From 008180962f62ba87cfff6b23cd6e95ed85adda4f Mon Sep 17 00:00:00 2001 From: Thorsten Vitt Date: Wed, 19 Apr 2017 16:32:42 +0200 Subject: [PATCH] Customizing the pipeline --- docs/CustomizingPipeline.ipynb | 2253 ++++++++++++++++++++++++++++++++ docs/GettingStarted.ipynb | 5 +- docs/index.rst | 1 + 3 files changed, 2258 insertions(+), 1 deletion(-) create mode 100644 docs/CustomizingPipeline.ipynb diff --git a/docs/CustomizingPipeline.ipynb b/docs/CustomizingPipeline.ipynb new file mode 100644 index 0000000..6c1f0cf --- /dev/null +++ b/docs/CustomizingPipeline.ipynb @@ -0,0 +1,2253 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Customizing the pipeline\n", + "\n", + "### Reading the corpus\n", + "\n", + "The Corpus class that represents the feature matrix which is the starting point for the analysis basically is a data frame that maps combinations of documents and features to feature counts. When _building_ the representation, it uses two helper classes to decouple aspects that are relevant for building the feature matrix: FeatureGenerator performs the actual reading and feature counting, DocumentDescriber provides labels and groupings for the documents that can be used in visualization and evaluation steps.\n", + "\n", + "When you just create a corpus using `delta.Corpus(directory)`, default versions will be used, but stuff can be customized:" + ] + }, + { + "cell_type": "code", + "execution_count": 1, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Help on function __init__ in module delta.corpus:\n", + "\n", + "__init__(self, subdir=None, file=None, corpus=None, feature_generator=FeatureGenerator(encoding='utf-8', token_pattern=regex.Regex('\\\\p{L}+', flags=regex.V0), lower_case=False, max_tokens=None, glob='*.txt', skip=None, ngrams=None), document_describer=, metadata=None, **kwargs)\n", + " Creates a new Corpus.\n", + " \n", + " Args:\n", + " subdir (str): Path to a subdirectory containing the (unprocessed) corpus data.\n", + " file (str): Path to a CSV file containing the feature vectors.\n", + " corpus (pandas.DataFrame): A dataframe or :class:`Corpus` from which to create a new corpus, as a copy.\n", + " feature_generator (FeatureGenerator): A customizeable helper class that will process a `subdir` to a feature matrix, if the `subdir` argument is also given.\n", + " metadata (dict): A dictionary with metadata to copy into the new corpus.\n", + " **kwargs: Additional keyword arguments will be set in the metadata record of the new corpus.\n", + "\n" + ] + } + ], + "source": [ + "import delta\n", + "help(delta.Corpus.__init__)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "E.g., to create a corpus that ignores case differences and reads only the first 5000 tokens of each text:" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
wirdmutterihnenbiszeitihrenetwaswurdedirkonnte
Arnim,-Ludwig-Achim-von_Armut Reichtum Schuld und Busse der Graefin Dolores1.04.018.09.05.07.08.09.02.02.0
Arnim,-Ludwig-Achim-von_Isabella von Aegypten8.04.06.08.04.08.06.06.09.013.0
Arnim,-Ludwig-Achim-von_Kronenwaechter 115.01.03.02.010.01.04.02.06.01.0
Dohm,-Hedwig_Christa Ruland2.017.06.08.02.020.013.011.01.07.0
Dohm,-Hedwig_Schicksale einer Seele8.022.05.017.07.03.07.09.07.06.0
\n", + "
