From 9dda28d2ff8c67d0efb30333cafa35995ed18521 Mon Sep 17 00:00:00 2001
From: Oleg Zendel <13455948+Zendelo@users.noreply.github.com>
Date: Tue, 27 Feb 2024 22:16:40 +1100
Subject: [PATCH] Analyses with qpptk@tirex in progress
work in progress
---
code/qpp_notebook.ipynb | 2519 +++++++++++++++++++++++++++++++++++++++
1 file changed, 2519 insertions(+)
create mode 100644 code/qpp_notebook.ipynb
diff --git a/code/qpp_notebook.ipynb b/code/qpp_notebook.ipynb
new file mode 100644
index 0000000..463e54c
--- /dev/null
+++ b/code/qpp_notebook.ipynb
@@ -0,0 +1,2519 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:44:19.496959Z",
+ "start_time": "2024-02-27T09:44:19.493171Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "PxFxlda_EpHQ",
+ "outputId": "b5f63c83-af04-4c91-e7ee-b4dc49d1dcdc"
+ },
+ "outputs": [],
+ "source": [
+ "# !pip3 install tira ir_datasets python-terrier"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:44:20.666891Z",
+ "start_time": "2024-02-27T09:44:19.498572Z"
+ },
+ "id": "SX4juBzqChpJ"
+ },
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "PyTerrier 0.10.0 has loaded Terrier 5.7 (built by craigm on 2022-11-10 18:30) and terrier-helper 0.0.7\n",
+ "\n",
+ "No etc/terrier.properties, using terrier.default.properties for bootstrap configuration.\n"
+ ]
+ }
+ ],
+ "source": [
+ "import pandas as pd\n",
+ "import pyterrier as pt\n",
+ "from tira.rest_api_client import Client\n",
+ "from tira.third_party_integrations import ensure_pyterrier_is_loaded\n",
+ "from tqdm import tqdm\n",
+ "\n",
+ "ensure_pyterrier_is_loaded()\n",
+ "tira = Client()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-O5UUtTZHxPG"
+ },
+ "source": [
+ "# First, Example for a single dataset"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:44:21.273458Z",
+ "start_time": "2024-02-27T09:44:21.242597Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 206
+ },
+ "id": "hEdqVc5iGGWK",
+ "outputId": "e967d78c-a3fe-48b2-8092-0172675d5a3b"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " qid | \n",
+ " query | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 | \n",
+ " 3990512 | \n",
+ " how can we get concentration onsomething | \n",
+ "
\n",
+ " \n",
+ " 1 | \n",
+ " 714612 | \n",
+ " why doesn t the water fall off earth if it s r... | \n",
+ "
\n",
+ " \n",
+ " 2 | \n",
+ " 2528767 | \n",
+ " how do i determine the charge of the iron ion ... | \n",
+ "
\n",
+ " \n",
+ " 3 | \n",
+ " 821387 | \n",
+ " i have mice how do i get rid of them humanely | \n",
+ "
\n",
+ " \n",
+ " 4 | \n",
+ " 1880028 | \n",
+ " what does see leaflet mean on ept pregnancy test | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
"
+ ],
+ "text/plain": [
+ " qid query\n",
+ "0 3990512 how can we get concentration onsomething\n",
+ "1 714612 why doesn t the water fall off earth if it s r...\n",
+ "2 2528767 how do i determine the charge of the iron ion ...\n",
+ "3 821387 i have mice how do i get rid of them humanely\n",
+ "4 1880028 what does see leaflet mean on ept pregnancy test"
+ ]
+ },
+ "execution_count": 4,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "dataset = 'antique-test-20230107-training'\n",
+ "pt_dataset = pt.get_dataset(f\"irds:ir-benchmarks/{dataset}\")\n",
+ "\n",
+ "pt_dataset.get_topics('query').head()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:44:26.021398Z",
+ "start_time": "2024-02-27T09:44:21.274610Z"
+ },
+ "id": "MQ6m5dptGi3C"
+ },
+ "outputs": [],
+ "source": [
+ "bm25 = tira.pt.from_submission('ir-benchmarks/tira-ir-starter/BM25 Re-Rank (tira-ir-starter-pyterrier)', dataset)\n",
+ "qpp_predictions = tira.pt.transform_queries('ir-benchmarks/qpptk/all-predictors', dataset)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:44:26.068925Z",
+ "start_time": "2024-02-27T09:44:26.023712Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 377
+ },
+ "id": "I6F83a5NHlNv",
+ "outputId": "7bf7dc2b-ce46-403d-b1c4-35c2e6585c7a"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " qid | \n",
