From bfd6ba6aad8cdfb89b8ae606e12ea024f0f06a54 Mon Sep 17 00:00:00 2001 From: Chunrui Huang Date: Mon, 19 Jul 2021 16:32:31 -0700 Subject: [PATCH] check in the changes in the notebook, have put the original clustering code in evaluation_pipeline --- tour_model_eval/evaluation_pipeline.py | 104 +++++++---- .../first_second_round_evaluation.ipynb | 167 ++++++++++-------- 2 files changed, 157 insertions(+), 114 deletions(-) diff --git a/tour_model_eval/evaluation_pipeline.py b/tour_model_eval/evaluation_pipeline.py index 6f007a8..08c66ff 100644 --- a/tour_model_eval/evaluation_pipeline.py +++ b/tour_model_eval/evaluation_pipeline.py @@ -1,11 +1,14 @@ import emission.analysis.modelling.tour_model.similarity as similarity import numpy as np +import emission.core.get_database as edb import emission.analysis.modelling.tour_model.get_request_percentage as grp import emission.analysis.modelling.tour_model.get_scores as gs import emission.analysis.modelling.tour_model.label_processing as lp import emission.analysis.modelling.tour_model.data_preprocessing as preprocess -import second_round_of_clustering as sr +import emission.analysis.modelling.tour_model.second_round_of_clustering as sr +import emission.analysis.modelling.tour_model.get_users as gu import pandas as pd +import jsonpickle as jpickle def second_round(bin_trips,filter_trips,first_labels,track,low,dist_pct,sim,kmeans): sec = sr.SecondRoundOfClustering(bin_trips,first_labels) @@ -122,43 +125,68 @@ def test(data,radius,low,dist_pct,kmeans): return homo_first,percentage_first,homo_second,percentage_second,scores -def main(uuid=None): - user = uuid +def main(all_users): radius = 100 - df = pd.DataFrame(columns=['user','user_id','percentage of 1st round','homogeneity socre of 1st round','percentage of 2nd round', - 'homogeneity socre of 2nd roun','scores','lower boundary','distance percentage']) - trips = preprocess.read_data(user) - filter_trips = preprocess.filter_data(trips, radius) - tune_idx, test_idx = preprocess.split_data(filter_trips) - tune_data = preprocess.get_subdata(filter_trips, test_idx) - test_data = preprocess.get_subdata(filter_trips, tune_idx) - - # tune data - for j in range(len(tune_data)): - low, dist_pct = tune(tune_data[j], radius, kmeans=False) - df.loc[j,'lower boundary']=low - df.loc[j,'distance percentage']=dist_pct - - - # testing - for k in range(len(test_data)): - low = df.loc[k,'lower boundary'] - dist_pct = df.loc[k,'distance percentage'] - - # for testing, we add kmeans to re-build the model - homo_first, percentage_first, homo_second, percentage_second, scores = test(test_data[k],radius,low, - dist_pct,kmeans=True) - - df.loc[k, 'percentage of 1st round'] = percentage_first - df.loc[k, 'homogeneity socre of 1st round'] = homo_first - df.loc[k, 'percentage of 2nd round'] = percentage_second - df.loc[k, 'homogeneity socre of 2nd round'] = homo_second - df.loc[k, 'scores'] = scores - df['user_id'] = user - df['user'] = 'user0' - - filename = "user_" + str(user) + ".csv" - df.to_csv(filename, index=True, index_label='split') + # get all/valid user list + user_ls, valid_users = gu.get_user_ls(all_users, radius) + all_filename = [] + for a in range(len(all_users)): + user = all_users[a] + df = pd.DataFrame(columns=['user','user_id','percentage of 1st round','homogeneity socre of 1st round', + 'percentage of 2nd round','homogeneity socre of 2nd roun','scores','lower boundary', + 'distance percentage']) + trips = preprocess.read_data(user) + filter_trips = preprocess.filter_data(trips, radius) + # filter out users that don't have enough valid labeled trips + if not gu.valid_user(filter_trips, trips): + continue + tune_idx, test_idx = preprocess.split_data(filter_trips) + # choose tuning/test set to run the model + # this step will use KFold (5 splits) to split the data into different subsets + # - tune: tuning set + # - test: test set + # Here we user a bigger part of the data for testing and a smaller part for tuning + tune_data = preprocess.get_subdata(filter_trips, test_idx) + test_data = preprocess.get_subdata(filter_trips, tune_idx) + + # tune data + for j in range(len(tune_data)): + # for tuning, we don't add kmeans for re-clustering. We just need to get tuning parameters + # - low: the lower boundary of the dendrogram. If the final distance of the dendrogram is lower than "low", + # this bin no need to be re-clutered. + # - dist_pct: the higher boundary of the dendrogram. If the final distance is higher than "low", + # the cutoff of the dendrogram is (the final distance of the dendrogram * dist_pct) + low, dist_pct = tune(tune_data[j], radius, kmeans=False) + df.loc[j,'lower boundary']=low + df.loc[j,'distance percentage']=dist_pct + + # testing + for k in range(len(test_data)): + low = df.loc[k,'lower boundary'] + dist_pct = df.loc[k,'distance percentage'] + + # for testing, we add kmeans to re-build the model + homo_first, percentage_first, homo_second, percentage_second, scores = test(test_data[k],radius,low, + dist_pct,kmeans=True) + df.loc[k, 'percentage of 1st round'] = percentage_first + df.loc[k, 'homogeneity socre of 1st round'] = homo_first + df.loc[k, 'percentage of 2nd round'] = percentage_second + df.loc[k, 'homogeneity socre of 2nd round'] = homo_second + df.loc[k, 'scores'] = scores + df['user_id'] = user + df['user']='user'+str(a+1) + + filename = "user_" + str(user) + ".csv" + all_filename.append(filename) + df.to_csv(filename, index=True, index_label='split') + + # collect filename in a file, use it to plot the scatter + collect_filename = jpickle.dumps(all_filename) + with open("collect_filename", "w") as fd: + fd.write(collect_filename) + if __name__ == '__main__': - main(uuid=None) + participant_uuid_obj = list(edb.get_profile_db().find({"install_group": "participant"}, {"user_id": 1, "_id": 0})) + all_users = [u["user_id"] for u in participant_uuid_obj] + main(all_users) diff --git a/tour_model_eval/first_second_round_evaluation.ipynb b/tour_model_eval/first_second_round_evaluation.ipynb index b510532..2510b0d 100644 --- a/tour_model_eval/first_second_round_evaluation.ipynb +++ b/tour_model_eval/first_second_round_evaluation.ipynb @@ -1,31 +1,39 @@ { "cells": [ + { + "cell_type": "markdown", + "id": "victorian-speech", + "metadata": {}, + "source": [ + "## This notebook is to show the evaluation (scatter plot) of two rounds of clustering" + ] + }, { "cell_type": "code", - "execution_count": null, + "execution_count": 1, "id": "mighty-ukraine", "metadata": {}, - "outputs": [], + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "storage not configured, falling back to sample, default configuration\n", + "Connecting to database URL localhost\n" + ] + } + ], "source": [ "import emission.core.get_database as edb\n", - "import emission.analysis.modelling.tour_model.similarity as similarity\n", - "import pandas as pd\n", - "import numpy as np\n", - "import emission.analysis.modelling.tour_model.get_request_percentage as grp\n", - "import emission.analysis.modelling.tour_model.get_scores as gs\n", - "import emission.analysis.modelling.tour_model.label_processing as lp\n", "import emission.analysis.modelling.tour_model.get_users as gu\n", - "import emission.analysis.modelling.tour_model.data_preprocessing as preprocess\n", - "import evaluation_pipeline as ep\n", + "import emission.analysis.modelling.tour_model.load_predict as predict\n", "import matplotlib.pyplot as plt\n", - "import get_plot as plot\n", - "import emission.core.common as ecc\n", - "import jsonpickle as jpickle" + "import emission.analysis.modelling.tour_model.get_plot as plot" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 2, "id": "cathedral-pointer", "metadata": {}, "outputs": [], @@ -36,7 +44,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 3, "id": "exotic-livestock", "metadata": {}, "outputs": [], @@ -46,72 +54,79 @@ }, { "cell_type": "code", - "execution_count": null, - "id": "focal-express", + "execution_count": 