Created
August 1, 2015 20:19
-
-
Save rbdixon/f86675aee788043824e4 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
{ | |
"cells": [ | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"# Question: How should I transform multiple key/value columns in a scikit-learn pipeline?\n", | |
"\n", | |
"See http://stackoverflow.com/questions/31749812/how-should-i-transform-multiple-key-value-columns-in-a-scikit-learn-pipeline/" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"Input data:" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 11, | |
"metadata": { | |
"collapsed": false | |
}, | |
"outputs": [ | |
{ | |
"name": "stdout", | |
"output_type": "stream", | |
"text": [ | |
" k1 v1 k2 v2\n", | |
"0 a 1 b 2\n", | |
"1 b 2 c 3\n" | |
] | |
} | |
], | |
"source": [ | |
"import pandas as pd\n", | |
"\n", | |
"D = pd.DataFrame([ ['a', 1, 'b', 2], ['b', 2, 'c', 3]], columns = ['k1', 'v1', 'k2', 'v2'])\n", | |
"print(D)" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"This is the type of output data that is required:" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 12, | |
"metadata": { | |
"collapsed": false | |
}, | |
"outputs": [ | |
{ | |
"name": "stdout", | |
"output_type": "stream", | |
"text": [ | |
"[{'a': 1, 'b': 2}, {'c': 3, 'b': 2}]\n" | |
] | |
}, | |
{ | |
"data": { | |
"text/plain": [ | |
"array([[ 1., 2., 0.],\n", | |
" [ 0., 2., 3.]])" | |
] | |
}, | |
"execution_count": 12, | |
"metadata": {}, | |
"output_type": "execute_result" | |
} | |
], | |
"source": [ | |
"from sklearn.feature_extraction import DictVectorizer\n", | |
"\n", | |
"row1 = {'a':1, 'b':2}\n", | |
"row2 = {'b':2, 'c':3}\n", | |
"data = [row1, row2]\n", | |
"print(data)\n", | |
"\n", | |
"DictVectorizer( sparse=False ).fit_transform(data)" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": { | |
"collapsed": true | |
}, | |
"source": [ | |
"# Solution\n", | |
"\n", | |
"Courtesy of [Mike](http://stackoverflow.com/users/2055368/mike): http://stackoverflow.com/a/31752733/1185562 and extended into a general pipeline transformer.\n", | |
"\n", | |
"Here is the transformer:" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 13, | |
"metadata": { | |
"collapsed": false | |
}, | |
"outputs": [], | |
"source": [ | |
"from sklearn.base import TransformerMixin\n", | |
"from sklearn.pipeline import Pipeline, FeatureUnion\n", | |
"\n", | |
"class KVExtractor(TransformerMixin):\n", | |
" def __init__(self, kvpairs):\n", | |
" self.kpairs = kvpairs\n", | |
" \n", | |
" def transform(self, X, *_):\n", | |
" result = []\n", | |
" for index, rowdata in X.iterrows():\n", | |
" rowdict = {}\n", | |
" for kvp in self.kpairs:\n", | |
" rowdict.update( { rowdata[ kvp[0] ]: rowdata[ kvp[1] ] } )\n", | |
" result.append(rowdict)\n", | |
" return result\n", | |
" \n", | |
" def fit(self, *_):\n", | |
" return self" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"Lets try it out:" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 14, | |
"metadata": { | |
"collapsed": false | |
}, | |
"outputs": [ | |
{ | |
"data": { | |
"text/plain": [ | |
"[{'a': 1, 'b': 2}, {'b': 2, 'c': 3}]" | |
] | |
}, | |
"execution_count": 14, | |
"metadata": {}, | |
"output_type": "execute_result" | |
} | |
], | |
"source": [ | |
"kvpairs = [ ['k1', 'v1'], ['k2', 'v2'] ]\n", | |
"KVExtractor( kvpairs ).transform(D)" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"Now try it out in a pipeline with `DictVectorizer`:" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 15, | |
"metadata": { | |
"collapsed": false | |
}, | |
"outputs": [ | |
{ | |
"name": "stdout", | |
"output_type": "stream", | |
"text": [ | |
" k1 v1 k2 v2\n", | |
"0 a 1 b 2\n", | |
"1 b 2 c 3\n", | |
"(2, 3)\n", | |
"[[ 1. 2. 0.]\n", | |
" [ 0. 2. 3.]]\n" | |
] | |
} | |
], | |
"source": [ | |
"pipeline = Pipeline(\n", | |
" [( 'kv', KVExtractor( kvpairs ) )] +\n", | |
" [( 'dv', DictVectorizer(sparse=False) )] +\n", | |
