def generatePrecisionRecallHistogramPlots(averageScores, projectRoot, experimentNames, fileName, title, xLabel):
"""
Generate histogram-style plots for the query generation keyword source experiments
"""
sortedMetrics = [
'recall',
'precision'
]
# Generate the header for the metrics
dataContent = "Experiment "
for experiment in experimentNames:
dataContent += '"' + splitCamelCase(experiment).title().replace('&', 'and') + '" '
dataContent += '\n'
# Build the results for this experiment
for metric in sortedMetrics:
dataContent += '"' + splitCamelCase(metric).title().replace('&', 'and') + '" '
for experiment in experimentNames:
dataContent += str(averageScores[experiment][metric]) + ' '
dataContent += "\n"
# Write it out to the input file to gnuplot
open("input.dat", 'w').write(dataContent)
# Get the configuration
configurationFile = projectRoot + '/analysis/configurations/query_precision_recall_histograms'
configuration = open(configurationFile).read()
configuration = configuration % (
projectRoot + '/analysis/output/retrieval/' + fileName + '.png',
title,
xLabel,
)
open('config', 'w').write(configuration)
# Run gnuplot
subprocess.Popen(['gnuplot', 'config']).communicate()
os.remove('config')
os.remove('input.dat')
# The order in which the groups of histograms should appear
sortedExperiments = [
'BM25Ranking',
experiment,
'BM25BaselineRanking'
]
# Sort the metrics too
sortedMetrics = averageScores[experiment].keys()
sortedMetrics.sort()
# Generate the header for the metrics
dataContent = "Experiment "
for metric in sortedMetrics:
dataContent += '"' + splitCamelCase(metric).title().replace('&', 'and') + '" '
dataContent += '\n'
# Build the results for this experiment
for drawnExperiment in sortedExperiments:
dataContent += '"' + experiments[drawnExperiment].replace('&', 'and') + '" '
for metric in sortedMetrics:
dataContent += str(averageScores[drawnExperiment][metric]) + ' '
dataContent += "\n"
# Write it out to the input file to gnuplot
open("input.dat", 'w').write(dataContent)
# Get the configuration
configurationFile = projectRoot + '/analysis/configurations/single_ranking_feature_clustered_histograms'
configuration = open(configurationFile).read()
averageScores = averageEntityScores(data)
# Concatenate the names of metrics for the horizontal
dataContent = "Experiment "
sortedAverageScores = [
('BM25Ranking', 'None'),
('DMOZSmoothed500BM25Ranking', '500 Documents'),
('DMOZSmoothed1000BM25Ranking', '1000 Documents'),
('DMOZSmoothed5000BM25Ranking', '5000 Documents'),
('DMOZSmoothed1000BM25Ranking', '10000 Documents')
]
sortedMetrics = averageScores[sortedAverageScores[0][0]].keys()
sortedMetrics.sort()
for metric in sortedMetrics:
dataContent += '"' + splitCamelCase(metric).title() + '" '
dataContent += '\n'
for experiment in sortedAverageScores:
dataContent += '"' + experiment[1] + '" '
for metric in sortedMetrics:
dataContent += str(averageScores[experiment[0]][metric]) + ' '
dataContent += "\n"
# Write it out to the input file to gnuplot
open("input.dat", 'w').write(dataContent)
# Get the configuration
configurationFile = projectRoot + '/analysis/configurations/smoothing_clustered_histograms'
configuration = open(configurationFile).read()
# Fill out the configuration