davinci-003 vs ada-002

txt2vec-openai-nearText.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import weaviate\n",
    "import uuid\n",
    "import os\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Client"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "client = weaviate.Client(\"https://foo.weaviate.network\",\n",
    "                         additional_headers={\"X-OpenAI-Api-Key\": os.environ[\"OPENAI_APIKEY\"]})\n",
    "client.schema.delete_all()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Upload some stuff to weaviate:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with davinci 003\n",
    "doc_class_schema = {\n",
    "    \"class\": \"Document\",\n",
    "    \"description\": \"A class called document\",\n",
    "    \"vectorizer\": \"text2vec-openai\",\n",
    "    \"moduleConfig\": {\n",
    "        \"text2vec-openai\": {\n",
    "            \"model\": \"davinci\",\n",
    "            \"modelVersion\": \"003\",\n",
    "            \"type\": \"text\"\n",
    "        }\n",
    "    },\n",
    "    \"properties\": [\n",
    "        {\n",
    "            \"dataType\": [\n",
    "                \"text\"\n",
    "            ],\n",
    "            \"description\": \"Content that will be vectorized\",\n",
    "            \"moduleConfig\": {\n",
    "                \"text2vec-openai\": {\n",
    "                    \"skip\": \"false\",\n",
    "                    \"vectorizePropertyName\": \"false\"\n",
    "                }\n",
    "            },\n",
    "            \"name\": \"content\"\n",
    "        }\n",
    "    ]\n",
    "}\n",
    "\n",
    "# with ada 002\n",
    "# doc_class_schema = {\n",
    "#     \"class\": \"Document\",\n",
    "#     \"description\": \"A class called document\",\n",
    "#     \"vectorizer\": \"text2vec-openai\",\n",
    "#     \"moduleConfig\": {\n",
    "#         \"text2vec-openai\": {\n",
    "#             \"model\": \"ada\",\n",
    "#             \"modelVersion\": \"002\",\n",
    "#             \"type\": \"text\"\n",
    "#         }\n",
    "#     },\n",
    "#     \"properties\": [\n",
    "#         {\n",
    "#             \"dataType\": [\n",
    "#                 \"text\"\n",
    "#             ],\n",
    "#             \"description\": \"Content that will be vectorized\",\n",
    "#             \"moduleConfig\": {\n",
    "#                 \"text2vec-openai\": {\n",
    "#                     \"skip\": \"false\",\n",
    "#                     \"vectorizePropertyName\": \"false\"\n",
    "#                 }\n",
    "#             },\n",
    "#             \"name\": \"content\"\n",
    "#         }\n",
    "#     ]\n",
    "# }\n",
    "\n",
    "client.schema.create_class(doc_class_schema)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "client.batch(batch_size=10, dynamic=True, num_workers=1)\n",
    "\n",
    "with client.batch as batch:\n",
    "    batch.add_data_object(\n",
    "        data_object={\n",
    "            \"content\": \"What accounting software do we use?\",\n",
    "        },\n",
    "        class_name=\"Document\"\n",
    "    )\n",
    "\n",
    "    batch.add_data_object(\n",
    "        data_object={\n",
    "            \"content\": \"What software do we use for accounting?\",\n",
    "        },\n",
    "        class_name=\"Document\"\n",
    "    )\n",
    "\n",
    "    batch.add_data_object(\n",
    "        data_object={\n",
    "            \"content\": \"What accounting software does the company use?\",\n",
    "        },\n",
    "        class_name=\"Document\"\n",
    "    )\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Queries"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get vector for \"What accounting software do we use?\":"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "where_filter = {\n",
    "    \"path\": [\"content\"],\n",
    "    \"operator\": \"Equal\",\n",
    "    \"valueText\": \"What accounting software do we use?\"\n",
    "}\n",
    "\n",
    "query_result = (\n",
    "    client.query\n",
    "    .get(\"Document\", \"content\")\n",
    "    .with_where(where_filter)\n",
    "    .with_additional([\"id\"])\n",
    "    .do()\n",
    ")\n",
    "\n",
    "uuid = query_result[\"data\"][\"Get\"][\"Document\"][0][\"_additional\"][\"id\"]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[-0.005881898,\n",
       " 0.01804839,\n",
       " 0.0030804658,\n",
       " -0.015143907,\n",
       " 0.011484764,\n",
       " -0.0017613986,\n",
       " -0.013659956,\n",
       " -0.0007653602,\n",
       " -0.01143403,\n",
       " 0.0058914106]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "where_filter = {\n",
    "    \"path\": [\"id\"],\n",
    "    \"operator\": \"Equal\",\n",
    "    \"valueString\": uuid\n",
    "}\n",
    "\n",
    "query_result = (\n",
    "    client.query\n",
    "    .get(\"Document\", \"content\")\n",
    "    .with_where(where_filter)\n",
    "    .with_additional([\"vector\"])\n",
    "    .do()\n",
    ")\n",
    "\n",
    "vector = query_result[\"data\"][\"Get\"][\"Document\"][0][\"_additional\"][\"vector\"]\n",
    "vector[:10]\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Find documents near `vector`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'_additional': {'distance': -4.7683716e-07}, 'content': 'What accounting software do we use?'}\n",
      "{'_additional': {'distance': 0.028079808}, 'content': 'What software do we use for accounting?'}\n",
      "{'_additional': {'distance': 0.052015245}, 'content': 'What accounting software does the company use?'}\n"
     ]
    }
   ],
   "source": [
    "nearVector = {\n",
    "    'vector': vector\n",
    "}\n",
    "\n",
    "result = (\n",
    "    client.query\n",
    "    .get(\"Document\", \"content\")\n",
    "    .with_additional('distance')\n",
    "    .with_near_vector(nearVector)\n",
    "    .do()\n",
    ")\n",
    "\n",
    "docs = result[\"data\"][\"Get\"][\"Document\"]\n",
    "\n",
    "for doc in docs:\n",
    "    print(doc)\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Find documents near \"What software do we use for accounting?\":"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "nearText = {\n",
    "    \"concepts\": [\"What software do we use for accounting\"],\n",
    "    \"distance\": 0.6\n",
    "}\n",
    "\n",
    "result = (\n",
    "    client.query\n",
    "    .get(\"Document\", [\"content\", \"_additional {distance} \"])\n",
    "    .with_near_text(nearText)\n",
    "    .do()\n",
    ")\n",
    "\n",
    "docs = result[\"data\"][\"Get\"][\"Document\"]\n",
    "\n",
    "for doc in docs:\n",
    "    print(doc)\n"
   ]
  }
 ],
 "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": "3.8.16"
  },
  "vscode": {
   "interpreter": {
    "hash": "949777d72b0d2535278d3dc13498b2535136f6dfe0678499012e853ee9abcab1"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}