4, + "id": "elementary-advocacy", "metadata": {}, "outputs": [], "source": [ "# get all/valid user list\n", "user_ls, valid_users = gu.get_user_ls(all_users, radius)\n", - "\n", - "all_filename = []\n", - "for a in range(len(all_users)):\n", - " df = pd.DataFrame(columns=['user','user_id','percentage of 1st round','homogeneity socre of 1st round','percentage of 2nd round',\n", - " 'homogeneity socre of 2nd roun','scores','lower boundary','distance percentage'])\n", - " user = all_users[a]\n", - " \n", - " trips = preprocess.read_data(user)\n", - " filter_trips = preprocess.filter_data(trips, radius)\n", - " print('user', a + 1, 'filter_trips len', len(filter_trips))\n", - "\n", - " # filter out users that don't have enough valid labeled trips\n", - " if not gu.valid_user(filter_trips, trips):\n", - " continue\n", - " tune_idx, test_idx = preprocess.split_data(filter_trips)\n", - "\n", - " # choose tuning/test set to run the model\n", - " # this step will use KFold (5 splits) to split the data into different subsets\n", - " # - tune: tuning set\n", - " # - test: test set\n", - " # Here we user a bigger part of the data for testing and a smaller part for tuning\n", - " tune_data = preprocess.get_subdata(filter_trips, test_idx)\n", - " test_data = preprocess.get_subdata(filter_trips, tune_idx)\n", - " \n", - " # tune data\n", - " for j in range(len(tune_data)):\n", - " # for tuning, we don't add kmeans for re-clustering. We just need to get tuning parameters\n", - " # - low: the lower boundary of the dendrogram. If the final distance of the dendrogram is lower than \"low\", \n", - " # this bin no need to be re-clutered.\n", - " # - dist_pct: the higher boundary of the dendrogram. If the final distance is higher than \"low\", \n", - " # the cutoff of the dendrogram is (the final distance of the dendrogram * dist_pct)\n", - " low,dist_pct = ep.tune(tune_data[j],radius,kmeans=False)\n", - " df.loc[j,'lower boundary']=low\n", - " df.loc[j,'distance percentage']=dist_pct\n", - "\n", - " # testing\n", - " for k in range(len(test_data)):\n", - " low = df.loc[k,'lower boundary']\n", - " dist_pct = df.loc[k,'distance percentage'] \n", - " # for testing, we add kmeans to re-build the model\n", - " homo_first, percentage_first, homo_second, percentage_second, scores = ep.test(test_data[k],radius,low,dist_pct,kmeans=True)\n", - " \n", - " df.loc[k,'percentage of 1st round']=percentage_first\n", - " df.loc[k,'homogeneity socre of 1st round']=homo_first\n", - " df.loc[k,'percentage of 2nd round']=percentage_second\n", - " df.loc[k,'homogeneity socre of 2nd round']=homo_second\n", - " df.loc[k,'scores']=scores\n", - " df['user_id']=user\n", - " df['user']='user'+str(a+1)\n", - " filename = \"user_\"+str(user)+\".csv\"\n", - " all_filename.append(filename)\n", - " df.to_csv(filename,index=True,index_label='split')\n", - " \n", - " \n", - " \n", - "# collect filename in a file, use it to plot the scatter\n", - "collect_filename = jpickle.dumps(all_filename)\n", - "with open(\"collect_filename\", \"w\") as fd:\n", - " fd.write(collect_filename)" + "# get all filenames of clustering result\n", + "collect_filename = predict.loadModelStage(\"collect_filename\")" + ] + }, + { + "cell_type": "markdown", + "id": "delayed-apache", + "metadata": {}, + "source": [ + "### Get scatter plot from the 1st round of clustering" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "id": "breeding-favor", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "plt.figure()\n", + "plot.get_scatter(valid_users, collect_filename, first_round=True, second_round=False)" + ] + }, + { + "cell_type": "markdown", + "id": "acknowledged-blackjack", + "metadata": {}, + "source": [ + "### Get scatter plot from the 2nd round of clustering" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "id": "pretty-software", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "plt.figure()\n", + "plot.get_scatter(valid_users, collect_filename, first_round=False, second_round=True)" ] } ],