traverse agglomerative clustering sklearn

Dendrogram.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle as pkl\n",
    "import numpy as np\n",
    "import itertools\n",
    "from collections import namedtuple"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# From the Documentation\n",
    "```\n",
    "children_array-like of shape (n_samples-1, 2)\n",
    "The children of each non-leaf node. \n",
    "Values less than n_samples correspond to leaves of the tree which are the original samples. \n",
    "A node i greater than or equal to n_samples is a non-leaf node and has children children_[i - n_samples].\n",
    "Alternatively at the i-th iteration, children[i][0] and children[i][1] are merged to form node n_samples + i\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/daniel.kohlsdorf/anaconda3/lib/python3.7/site-packages/sklearn/base.py:318: UserWarning: Trying to unpickle estimator AgglomerativeClustering from version 0.21.3 when using version 0.22.2.post1. This might lead to breaking code or invalid results. Use at your own risk.\n",
      "  UserWarning)\n"
     ]
    }
   ],
   "source": [
    "agg       = pkl.load(open('models/v2_lstm_v5/agg.pkl', 'rb'))\n",
    "tree      = agg.children_ \n",
    "n_samples = agg.n_leaves_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13897\n",
      "11274\n"
     ]
    }
   ],
   "source": [
    "ii = itertools.count(n_samples)\n",
    "trees = {}\n",
    "for x in tree:\n",
    "    trees[next(ii)] = {'left': x[0], 'right':x[1]}\n",
    "print(len(trees))\n",
    "print(agg.n_clusters_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Node(namedtuple(\"Node\", \"i l r\")):\n",
    "    \n",
    "    def depth(self):\n",
    "        l = 0\n",
    "        r = 0\n",
    "        if isinstance(self.l, Node):\n",
    "            l =  1 + self.l.depth() \n",
    "        if isinstance(self.r, Node):\n",
    "            r =  1 + self.r.depth() \n",
    "        return max(l, r)\n",
    "        \n",
    "        \n",
    "def iterate_tree(node, trees, n_leafs, closed):\n",
    "    closed.add(node)\n",
    "    left    = trees[node]['left']\n",
    "    right   = trees[node]['right']\n",
    "    if left >= n_leafs:\n",
    "        left  = iterate_tree(left, trees, n_leafs, closed)\n",
    "    if right >= n_leafs:\n",
    "        right = iterate_tree(right, trees, n_leafs, closed)\n",
    "    return Node(node, left, right)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ".\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "444"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "closed   = set([])\n",
    "openl    = [k for k,_ in trees.items()]\n",
    "clusters = []\n",
    "while len(openl) > 0:\n",
    "    print('.')\n",
    "    max_node = max(openl)\n",
    "    cluster = iterate_tree(max_node, trees, n_samples, closed)\n",
    "    openl = [k for k,_ in trees.items() if k not in closed]\n",
    "    clusters.append(cluster)\n",
    "assert(len(closed) == len(trees))\n",
    "clusters[0].depth()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}