" + ], + "text/plain": [ + " wird mutter ihnen bis \\\n", + "Arnim,-Ludwig-Achim-von_Armut Reichtum Schuld u... 1.0 4.0 18.0 9.0 \n", + "Arnim,-Ludwig-Achim-von_Isabella von Aegypten 8.0 4.0 6.0 8.0 \n", + "Arnim,-Ludwig-Achim-von_Kronenwaechter 1 15.0 1.0 3.0 2.0 \n", + "Dohm,-Hedwig_Christa Ruland 2.0 17.0 6.0 8.0 \n", + "Dohm,-Hedwig_Schicksale einer Seele 8.0 22.0 5.0 17.0 \n", + "\n", + " zeit ihren etwas wurde \\\n", + "Arnim,-Ludwig-Achim-von_Armut Reichtum Schuld u... 5.0 7.0 8.0 9.0 \n", + "Arnim,-Ludwig-Achim-von_Isabella von Aegypten 4.0 8.0 6.0 6.0 \n", + "Arnim,-Ludwig-Achim-von_Kronenwaechter 1 10.0 1.0 4.0 2.0 \n", + "Dohm,-Hedwig_Christa Ruland 2.0 20.0 13.0 11.0 \n", + "Dohm,-Hedwig_Schicksale einer Seele 7.0 3.0 7.0 9.0 \n", + "\n", + " dir konnte \n", + "Arnim,-Ludwig-Achim-von_Armut Reichtum Schuld u... 2.0 2.0 \n", + "Arnim,-Ludwig-Achim-von_Isabella von Aegypten 9.0 13.0 \n", + "Arnim,-Ludwig-Achim-von_Kronenwaechter 1 6.0 1.0 \n", + "Dohm,-Hedwig_Christa Ruland 1.0 7.0 \n", + "Dohm,-Hedwig_Schicksale einer Seele 7.0 6.0 " + ] + }, + "execution_count": 2, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "corpus = delta.Corpus('../../refcor/German', feature_generator=delta.FeatureGenerator(lower_case=True, max_tokens=5000))\n", + "corpus.ix[:5,100:110]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "It is also possible to roll your own feature generator. As an example, we would like to use lemmas instead of word forms as features. For that, we would like to read a version of our corpus that has been preprocessed using the DARIAH DKPro wrapper.\n", + "\n", + "A file like this is essentially a table that contains a line per token with various analyses in the columns:" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
SectionIdParagraphIdSentenceIdTokenIdBeginEndTokenLemmaCPOSPOS...MorphologyHyphenationDependencyHeadDependencyRelationNamedEntityQuoteMarkerCoreferenceChainIdsSyntaxTreePredicateSemanticArgumentIndex
1003_824100358265836ThatsachenThatsachenNNNN..._That-sa-chen1002NK_0____
1004_824100458375844annimmtannehmenVVVFIN..._an-nimmt991CJ_0____
1005_824100558445845,,PUNC$,..._,1004--_0____
1006_824100658465848sosoADVADV..._so1007MO_0____
1007_824100758495853wirdwerdenVVAFIN..._wird-1--_0____
1008_824100858545856ererPRPPER..._er1007SB_0____
1009_824100958575861sichsichPRPRF..._sich1030OA_0____
1010_824101058625866dochdochADVADV..._doch1030MO_0____
1011_824101158665867,,PUNC$,..._,1010--_0____
1012_824101258685872ebenebenADVADV..._eben1030MO_0____
1013_824101358735875sosoADVADV..._so1014MO_0____
1014_824101458765881wenigwenigePRPIS..._we-nig1030MO_0____
1015_824101558815882,,PUNC$,..._,1014--_0____
1016_824101658835886aufaufPPAPPR..._auf1030MO_0____
1017_824101758875890diedieARTART..._die1019NK_0____
1018_824101858915904verdrießlicheverdrießlichADJADJA..._ver-drieß-li-che1019NK_0____
1019_824101959055915BetheurungBetheurungNNNN..._Be-theu-rung1016NK_0____
1020_824102059165919derdieARTART..._der1021NK_0____
1021_824102159205929WärterinnWärterinnNNNN..._Wär-te-rinn1019AG_0____
1022_824102259295930,,PUNC$,..._,1021--_0____
1023_824102359315934daßdassCONJKOUS..._daß1027CP_0____
1024_824102459355941diesesdiesPRPDAT..._die-ses1027NK_0____
1025_824102559425947allesallePRPIAT..._al-les1026NK_0____
1026_824102659485958FieberwahnFieberwahnNNNN..._Fie-ber-wahn1027NK_0____
1027_824102759595962seyseyNPNE..._sey1013CC_0____
1028_824102859625963,,PUNC$,..._,1027--_0____
1029_824102959645979schlechterdingsschlechterdingsADVADV..._schlech-ter-dings1030MO_0____
1030_824103059805989verlassenverlassenVVVPP..._ver-las-sen1007OC_0____
1031_824103159895990;;PUNC$...._;1030--_0____
1032_824103259915993ererPRPPER..._er1033SB_0____
..................................................................
