word2vec 001.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\bridge\\Anaconda3\\envs\\env01\\lib\\site-packages\\gensim\\utils.py:860: UserWarning: detected Windows; aliasing chunkize to chunkize_serial\n",
      "  warnings.warn(\"detected Windows; aliasing chunkize to chunkize_serial\")\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import gensim\n",
    "from matplotlib import pyplot as plt\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "train_orders = pd.read_csv(\"../Untitled Folder/order_products__train.csv\")\n",
    "prior_orders = pd.read_csv(\"../Untitled Folder/order_products__prior.csv\")\n",
    "products = pd.read_csv(\"../Untitled Folder/products.csv\").set_index('product_id')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1384617, 4)\n",
      "(32434489, 4)\n"
     ]
    }
   ],
   "source": [
    "print(train_orders.shape)\n",
    "print(prior_orders.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "131209 unique orders in train\n",
      "3214874 unique orders in prior\n",
      "39123 unique products in train\n",
      "49677 unique products in prior\n"
     ]
    }
   ],
   "source": [
    "print(len(np.unique(train_orders.order_id)), \"unique orders in train\")\n",
    "print(len(np.unique(prior_orders.order_id)), \"unique orders in prior\")\n",
    "print(len(np.unique(train_orders.product_id)), \"unique products in train\")\n",
    "print(len(np.unique(prior_orders.product_id)), \"unique products in prior\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "train_orders[\"product_id\"] = train_orders[\"product_id\"].astype(str)\n",
    "prior_orders[\"product_id\"] = prior_orders[\"product_id\"].astype(str)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 4min 3s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "train_products = train_orders.groupby(\"order_id\").apply(lambda order: order['product_id'].tolist())\n",
    "prior_products = prior_orders.groupby(\"order_id\").apply(lambda order: order['product_id'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "order_id\n",
       "1     [49302, 11109, 10246, 49683, 43633, 13176, 472...\n",
       "36    [39612, 19660, 49235, 43086, 46620, 34497, 486...\n",
       "38    [11913, 18159, 4461, 21616, 23622, 32433, 2884...\n",
       "96    [20574, 30391, 40706, 25610, 27966, 24489, 39275]\n",
       "98    [8859, 19731, 43654, 13176, 4357, 37664, 34065...\n",
       "dtype: object"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_products.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "order_id\n",
       "2    [33120, 28985, 9327, 45918, 30035, 17794, 4014...\n",
       "3    [33754, 24838, 17704, 21903, 17668, 46667, 174...\n",
       "4    [46842, 26434, 39758, 27761, 10054, 21351, 225...\n",
       "5    [13176, 15005, 47329, 27966, 23909, 48370, 132...\n",
       "6                                [40462, 15873, 41897]\n",
       "dtype: object"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prior_products.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 1.17 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "sentences = prior_products.append(train_products)\n",
    "longest = np.max(sentences.apply(len))\n",
    "sentences = sentences.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3346083,)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(sentences)\n",
    "sentences.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ list(['33120', '28985', '9327', '45918', '30035', '17794', '40141', '1819', '43668']),\n",
       "       list(['33754', '24838', '17704', '21903', '17668', '46667', '17461', '32665']),\n",
       "       list(['49235', '13565', '14233', '35548']),\n",
       "       list(['35951', '16953', '4724'])], dtype=object)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sentences[np.r_[0:2, -2:0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 4min 45s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "model = gensim.models.Word2Vec(sentences, size=100, window=longest, min_count=2, workers=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 13 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "vocab = list(model.wv.vocab.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['33120',\n",
       " '28985',\n",
       " '9327',\n",
       " '45918',\n",
       " '30035',\n",
       " '17794',\n",
       " '40141',\n",
       " '1819',\n",
       " '43668',\n",
       " '33754']"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#vocab[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('49075', 0.7379200458526611),\n",
       " ('5646', 0.7055116891860962),\n",
       " ('28289', 0.6989336609840393),\n",
       " ('11090', 0.696395754814148),\n",
       " ('17878', 0.6666847467422485),\n",
       " ('10017', 0.6618810892105103),\n",
       " ('4799', 0.6419677734375),\n",
       " ('41328', 0.6361633539199829),\n",
       " ('33401', 0.6323817372322083),\n",
       " ('21543', 0.627264142036438)]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#display(model.most_similar('33120'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'inverse_dictionary' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<timed exec>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'inverse_dictionary' is not defined"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "def tfidf_word2vec_transform(model, embed_dim, inv_dic_tfidf, score_tfidf):\n",
    "    tfidf_word2vec = np.zeros((score_tfidf.shape[0], embed_dim))\n",
    "    for i in range(score_tfidf.shape[0]):\n",
    "        doc_word2vec = np.zeros(embed_dim)\n",
    "        for j in range(score_tfidf.shape[1]):\n",
    "            doc_word2vec += score_tfidf[i, j] * model[inv_dic_tfidf[j]]\n",
    "        tfidf_word2vec[i, :] = doc_word2vec\n",
    "    return tfidf_word2vec\n",
    "\n",
    "X_train_w2v = tfidf_word2vec_transform(model, 64, inverse_dictionary, X_train)\n",
    "X_test_w2v = tfidf_word2vec_transform(model, 64, inverse_dictionary, X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 329 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "pca = PCA(n_components=2)\n",
    "pca.fit(model.wv.syn0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 0 ns\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "def get_batch(vocab, model, n_batches=3):\n",
    "    output = list()\n",
    "    for i in range(0, n_batches):\n",
    "        rand_int = np.random.randint(len(vocab), size=1)[0]\n",
    "        suggestions = model.most_similar(positive=[vocab[rand_int]], topn=5)\n",
    "        suggest = list()\n",
    "        for i in suggestions:\n",
    "            suggest.append(i[0])\n",
    "        output += suggest\n",
    "        output.append(vocab[rand_int])\n",
    "    return output\n",
    "\n",
    "def plot_with_labels(low_dim_embs, labels, filename='tsne.png'):\n",
    "    \"\"\"From Tensorflow's tutorial.\"\"\"\n",
    "    assert low_dim_embs.shape[0] >= len(labels), \"More labels than embeddings\"\n",
    "    plt.figure(figsize=(18, 18))  #in inches\n",
    "    for i, label in enumerate(labels):\n",
    "        x, y = low_dim_embs[i,:]\n",
    "        plt.scatter(x, y)\n",
    "        plt.annotate(label,\n",
    "                     xy=(x, y),\n",
    "                     xytext=(5, 2),\n",
    "                     textcoords='offset points',\n",
    "                     ha='right',\n",
    "                     va='bottom')\n",
    "#     plt.savefig(filename)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28f2f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%time\n",
    "embeds = []\n",
    "labels = []\n",
    "for item in get_batch(vocab, model, n_batches=3):\n",
    "    embeds.append(model[item])\n",
    "    labels.append(products.loc[int(item)]['product_name'])\n",
    "embeds = np.array(embeds)\n",
    "embeds = pca.fit_transform(embeds)\n",
    "plot_with_labels(embeds, labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "model.save(\"product2vec.model\")"
   ]
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}