value=[[1, 2, 3], [4, 5, 6]],\n",
    "            # NOTE that lists allow mixing of dtypes\n",
    "        ),\n",
    "    ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "62e532fb-d22e-4998-87e0-1c2a144c1129",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_spec3 = NWBGroupSpec(\n",
    "    neurodata_type_def='TestDefaultValueFloat2D',\n",
    "    neurodata_type_inc=\"NWBDataInterface\",\n",
    "    doc=\"test\",\n",
    "    attributes = [\n",
    "        NWBAttributeSpec(\n",
    "            name=\"test_attribute\",\n",
    "            dtype=\"float\",\n",
    "            doc=\"test\",\n",
    "            shape=[None, None],\n",
    "            default_value=[[1., 2., 3.], [4.1, 5.1, 6.1]],\n",
    "        ),\n",
    "    ],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "873d8a3e-3502-4ebc-bfe2-5f92958f33de",
   "metadata": {},
   "outputs": [],
   "source": [
    "new_data_types = [test_spec1, test_spec2, test_spec3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7e94d8a3-526c-4f78-a0a4-ad5a7801ce07",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "output_dir = \"ndx-spec-non-string-value\"\n",
    "Path(output_dir).mkdir(exist_ok=True)\n",
    "export_spec(ns_builder, new_data_types, output_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "87d5908b-680d-409d-9d30-9b9b8b6a21fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pynwb import get_class, load_namespaces\n",
    "load_namespaces(\"ndx-spec-non-string-value/ndx-spec-non-string-value.namespace.yaml\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "98f5cd33-c89c-434d-84e5-5363b21a6c96",
   "metadata": {},
   "outputs": [],
   "source": [
    "TestDefaultValueIntScalar = get_class(\"TestDefaultValueIntScalar\", \"ndx-spec-non-string-value\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b646b5af-75ea-4ca8-9e17-96c499ed3e3d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "obj abc.TestDefaultValueIntScalar at 0x4815574528\n",
       "Fields:\n",
       "  test_attribute: 10"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "TestDefaultValueIntScalar(name=\"obj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7100f942-5746-4f3b-b51e-a7d97576132b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "obj abc.TestDefaultValueIntScalar at 0x4815576688\n",
       "Fields:\n",
       "  test_attribute: 20"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "TestDefaultValueIntScalar(name=\"obj\", test_attribute=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "baecc893-07e6-4878-bddc-4441b057d1fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# TestDefaultValueIntScalar(name=\"obj\", test_attribute=\"20\")\n",
    "# generates error:\n",
    "# TypeError: CustomClassGenerator.set_init.<locals>.__init__: incorrect type for 'test_attribute' (got 'str', expected 'int, int32 or int64')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "29df90cc-7f7d-4628-9c44-f6502bf2e14f",
   "metadata": {},
   "outputs": [],
   "source": [
    "TestDefaultValueInt2D = get_class(\"TestDefaultValueInt2D\", \"ndx-spec-non-string-value\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d5ebc1d3-386d-485d-90db-a81f01f4b6de",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "obj abc.TestDefaultValueInt2D at 0x4815575344\n",
       "Fields:\n",
       "  test_attribute: [[1 2 3]\n",
       " [4 5 6]]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "TestDefaultValueInt2D(name=\"obj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "fcf3ce5d-11e7-482c-ad55-76915139970f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[1, 2, 3], [4, 5, 6]]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "TestDefaultValueInt2D(name=\"obj\").test_attribute"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "12ed44c2-b2c8-4588-bec9-14cd67c57fb5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "errors: []\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.dtype[int64]'>\n",
      "[[1 2 3]\n",
      " [4 5 6]]\n"
     ]
    }
   ],
   "source": [
    "from pynwb import NWBFile, NWBHDF5IO, validate\n",
    "import datetime\n",
    "\n",
    "nwbfile = NWBFile(\n",
    "    session_description=\"session_description\",\n",
    "    identifier=\"identifier\",\n",
    "    session_start_time=datetime.datetime.now(datetime.timezone.utc),\n",
    ")\n",
    "\n",
    "obj = TestDefaultValueInt2D(name=\"obj\")\n",
    "nwbfile.add_acquisition(obj)\n",
    "\n",
    "filename = \"ndx-spec-non-string-value/test_out.nwb\"\n",
    "\n",
    "with NWBHDF5IO(filename, \"w\") as io:\n",
    "    io.write(nwbfile)\n",
    "\n",
    "with NWBHDF5IO(filename, \"r\") as io:\n",
    "    errors = validate(io)\n",
    "    print(\"errors:\", errors)\n",
    "    nwbfile = io.read()\n",
    "    print(type(nwbfile.acquisition[\"obj\"].test_attribute))\n",
    "    print(type(nwbfile.acquisition[\"obj\"].test_attribute.dtype))\n",
    "    print(nwbfile.acquisition[\"obj\"].test_attribute)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "120988e2-41bd-4c3b-8384-c35efbc9640c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/rly/mambaforge/envs/test/lib/python3.10/site-packages/hdmf/build/objectmapper.py:260: DtypeConversionWarning: Spec 'TestDefaultValueInt2D/test_attribute': Value with data type float64 is being converted to data type int64 as specified.\n",
      "  warnings.warn(full_warning_msg, DtypeConversionWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "errors: []\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.dtype[int64]'>\n",
      "[[1 2]\n",
      " [3 4]]\n"
     ]
    }
   ],
   "source": [
    "from pynwb import NWBFile, NWBHDF5IO, validate\n",
    "import datetime\n",
    "\n",
    "nwbfile = NWBFile(\n",
    "    session_description=\"session_description\",\n",
    "    identifier=\"identifier\",\n",
    "    session_start_time=datetime.datetime.now(datetime.timezone.utc),\n",
    ")\n",
    "\n",
    "obj = TestDefaultValueInt2D(name=\"obj\", test_attribute=[[1, 2.1], [3, 4]])\n",
    "nwbfile.add_acquisition(obj)\n",
    "\n",
    "filename = \"ndx-spec-non-string-value/test_out.nwb\"\n",
    "\n",
    "with NWBHDF5IO(filename, \"w\") as io:\n",
    "    io.write(nwbfile)\n",
    "\n",
    "with NWBHDF5IO(filename, \"r\") as io:\n",
    "    errors = validate(io)\n",
    "    print(\"errors:\", errors)\n",
    "    nwbfile = io.read()\n",
    "    print(type(nwbfile.acquisition[\"obj\"].test_attribute))\n",
    "    print(type(nwbfile.acquisition[\"obj\"].test_attribute.dtype))\n",
    "    print(nwbfile.acquisition[\"obj\"].test_attribute)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "ecbdc40a-9a5b-4781-aed6-1ba4f76e2bb6",
   "metadata": {},
   "outputs": [],
   "source": [
    "TestDefaultValueFloat2D = get_class(\"TestDefaultValueFloat2D\", \"ndx-spec-non-string-value\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "1bad18ec-6d9d-4e39-bd7a-464a54c30006",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "obj abc.TestDefaultValueFloat2D at 0x4816245264\n",
       "Fields:\n",
       "  test_attribute: [[1.  2.  3. ]\n",
       " [4.1 5.1 6.1]]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "TestDefaultValueFloat2D(name=\"obj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "b4661159-6e41-4cbb-84a4-ff65393cfc24",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "float"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(TestDefaultValueFloat2D(name=\"obj\").test_attribute[0][0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cf3aefd9-a0a7-48a0-9b46-7fd28563df07",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}