9980_2631199805549155495seinseinPRPPOSAT..._sein9981NK_0____
9981_2631199815549655503GesichtGesichtNNNN..._Ge-sicht9982OA_0____
9982_2631199825550455510flammtflammenVVVFIN..._flammt9976MO_0____
9983_2631199835551055511,,PUNC$,..._,9982--_0____
9984_2631199845551255517LuiseLuiseNPNE..._Lui-se9985OAB-PER0____
9985_2631199855551855525schlägtschlagenVVVFIN..._schlägt9976CJ_0____
9986_2631199865552655529diedieARTART..._die9987NK_0____
9987_2631199875553055535AugenAugeNNNN..._Au-gen9985OA_0____
9988_2631199885553655542niederniederVPTKVZ..._nie-der9985SVP_0____
9989_2631199895554255543,,PUNC$,..._,9988--_0____
9990_2631199905554455547undundCONJKON..._und9985CD_0____
9991_2631199915554855559beschuldigtbeschuldigenVVVFIN..._be-schul-digt9990CJ_0____
9992_2631199925556055564sichsichPRPRF..._sich9991OA_0____
9993_2631199935556555568derdieARTART..._der9994NK_0____
9994_2631199945556955582UndankbarkeitUndankbarkeitNNNN..._Un-dank-bar-keit9991OG_0____
9995_2631199955558255583..PUNC$...._.9994--_0____
9996_2631299965558455587SieSie|siePRPPER..._Sie9997SB_0____
9997_2631299975558855594wolltewollenVVMFIN..._woll-te-1--_0____
9998_2631299985559555605wenigstenswenigstensADVADV..._we-nigs-tens9997MO_0____
9999_2631299995560655610ihreihrPRPPOSAT..._ih-re10000NK_0____
10000_26312100005561155617SchuldSchuldNNNN..._Schuld10004OA_0____
10001_26312100015561855623durchdurchPPAPPR..._durch10004MO_0____
10002_26312100025562455629keinekeinePRPIAT..._kei-ne10003NK_0____
10003_26312100035563055639HeucheleyHeucheleyNNNN..._Heu-che-ley10001NK_0____
10004_26312100045564055649vermehrenvermehrenVVVINF..._ver-meh-ren9997OC_0____
10005_26312100055564955650;;PUNC$...._;10004--_0____
10006_26312100065565155654siesiePRPPER..._sie10007SB_0____
10007_26312100075565555660sagtesagenVVVFIN..._sag-te9997CJ_0____
10008_26312100085566155664ihmerPRPPER..._ihm10007DA_0____
10009_26312100095566555669keinkeinePRPIAT..._kein10010NK_0____
\n", + "

9007 rows × 21 columns

\n", + "
" + ], + "text/plain": [ + " SectionId ParagraphId SentenceId TokenId Begin End \\\n", + "1003 _ 8 24 1003 5826 5836 \n", + "1004 _ 8 24 1004 5837 5844 \n", + "1005 _ 8 24 1005 5844 5845 \n", + "1006 _ 8 24 1006 5846 5848 \n", + "1007 _ 8 24 1007 5849 5853 \n", + "1008 _ 8 24 1008 5854 5856 \n", + "1009 _ 8 24 1009 5857 5861 \n", + "1010 _ 8 24 1010 5862 5866 \n", + "1011 _ 8 24 1011 5866 5867 \n", + "1012 _ 8 24 1012 5868 5872 \n", + "1013 _ 8 24 1013 5873 5875 \n", + "1014 _ 8 24 1014 5876 5881 \n", + "1015 _ 8 24 1015 5881 5882 \n", + "1016 _ 8 24 1016 5883 5886 \n", + "1017 _ 8 24 1017 5887 5890 \n", + "1018 _ 8 24 1018 5891 5904 \n", + "1019 _ 8 24 1019 5905 5915 \n", + "1020 _ 8 24 1020 5916 5919 \n", + "1021 _ 8 24 1021 5920 5929 \n", + "1022 _ 8 24 1022 5929 5930 \n", + "1023 _ 8 24 1023 5931 5934 \n", + "1024 _ 8 24 1024 5935 5941 \n", + "1025 _ 8 24 1025 5942 5947 \n", + "1026 _ 8 24 1026 5948 5958 \n", + "1027 _ 8 24 1027 5959 5962 \n", + "1028 _ 8 24 1028 5962 5963 \n", + "1029 _ 8 24 1029 5964 5979 \n", + "1030 _ 8 24 1030 5980 5989 \n", + "1031 _ 8 24 1031 5989 5990 \n", + "1032 _ 8 24 1032 5991 5993 \n", + "... ... ... ... ... ... ... \n", + "9980 _ 26 311 9980 55491 55495 \n", + "9981 _ 26 311 9981 55496 55503 \n", + "9982 _ 26 311 9982 55504 55510 \n", + "9983 _ 26 311 9983 55510 55511 \n", + "9984 _ 26 311 9984 55512 55517 \n", + "9985 _ 26 311 9985 55518 55525 \n", + "9986 _ 26 311 9986 55526 55529 \n", + "9987 _ 26 311 9987 55530 55535 \n", + "9988 _ 26 311 9988 55536 55542 \n", + "9989 _ 26 311 9989 55542 55543 \n", + "9990 _ 26 311 9990 55544 55547 \n", + "9991 _ 26 311 9991 55548 55559 \n", + "9992 _ 26 311 9992 55560 55564 \n", + "9993 _ 26 311 9993 55565 55568 \n", + "9994 _ 26 311 9994 55569 55582 \n", + "9995 _ 26 311 9995 55582 55583 \n", + "9996 _ 26 312 9996 55584 55587 \n", + "9997 _ 26 312 9997 55588 55594 \n", + "9998 _ 26 312 9998 55595 55605 \n", + "9999 _ 26 312 9999 55606 55610 \n", + "10000 _ 26 312 10000 55611 55617 \n", + "10001 _ 26 312 10001 55618 55623 \n", + "10002 _ 26 312 10002 55624 55629 \n", + "10003 _ 26 312 10003 55630 55639 \n", + "10004 _ 26 312 10004 55640 55649 \n", + "10005 _ 26 312 10005 55649 55650 \n", + "10006 _ 26 312 10006 55651 55654 \n", + "10007 _ 26 312 10007 55655 55660 \n", + "10008 _ 26 312 10008 55661 55664 \n", + "10009 _ 26 312 10009 55665 55669 \n", + "\n", + " Token Lemma CPOS POS ... \\\n", + "1003 Thatsachen Thatsachen NN NN ... \n", + "1004 annimmt annehmen V VVFIN ... \n", + "1005 , , PUNC $, ... \n", + "1006 so so ADV ADV ... \n", + "1007 wird werden V VAFIN ... \n", + "1008 er er PR PPER ... \n", + "1009 sich sich PR PRF ... \n", + "1010 doch doch ADV ADV ... \n", + "1011 , , PUNC $, ... \n", + "1012 eben eben ADV ADV ... \n", + "1013 so so ADV ADV ... \n", + "1014 wenig wenige PR PIS ... \n", + "1015 , , PUNC $, ... \n", + "1016 auf auf PP APPR ... \n", + "1017 die die ART ART ... \n", + "1018 verdrießliche verdrießlich ADJ ADJA ... \n", + "1019 Betheurung Betheurung NN NN ... \n", + "1020 der die ART ART ... \n", + "1021 Wärterinn Wärterinn NN NN ... \n", + "1022 , , PUNC $, ... \n", + "1023 