Add ecoinvent migrations

dev/Generate ecoinvent mappings.ipynb

@@ -0,0 +1,187 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "03251142-6347-4c3b-9dab-ca61b3b8a552",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import randonneur as rd\n",
+    "from pathlib import Path\n",
+    "import randonneur_data as rdata"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "3065cefd-ec8a-4d69-bc76-b45973abeba7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "registry = rdata.Registry()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "8b1cda58-092e-4897-b1bd-821ce19649f3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def write_for_version(version, base_path, filename, sheet_name, combined):\n",
+    "    df = pd.read_excel(\n",
+    "        base_path / filename,\n",
+    "        sheet_name\n",
+    "    )    \n",
+    "    data = [\n",
+    "        {\n",
+    "            'source': {\n",
+    "                'identifier': row['Simapro Process Identifier'],\n",
+    "                'name': row['Simapro Process Name'],\n",
+    "                'platform_id': row['PlatformID'],\n",
+    "            },\n",
+    "            'target': {\n",
+    "                'filename': row[combined].lower() + \".spold\",\n",
+    "                'name': row['Activity Name'],\n",
+    "                'location': row['Geography'],\n",
+    "                'reference product': row['Reference Product Name'],\n",
+    "                'unit': row['Unit'],\n",
+    "            }\n",
+    "        } \n",
+    "        for row in df.to_dict(orient='records')\n",
+    "    ]\n",
+    "    dp = rd.Datapackage(\n",
+    "        name=f\"simapro-ecoinvent-{version}-cutoff\",\n",
+    "        description=f\"Data migration file from SimaPro 9 to ecoinvent-{version}-cutoff generated by PRé and provided via request at https://support.simapro.com/s/contactsupport\",\n",
+    "        contributors=[\n",
+    "            {\"title\": \"PRé\", \"path\": \"https://pre-sustainability.com/\", \"role\": \"author\"},\n",
+    "            {\"title\": \"Chris Mutel\", \"path\": \"https://chris.mutel.org/\", \"role\": \"wrangler\"},\n",
+    "        ],\n",
+    "        mapping_source=rd.MappingConstants.SIMAPRO_CSV,\n",
+    "        mapping_target=rd.MappingConstants.ECOSPOLD2,\n",
+    "        version=\"2.0.0\",\n",
+    "        source_id=\"SimaPro-9\",\n",
+    "        target_id=f\"ecoinvent-{version}-cutoff\"\n",
+    "    )\n",
+    "    dp.add_data(\"replace\", data)\n",
+    "    return registry.add_file(dp.to_json(Path(f\"simapro-ecoinvent-{version}-cutoff.json\")), replace=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "9afb0067-7343-46b4-beed-1fe7491281d9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "base_path = Path('/Users/cmutel/Projects/SimaPro - ecoinvent mapping/')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "8c30c0c9-9296-4527-aa74-c81fa0c9e12f",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "PosixPath('/Users/cmutel/Code/randonneur_data/randonneur_data/data/simapro-ecoinvent-3.10-cutoff.xz')"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "write_for_version(\n",
+    "    version=\"3.10\", \n",
+    "    base_path=base_path, \n",
+    "    filename=\"ecoinvent 3.10 Cut-off - SimaPro mapping.xlsx\", \n",
+    "    sheet_name=\"Mapping_Results_2024-04-25\", \n",
+    "    combined='Activity_UUID_Product_UUID',\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dbd86fa4-6142-461f-aa90-0aa650a44e0d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "write_for_version(\n",
+    "    version=\"3.8\", \n",
+    "    base_path=base_path, \n",
+    "    filename='ecoinvent cut-off 3.8 - SimaPro mapping file.xlsx', \n",
+    "    sheet_name=\"ecoinventCut-off38-SimaPro\", \n",
+    "    combined='Activity UUID & Product UUID'\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "08b41875-d606-4320-961d-15ce1bd94124",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "write_for_version(\n",
+    "    version=\"3.9.1\", \n",
+    "    base_path=base_path, \n",
+    "    filename='Ecoinvent_391_SimaPro_Mapping_Results_2023-10-03.xlsx', \n",
+    "    sheet_name=\"Mapping_Results_2023-10-03\", \n",
+    "    combined='Activity_UUID_Product_UUID'\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "66d21bb6-14f5-41d2-9e80-f9fd8256fe0d",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "`randonneur_data` registry with 3 files:\n",
+       "\tsimapro-ecoinvent-3.10-cutoff\n",
+       "\tsimapro-ecoinvent-3.8-cutoff\n",
+       "\tsimapro-ecoinvent-3.9.1-cutoff"
+      ]
+     },
+     "execution_count": 23,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "registry"
+   ]
+  }
+ ],
+ "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": "3.10.14"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

randonneur_data/__init__.py

@@ -3,7 +3,7 @@
     "Registry",
 )
 
-__version__ = "0.0.1"
+__version__ = "0.0.2"
 
 
 import json