Text Extensions for Pandas: Tips and Techniques for Extending Pandas, Part 1 Blog

tep_extending_pandas_blog1_2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'text_extensions_for_pandas.array.span.SpanArray'>\n",
      "<class 'text_extensions_for_pandas.array.tensor.TensorArray'>\n"
     ]
    }
   ],
   "source": [
    "# Extract `SpanArray` and `TensorArray` from columns.\n",
    "spans = df[\"span\"].array\n",
    "embeddings = df[\"embedding\"].array\n",
    "print(f\"{type(spans)}\\n{type(embeddings)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Column 'embedding'\n",
      "length: 9\n",
      "array shape: (9, 768)\n",
      "tensor shape: (768,)\n"
     ]
    }
   ],
   "source": [
    "# Print some info about the \"embedding\" column\n",
    "length = len(embeddings)\n",
    "tensor_array_shape = embeddings.numpy_shape\n",
    "tensor_shape = embeddings[0].to_numpy().shape\n",
    "print(f\"Column 'embedding'\\nlength: {length}\\narray shape: {tensor_array_shape}\\ntensor shape: {tensor_shape}\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}