daß dass CONJ KOUS ... \n", + "1024 dieses dies PR PDAT ... \n", + "1025 alles alle PR PIAT ... \n", + "1026 Fieberwahn Fieberwahn NN NN ... \n", + "1027 sey sey NP NE ... \n", + "1028 , , PUNC $, ... \n", + "1029 schlechterdings schlechterdings ADV ADV ... \n", + "1030 verlassen verlassen V VVPP ... \n", + "1031 ; ; PUNC $. ... \n", + "1032 er er PR PPER ... \n", + "... ... ... ... ... ... \n", + "9980 sein sein PR PPOSAT ... \n", + "9981 Gesicht Gesicht NN NN ... \n", + "9982 flammt flammen V VVFIN ... \n", + "9983 , , PUNC $, ... \n", + "9984 Luise Luise NP NE ... \n", + "9985 schlägt schlagen V VVFIN ... \n", + "9986 die die ART ART ... \n", + "9987 Augen Auge NN NN ... \n", + "9988 nieder nieder V PTKVZ ... \n", + "9989 , , PUNC $, ... \n", + "9990 und und CONJ KON ... \n", + "9991 beschuldigt beschuldigen V VVFIN ... \n", + "9992 sich sich PR PRF ... \n", + "9993 der die ART ART ... \n", + "9994 Undankbarkeit Undankbarkeit NN NN ... \n", + "9995 . . PUNC $. ... \n", + "9996 Sie Sie|sie PR PPER ... \n", + "9997 wollte wollen V VMFIN ... \n", + "9998 wenigstens wenigstens ADV ADV ... \n", + "9999 ihre ihr PR PPOSAT ... \n", + "10000 Schuld Schuld NN NN ... \n", + "10001 durch durch PP APPR ... \n", + "10002 keine keine PR PIAT ... \n", + "10003 Heucheley Heucheley NN NN ... \n", + "10004 vermehren vermehren V VVINF ... \n", + "10005 ; ; PUNC $. ... \n", + "10006 sie sie PR PPER ... \n", + "10007 sagte sagen V VVFIN ... \n", + "10008 ihm er PR PPER ... \n", + "10009 kein keine PR PIAT ... \n", + "\n", + " Morphology Hyphenation DependencyHead DependencyRelation \\\n", + "1003 _ That-sa-chen 1002 NK \n", + "1004 _ an-nimmt 991 CJ \n", + "1005 _ , 1004 -- \n", + "1006 _ so 1007 MO \n", + "1007 _ wird -1 -- \n", + "1008 _ er 1007 SB \n", + "1009 _ sich 1030 OA \n", + "1010 _ doch 1030 MO \n", + "1011 _ , 1010 -- \n", + "1012 _ eben 1030 MO \n", + "1013 _ so 1014 MO \n", + "1014 _ we-nig 1030 MO \n", + "1015 _ , 1014 -- \n", + "1016 _ auf 1030 MO \n", + "1017 _ die 1019 NK \n", + "1018 _ ver-drieß-li-che 1019 NK \n", + "1019 _ Be-theu-rung 1016 NK \n", + "1020 _ der 1021 NK \n", + "1021 _ Wär-te-rinn 1019 AG \n", + "1022 _ , 1021 -- \n", + "1023 _ daß 1027 CP \n", + "1024 _ die-ses 1027 NK \n", + "1025 _ al-les 1026 NK \n", + "1026 _ Fie-ber-wahn 1027 NK \n", + "1027 _ sey 1013 CC \n", + "1028 _ , 1027 -- \n", + "1029 _ schlech-ter-dings 1030 MO \n", + "1030 _ ver-las-sen 1007 OC \n", + "1031 _ ; 1030 -- \n", + "1032 _ er 1033 SB \n", + "... ... ... ... ... \n", + "9980 _ sein 9981 NK \n", + "9981 _ Ge-sicht 9982 OA \n", + "9982 _ flammt 9976 MO \n", + "9983 _ , 9982 -- \n", + "9984 _ Lui-se 9985 OA \n", + "9985 _ schlägt 9976 CJ \n", + "9986 _ die 9987 NK \n", + "9987 _ Au-gen 9985 OA \n", + "9988 _ nie-der 9985 SVP \n", + "9989 _ , 9988 -- \n", + "9990 _ und 9985 CD \n", + "9991 _ be-schul-digt 9990 CJ \n", + "9992 _ sich 9991 OA \n", + "9993 _ der 9994 NK \n", + "9994 _ Un-dank-bar-keit 9991 OG \n", + "9995 _ . 9994 -- \n", + "9996 _ Sie 9997 SB \n", + "9997 _ woll-te -1 -- \n", + "9998 _ we-nigs-tens 9997 MO \n", + "9999 _ ih-re 10000 NK \n", + "10000 _ Schuld 10004 OA \n", + "10001 _ durch 10004 MO \n", + "10002 _ kei-ne 10003 NK \n", + "10003 _ Heu-che-ley 10001 NK \n", + "10004 _ ver-meh-ren 9997 OC \n", + "10005 _ ; 10004 -- \n", + "10006 _ sie 10007 SB \n", + "10007 _ sag-te 9997 CJ \n", + "10008 _ ihm 10007 DA \n", + "10009 _ kein 10010 NK \n", + "\n", + " NamedEntity QuoteMarker CoreferenceChainIds SyntaxTree Predicate \\\n", + "1003 _ 0 _ _ _ \n", + "1004 _ 0 _ _ _ \n", + "1005 _ 0 _ _ _ \n", + "1006 _ 0 _ _ _ \n", + "1007 _ 0 _ _ _ \n", + "1008 _ 0 _ _ _ \n", + "1009 _ 0 _ _ _ \n", + "1010 _ 0 _ _ _ \n", + "1011 _ 0 _ _ _ \n", + "1012 _ 0 _ _ _ \n", + "1013 _ 0 _ _ _ \n", + "1014 _ 0 _ _ _ \n", + "1015 _ 0 _ _ _ \n", + "1016 _ 0 _ _ _ \n", + "1017 _ 0 _ _ _ \n", + "1018 _ 0 _ _ _ \n", + "1019 _ 0 _ _ _ \n", + "1020 _ 0 _ _ _ \n", + "1021 _ 0 _ _ _ \n", + "1022 _ 0 _ _ _ \n", + "1023 _ 0 _ _ _ \n", + "1024 _ 0 _ _ _ \n", + "1025 _ 0 _ _ _ \n", + "1026 _ 0 _ _ _ \n", + "1027 _ 0 _ _ _ \n", + "1028 _ 0 _ _ _ \n", + "1029 _ 0 _ _ _ \n", + "1030 _ 0 _ _ _ \n", + "1031 _ 0 _ _ _ \n", + "1032 _ 0 _ _ _ \n", + "... ... ... ... ... ... \n", + "9980 _ 0 _ _ _ \n", + "9981 _ 0 _ _ _ \n", + "9982 _ 0 _ _ _ \n", + "9983 _ 0 _ _ _ \n", + "9984 B-PER 0 _ _ _ \n", + "9985 _ 0 _ _ _ \n", + "9986 _ 0 _ _ _ \n", + "9987 _ 0 _ _ _ \n", + "9988 _ 0 _ _ _ \n", + "9989 _ 0 _ _ _ \n", + "9990 _ 0 _ _ _ \n", + "9991 _ 0 _ _ _ \n", + "9992 _ 0 _ _ _ \n", + "9993 _ 0 _ _ _ \n", + "9994 _ 0 _ _ _ \n", + "9995 _ 0 _ _ _ \n", + "9996 _ 0 _ _ _ \n", + "9997 _ 0 _ _ _ \n", + "9998 _ 0 _ _ _ \n", + "9999 _ 0 _ _ _ \n", + "10000 _ 0 _ _ _ \n", + "10001 _ 0 _ _ _ \n", + "10002 _ 0 _ _ _ \n", + "10003 _ 0 _ _ _ \n", + "10004 _ 0 _ _ _ \n", + "10005 _ 0 _ _ _ \n", + "10006 _ 0 _ _ _ \n", + "10007 _ 0 _ _ _ \n", + "10008 _ 0 _ _ _ \n", + "10009 _ 0 _ _ _ \n", + "\n", + " SemanticArgumentIndex \n", + "1003 _ \n", + "1004 _ \n", + "1005 _ \n", + "1006 _ \n", + "1007 _ \n", + "1008 _ \n", + "1009 _ \n", + "1010 _ \n", + "1011 _ \n", + "1012 _ \n", + "1013 _ \n", + "1014 _ \n", + "1015 _ \n", + "1016 _ \n", + "1017 _ \n", + "1018 _ \n", + "1019 _ \n", + "1020 _ \n", + "1021 _ \n", + "1022 _ \n", + "1023 _ \n", + "1024 _ \n", + "1025 _ \n", + "1026 _ \n", + "1027 _ \n", + "1028 _ \n", + "1029 _ \n", + "1030 _ \n", + "1031 _ \n", + "1032 _ \n", + "... ... \n", + "9980 _ \n", + "9981 _ \n", + "9982 _ \n", + "9983 _ \n", + "9984 _ \n", + "9985 _ \n", + "9986 _ \n", + "9987 _ \n", + "9988 _ \n", + "9989 _ \n", + "9990 _ \n", + "9991 _ \n", + "9992 _ \n", + "9993 _ \n", + "9994 _ \n", + "9995 _ \n", + "9996 _ \n", + "9997 _ \n", + "9998 _ \n", + "9999 _ \n", + "10000 _ \n", + "10001 _ \n", + "10002 _ \n", + "10003 _ \n", + "10004 _ \n", + "10005 _ \n", + "10006 _ \n", + "10007 _ \n", + "10008 _ \n", + "10009 _ \n", + "\n", + "[9007 rows x 21 columns]" + ] + }, + "execution_count": 3, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "import pandas as pd\n", + "import csv\n", + "dof_example = pd.read_table('../../refcor/dof/German/Huber,-Therese_Luise.txt.csv', quoting=csv.QUOTE_NONE, sep='\\t')\n", + "dof_example.iloc[1003:10010]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "So, here is a simple lemma-based feature generator:" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": { + "collapsed": false + }, + "outputs": [], + "source": [ + "class DOFLemmaFG(delta.FeatureGenerator):\n", + " \n", + " def process_file(self, filename):\n", + " dof = pd.read_table(filename, sep='\\t', quoting=csv.QUOTE_NONE)\n", + " tokens = dof[dof.CPOS != 'PUNC'] # only non-punctuation tokens\n", + " counts = tokens.Lemma.value_counts() # count the different values -> Series\n", + " counts.name = self.get_name(filename)\n", + " return counts" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Now, we can pass this in to Corpus:" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
liebglaubenMutterobbeide
Arnim,-Ludwig-Achim-von_Armut%20Reichtum%20Schuld%20und%20Buße%20der%20Gräfin%20Dolores.txt152.0124.095.0107.0188.0
Arnim,-Ludwig-Achim-von_Isabella%20von%20Ägypten.txt48.031.024.042.049.0
Arnim,-Ludwig-Achim-von_Kronenwächter%201.txt96.090.0188.0138.0161.0
Dohm,-Hedwig_Christa%20Ruland.txt68.056.067.070.020.0
Dohm,-Hedwig_Schicksale%20einer%20Seele.txt85.097.0164.082.017.0
\n", + "