" []\n", | |
")\n", | |
"print(D)\n", | |
"A=pipeline.fit_transform(D)\n", | |
"print A.shape\n", | |
"print A" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"Try a new key without transforming:" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 16, | |
"metadata": { | |
"collapsed": false | |
}, | |
"outputs": [ | |
{ | |
"name": "stdout", | |
"output_type": "stream", | |
"text": [ | |
" k1 v1 k2 v2\n", | |
"0 a 1 x 2\n", | |
"1 b 2 c 3\n", | |
"[[ 1. 0. 0.]\n", | |
" [ 0. 2. 3.]]\n" | |
] | |
} | |
], | |
"source": [ | |
"D['k2'] = ['x', 'c']\n", | |
"print D\n", | |
"print pipeline.transform(D)" | |
] | |
}, | |
{ | |
"cell_type": "markdown", | |
"metadata": {}, | |
"source": [ | |
"Perfect!" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 17, | |
"metadata": { | |
"collapsed": false | |
}, | |
"outputs": [ | |
{ | |
"ename": "AttributeError", | |
"evalue": "'KVExtractor' object has no attribute 'inverse_transform'", | |
"output_type": "error", | |
"traceback": [ | |
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", | |
"\u001b[0;31mAttributeError\u001b[0m Traceback (most recent call last)", | |
"\u001b[0;32m<ipython-input-17-9e5154ad285e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mpipeline\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minverse_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mA\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m", | |
"\u001b[0;32m/opt/boxen/homebrew/lib/python2.7/site-packages/sklearn/utils/metaestimators.pyc\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m 35\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_attribute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 36\u001b[0m \u001b[0;31m# lambda, but not partial, allows help() to work with update_wrapper\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 37\u001b[0;31m \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 38\u001b[0m \u001b[0;31m# update the docstring of the returned function\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 39\u001b[0m \u001b[0mupdate_wrapper\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", | |
"\u001b[0;32m/opt/boxen/homebrew/lib/python2.7/site-packages/sklearn/pipeline.pyc\u001b[0m in \u001b[0;36minverse_transform\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m 265\u001b[0m \u001b[0mXt\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mX\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 266\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstep\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msteps\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 267\u001b[0;31m \u001b[0mXt\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstep\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minverse_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 268\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mXt\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 269\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n", | |
"\u001b[0;31mAttributeError\u001b[0m: 'KVExtractor' object has no attribute 'inverse_transform'" | |
] | |
} | |
], | |
"source": [ | |
"pipeline.inverse_transform(A)" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": null, | |
"metadata": { | |
"collapsed": true | |
}, | |
"outputs": [], | |
"source": [] | |
} | |
], | |
"metadata": { | |
"kernelspec": { | |
"display_name": "Python 2", | |
"language": "python", | |
"name": "python2" | |
}, | |
"language_info": { | |
"codemirror_mode": { | |
"name": "ipython", | |
"version": 2 | |
}, | |
"file_extension": ".py", | |
"mimetype": "text/x-python", | |
"name": "python", | |
"nbconvert_exporter": "python", | |
"pygments_lexer": "ipython2", | |
"version": "2.7.10" | |
} | |
}, | |
"nbformat": 4, | |
"nbformat_minor": 0 | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Explained in an elegant manner. Thank you!