datreant tag selection parser

datreant-parser.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import datreant.core as dtr\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "for treetype in ('birch', 'oak'):\n",
    "    for branch in ('first', 'random'):\n",
    "        for sel in ('all', 'wood', 'core'):\n",
    "            folder = os.path.join('forest', '{}_{}_{}'.format(treetype, branch, sel))\n",
    "            tree = dtr.Treant(folder)\n",
    "            tree.tags = [treetype, branch, sel]\n",
    "        folder = os.path.join('forest', '{}_{}'.format(treetype, branch))\n",
    "        tree = dtr.Treant(folder)\n",
    "        tree.tags = [treetype, branch]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "bundle = dtr.discover('forest/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "birch_first\t  birch_random\t     oak_first\t     oak_random\r\n",
      "birch_first_all   birch_random_all   oak_first_all   oak_random_all\r\n",
      "birch_first_core  birch_random_core  oak_first_core  oak_random_core\r\n",
      "birch_first_wood  birch_random_wood  oak_first_wood  oak_random_wood\r\n"
     ]
    }
   ],
   "source": [
    "!ls forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_first\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_random\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_random_core\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_first_core\n"
     ]
    }
   ],
   "source": [
    "treants = bundle[bundle.tags[['oak', {'all'}, {'wood'}]]]\n",
    "\n",
    "for t in treants:\n",
    "    print(t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from pyparsing import (CaselessLiteral, Word, alphas, quotedString,\n",
    "                       removeQuotes, operatorPrecedence, opAssoc, stringEnd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class UnaryOperation(object):\n",
    "    def __init__(self, t):\n",
    "        self.op, self.a = t[0]\n",
    "\n",
    "\n",
    "class BinaryOperation(object):\n",
    "    def __init__(self, t):\n",
    "        self.op = t[0][1]\n",
    "        self.operands = t[0][0::2]\n",
    "\n",
    "\n",
    "class SearchAnd(BinaryOperation):\n",
    "    def generateTagExpression(self):\n",
    "        return list(oper.generateTagExpression() for oper in self.operands)\n",
    "\n",
    "    def __repr__(self):\n",
    "        return \"AND:(%s)\" % (\",\".join(str(oper) for oper in self.operands))\n",
    "\n",
    "\n",
    "class SearchOr(BinaryOperation):\n",
    "    def generateTagExpression(self):\n",
    "        return tuple(oper.generateTagExpression() for oper in self.operands)\n",
    "\n",
    "\n",
    "    def __repr__(self):\n",
    "        return \"OR:(%s)\" % (\",\".join(str(oper) for oper in self.operands))\n",
    "\n",
    "\n",
    "class SearchNot(UnaryOperation):\n",
    "    def generateTagExpression(self):\n",
    "        exps = self.a.generateTagExpression()\n",
    "        if isinstance(exps, list):\n",
    "            return [{e} for e in exps]\n",
    "        elif isinstance(exps, tuple):\n",
    "            return tuple({e} for e in exps)\n",
    "        else:\n",
    "            return {self.a.generateTagExpression()}\n",
    "\n",
    "    def __repr__(self):\n",
    "        return \"NOT:(%s)\" % str(self.a)\n",
    "\n",
    "\n",
    "class SearchTerm(object):\n",
    "    def __init__(self, tokens):\n",
    "        self.term = tokens[0]\n",
    "\n",
    "    def generateTagExpression(self):\n",
    "        return \"{}\".format(self.term)\n",
    "\n",
    "    def __repr__(self):\n",
    "        return self.term\n",
    "\n",
    "\n",
    "# define the grammar\n",
    "and_ = CaselessLiteral(\"and\")\n",
    "or_ = CaselessLiteral(\"or\")\n",
    "not_ = CaselessLiteral(\"not\")\n",
    "searchTerm = Word(alphas) | quotedString.setParseAction(removeQuotes)\n",
    "searchTerm.setParseAction(SearchTerm)\n",
    "searchExpr = operatorPrecedence(searchTerm, [\n",
    "    (not_, 1, opAssoc.RIGHT, SearchNot),\n",
    "    (and_, 2, opAssoc.LEFT, SearchAnd),\n",
    "    (or_, 2, opAssoc.LEFT, SearchOr),\n",
    "])\n",
    "Parser = searchExpr + stringEnd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AND:(oak,NOT:(AND:(all,wood)))\n",
      "['oak', [{'all'}, {'wood'}]]\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_first\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_random\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_random_core\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_first_core\n"
     ]
    }
   ],
   "source": [
    "sle_str = 'oak and not all and not wood'\n",
    "sel_str = 'not (all and wood) and oak'\n",
    "sel_str = 'oak and not (all and wood)'\n",
    "\n",
    "\n",
    "sel_ast = Parser.parseString(sel_str)[0]\n",
    "print(sel_ast)\n",
    "sel = sel_ast.generateTagExpression()\n",
    "print(sel)\n",
    "\n",
    "#treants = bundle[bundle.tags[['oak', {'all'}, {'wood'}]]]\n",
    "treants = bundle[bundle.tags[sel]]\n",
    "\n",
    "for t in treants:\n",
    "    print(t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AND:(oak,OR:(all,wood))\n",
      "['oak', ('all', 'wood')]\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_random_all\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_first_wood\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_random_wood\n",
      "/home/max/Academia/Projects/datreant-pyparsing/forest/oak_first_all\n"
     ]
    }
   ],
   "source": [
    "sle_str = 'oak and not all and not wood'\n",
    "sel_str = 'not (all and wood) and oak'\n",
    "#sel_str = 'oak and not (all and wood)'\n",
    "sel_str = 'oak and (all or wood)'\n",
    "\n",
    "\n",
    "sel_ast = Parser.parseString(sel_str)[0]\n",
    "print(sel_ast)\n",
    "sel = sel_ast.generateTagExpression()\n",
    "print(sel)\n",
    "\n",
    "#treants = bundle[bundle.tags[['oak', {'all'}, {'wood'}]]]\n",
    "treants = bundle[bundle.tags[sel]]\n",
    "\n",
    "for t in treants:\n",
    "    print(t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.6.1"
  },
  "toc": {
   "colors": {
    "hover_highlight": "#DAA520",
    "navigate_num": "#000000",
    "navigate_text": "#333333",
    "running_highlight": "#FF0000",
    "selected_highlight": "#FFD700",
    "sidebar_border": "#EEEEEE",
    "wrapper_background": "#FFFFFF"
   },
   "moveMenuLeft": true,
   "nav_menu": {
    "height": "12px",
    "width": "252px"
   },
   "navigate_menu": true,
   "number_sections": true,
   "sideBar": true,
   "threshold": 4,
   "toc_cell": false,
   "toc_section_display": "block",
   "toc_window_display": false,
   "widenNotebook": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}