notebooks/generative_ai/large_language_models.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import bigframes.pandas\n",
    "import pandas as pd\n",
    "from bigframes.ml.llm import GeminiTextGenerator"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_604997/3896046356.py:1: ApiDeprecationWarning: gemini-pro and gemini-1.5-X are going to be deprecated. Use gemini-2.0-X (https://cloud.google.com/python/docs/reference/bigframes/latest/bigframes.ml.llm.GeminiTextGenerator) instead. \n",
      "  model = GeminiTextGenerator(model_name=\"gemini-1.5-flash-002\")\n",
      "/usr/local/google/home/shuowei/src/python-bigquery-dataframes/bigframes/ml/llm.py:981: DefaultLocationWarning: No explicit location is set, so using location US for the session.\n",
      "  self.session = session or global_session.get_global_session()\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Query job dd2da3cc-27c3-4c6f-9936-4f7769c85090 is DONE. 0 Bytes processed. <a target=\"_blank\" href=\"https://console.cloud.google.com/bigquery?project=bigframes-dev&j=bq:US:dd2da3cc-27c3-4c6f-9936-4f7769c85090&page=queryresults\">Open Job</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Query job 00947011-4d7c-42fa-ae19-3b684976cec6 is DONE. 0 Bytes processed. <a target=\"_blank\" href=\"https://console.cloud.google.com/bigquery?project=bigframes-dev&j=bq:US:00947011-4d7c-42fa-ae19-3b684976cec6&page=queryresults\">Open Job</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = GeminiTextGenerator(model_name=\"gemini-1.5-flash-002\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create Sample Data\n",
    "\n",
    "Read as a BigQuery DataFrames."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(\n",
    "        {\n",
    "            \"prompt\": [\"What is BigQuery?\", \"What is BQML?\", \"What is BigQuery DataFrame?\"],\n",
    "        })\n",
    "bf_df = bigframes.pandas.read_pandas(df)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "Query job 562ca203-3b53-4409-9a23-0a80d3840fcc is DONE. 0 Bytes processed. <a target=\"_blank\" href=\"https://console.cloud.google.com/bigquery?project=bigframes-dev&j=bq:US:562ca203-3b53-4409-9a23-0a80d3840fcc&page=queryresults\">Open Job</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/google/home/shuowei/src/python-bigquery-dataframes/bigframes/core/array_value.py:114: PreviewWarning: JSON column interpretation as a custom PyArrow extention in\n",
      "`db_dtypes` is a preview feature and subject to change.\n",
      "  warnings.warn(msg, bfe.PreviewWarning)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Query job 5a6ceff2-53b5-4a4a-83ff-31bffab1b8b8 is DONE. 14.0 kB processed. <a target=\"_blank\" href=\"https://console.cloud.google.com/bigquery?project=bigframes-dev&j=bq:US:5a6ceff2-53b5-4a4a-83ff-31bffab1b8b8&page=queryresults\">Open Job</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ml_generate_text_llm_result</th>\n",
       "      <th>ml_generate_text_rai_result</th>\n",
       "      <th>ml_generate_text_status</th>\n",
       "      <th>prompt</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BigQuery is a serverless, highly scalable, and...</td>\n",
       "      <td>&lt;NA&gt;</td>\n",
       "      <td></td>\n",
       "      <td>What is BigQuery?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BQML stands for **BigQuery Machine Learning**....</td>\n",
       "      <td>&lt;NA&gt;</td>\n",
       "      <td></td>\n",
       "      <td>What is BQML?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BigQuery DataFrames is a Python client library...</td>\n",
       "      <td>&lt;NA&gt;</td>\n",
       "      <td></td>\n",
       "      <td>What is BigQuery DataFrame?</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         ml_generate_text_llm_result  \\\n",
       "0  BigQuery is a serverless, highly scalable, and...   \n",
       "1  BQML stands for **BigQuery Machine Learning**....   \n",
       "2  BigQuery DataFrames is a Python client library...   \n",
       "\n",
       "  ml_generate_text_rai_result ml_generate_text_status  \\\n",
       "0                        <NA>                           \n",
       "1                        <NA>                           \n",
       "2                        <NA>                           \n",
       "\n",
       "                        prompt  \n",
       "0            What is BigQuery?  \n",
       "1                What is BQML?  \n",
       "2  What is BigQuery DataFrame?  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred = model.predict(bf_df).to_pandas()\n",
    "pred"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fetch Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"## BigQuery: A serverless data warehouse for large-scale data analysis\\n\\nBigQuery is a serverless, highly-scalable data warehouse designed for analyzing large datasets. It's a cloud-based service offered by Google Cloud Platform (GCP), allowing users to store, manage, and analyze massive amounts of data without managing infrastructure. \\n\\nHere are some key features of BigQuery:\\n\\n**Serverless:** You don't need to worry about provisioning, managing, or scaling servers. BigQuery handles all of this automatically, letting you focus on analyzing your data.\\n\\n**Highly-scalable:** BigQuery can handle datasets of any size, from gigabytes to petabytes. It can also scale up and down automatically to meet your processing needs.\\n\\n**Cost-effective:** You only pay for the resources you use, and there are no upfront costs. Additionally, BigQuery offers several pricing models to fit your needs, including on-demand, flat-rate, and flexible slots.\\n\\n**Easy to use:** BigQuery uses SQL, a standard query language, making it easy to analyze your data. No need to learn a new programming language.\\n\\n**Integrated with GCP:** BigQuery integrates seamlessly with other GCP services, such as Google Cloud Storage, Dataflow, and Kubernetes. This allows you to build powerful data pipelines and workflows.\\n\\n**Secure:** BigQuery uses industry-standard security practices to protect your data.\\n\\nHere are some use cases for BigQuery:\\n\\n* **Data warehousing and analytics:** Store and analyze large datasets for business intelligence and reporting.\\n* **Machine learning:** Train and deploy machine learning models on your data.\\n* **Data integration:** Combine data from multiple sources for analysis.\\n* **Real-time analytics:** Analyze data in real-time for insights and decision-making.\\n\\n**Here are some additional resources that you may find helpful:**\\n\\n* **BigQuery website:** https://cloud.google.com/bigquery\\n* **BigQuery documentation:** https://cloud.google.com/bigquery/docs\\n* **BigQuery tutorial:** https://cloud.google.com/bigquery/docs/tutorials\\n* **BigQuery pricing:** https://cloud.google.com/bigquery/pricing\\n\\nI hope this gives you a good overview of BigQuery. Please let me know if you have any other questions.\""
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred.iloc[0, 0]"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "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": "3.10.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}