" + ], + "text/plain": [ + " lieb glauben Mutter \\\n", + "Arnim,-Ludwig-Achim-von_Armut%20Reichtum%20Schu... 152.0 124.0 95.0 \n", + "Arnim,-Ludwig-Achim-von_Isabella%20von%20Ägypte... 48.0 31.0 24.0 \n", + "Arnim,-Ludwig-Achim-von_Kronenwächter%201.txt 96.0 90.0 188.0 \n", + "Dohm,-Hedwig_Christa%20Ruland.txt 68.0 56.0 67.0 \n", + "Dohm,-Hedwig_Schicksale%20einer%20Seele.txt 85.0 97.0 164.0 \n", + "\n", + " ob beide \n", + "Arnim,-Ludwig-Achim-von_Armut%20Reichtum%20Schu... 107.0 188.0 \n", + "Arnim,-Ludwig-Achim-von_Isabella%20von%20Ägypte... 42.0 49.0 \n", + "Arnim,-Ludwig-Achim-von_Kronenwächter%201.txt 138.0 161.0 \n", + "Dohm,-Hedwig_Christa%20Ruland.txt 70.0 20.0 \n", + "Dohm,-Hedwig_Schicksale%20einer%20Seele.txt 82.0 17.0 " + ] + }, + "execution_count": 5, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "lemma_corpus = delta.Corpus('../../refcor/dof/German', feature_generator=DOFLemmaFG(glob='*.txt.csv'))\n", + "lemma_corpus.iloc[0:5,125:130]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Rest of the steps can be performed as before:" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAAEGCAYAAABGnrPVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "c2500 = lemma_corpus.get_mfw_table(2500)\n", + "distances = delta.functions.cosine_delta(c2500)\n", + "clustering = delta.Clustering(distances)\n", + "delta.Dendrogram(clustering).show()" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "data": { + "text/plain": [ + "F-Ratio 0.347514\n", + "Fisher's LD 2.159179\n", + "Simple Score 3.828408\n", + "Cluster Errors 1.000000\n", + "Adjusted Rand Index 0.965961\n", + "Homogeneity 0.990683\n", + "Completeness 0.992079\n", + "V Measure 0.991380\n", + "Purity 0.986667\n", + "Entropy 0.009317\n", + "dtype: float64" + ] + }, + "execution_count": 7, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "pd.concat((distances.evaluate(), clustering.fclustering().evaluate()))" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "pydelta-next", + "language": "python", + "name": "pydelta-next" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.5.3rc1" + } + }, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/docs/GettingStarted.ipynb b/docs/GettingStarted.ipynb index 675267b..aeeedd8 100644 --- a/docs/GettingStarted.ipynb +++ b/docs/GettingStarted.ipynb @@ -1,11 +1,14 @@ { "cells": [ { - "cell_type": "markdown", + "cell_type": "code", + "execution_count": null, "metadata": { + "collapsed": true, "deletable": true, "editable": true }, + "outputs": [], "source": [ "## Getting Started with pyDelta's _next_ branch\n", "\n", diff --git a/docs/index.rst b/docs/index.rst index 83c0dcf..50aa49a 100644 --- a/docs/index.rst +++ b/docs/index.rst @@ -16,6 +16,7 @@ Contents: README GettingStarted.ipynb + CustomizingPipeline.ipynb concept delta