+ " query | \n",
+ " max-idf | \n",
+ " avg-idf | \n",
+ " scq | \n",
+ " max-scq | \n",
+ " avg-scq | \n",
+ " var | \n",
+ " max-var | \n",
+ " avg-var | \n",
+ " ... | \n",
+ " smv+50 | \n",
+ " clarity+50+100 | \n",
+ " wig+100 | \n",
+ " nqc+100 | \n",
+ " smv+100 | \n",
+ " clarity+100+100 | \n",
+ " wig+1000 | \n",
+ " nqc+1000 | \n",
+ " smv+1000 | \n",
+ " clarity+1000+100 | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 | \n",
+ " 3990512 | \n",
+ " how can we get concentration onsomething | \n",
+ " 11.116584 | \n",
+ " 7.046932 | \n",
+ " 154.548142 | \n",
+ " 48.116900 | \n",
+ " 38.637036 | \n",
+ " 11.621307 | \n",
+ " 6.724458 | \n",
+ " 1.936884 | \n",
+ " ... | \n",
+ " 0.030120 | \n",
+ " 3.795747 | \n",
+ " 1.823258 | \n",
+ " 0.039069 | \n",
+ " 0.035748 | \n",
+ " 3.731287 | \n",
+ " 0.972405 | \n",
+ " 0.019728 | \n",
+ " 0.010170 | \n",
+ " 3.389977 | \n",
+ "
\n",
+ " \n",
+ " 1 | \n",
+ " 714612 | \n",
+ " why doesn t the water fall off earth if it s r... | \n",
+ " 12.215196 | \n",
+ " 7.627986 | \n",
+ " 380.326020 | \n",
+ " 47.708207 | \n",
+ " 38.032602 | \n",
+ " 15.536635 | \n",
+ " 5.230969 | \n",
+ " 1.294720 | \n",
+ " ... | \n",
+ " 0.004710 | \n",
+ " 4.267368 | \n",
+ " 1.538839 | \n",
+ " 0.011142 | \n",
+ " 0.008922 | \n",
+ " 4.106367 | \n",
+ " 0.695940 | \n",
+ " 0.009800 | \n",
+ " 0.006195 | \n",
+ " 3.815700 | \n",
+ "
\n",
+ " \n",
+ " 2 | \n",
+ " 2528767 | \n",
+ " how do i determine the charge of the iron ion ... | \n",
+ " 12.908343 | \n",
+ " 8.712145 | \n",
+ " 442.767537 | \n",
+ " 52.221022 | \n",
+ " 40.251594 | \n",
+ " 20.639520 | \n",
+ " 3.725044 | \n",
+ " 1.876320 | \n",
+ " ... | \n",
+ " 0.016734 | \n",
+ " 4.314736 | \n",
+ " 2.091532 | \n",
+ " 0.017572 | \n",
+ " 0.015034 | \n",
+ " 4.305586 | \n",
+ " 0.928886 | \n",
+ " 0.012062 | \n",
+ " 0.007967 | \n",
+ " 4.294343 | \n",
+ "
\n",
+ " \n",
+ " 3 | \n",
+ " 821387 | \n",
+ " i have mice how do i get rid of them humanely | \n",
+ " 12.908343 | \n",
+ " 8.890641 | \n",
+ " 340.263514 | \n",
+ " 50.394187 | \n",
+ " 37.807057 | \n",
+ " 11.381356 | \n",
+ " 2.897810 | \n",
+ " 1.138136 | \n",
+ " ... | \n",
+ " 0.021428 | \n",
+ " 3.532815 | \n",
+ " 2.210218 | \n",
+ " 0.021395 | \n",
+ " 0.017602 | \n",
+ " 3.516496 | \n",
+ " 0.805009 | \n",
+ " 0.015333 | \n",
+ " 0.010520 | \n",
+ " 3.508216 | \n",
+ "
\n",
+ " \n",
+ " 4 | \n",
+ " 1880028 | \n",
+ " what does see leaflet mean on ept pregnancy test | \n",
+ " 12.908343 | \n",
+ " 6.848213 | \n",
+ " 339.046657 | \n",
+ " 49.747516 | \n",
+ " 37.671851 | \n",
+ " 11.333447 | \n",
+ " 3.290287 | \n",
+ " 1.259272 | \n",
+ " ... | \n",
+ " 0.011415 | \n",
+ " 4.770863 | \n",
+ " 1.538375 | \n",
+ " 0.016755 | \n",
+ " 0.011784 | \n",
+ " 4.750485 | \n",
+ " 0.881017 | \n",
+ " 0.009732 | \n",
+ " 0.005666 | \n",
+ " 4.640350 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
5 rows × 34 columns
\n",
+ "
"
+ ],
+ "text/plain": [
+ " qid query max-idf \\\n",
+ "0 3990512 how can we get concentration onsomething 11.116584 \n",
+ "1 714612 why doesn t the water fall off earth if it s r... 12.215196 \n",
+ "2 2528767 how do i determine the charge of the iron ion ... 12.908343 \n",
+ "3 821387 i have mice how do i get rid of them humanely 12.908343 \n",
+ "4 1880028 what does see leaflet mean on ept pregnancy test 12.908343 \n",
+ "\n",
+ " avg-idf scq max-scq avg-scq var max-var avg-var \\\n",
+ "0 7.046932 154.548142 48.116900 38.637036 11.621307 6.724458 1.936884 \n",
+ "1 7.627986 380.326020 47.708207 38.032602 15.536635 5.230969 1.294720 \n",
+ "2 8.712145 442.767537 52.221022 40.251594 20.639520 3.725044 1.876320 \n",
+ "3 8.890641 340.263514 50.394187 37.807057 11.381356 2.897810 1.138136 \n",
+ "4 6.848213 339.046657 49.747516 37.671851 11.333447 3.290287 1.259272 \n",
+ "\n",
+ " ... smv+50 clarity+50+100 wig+100 nqc+100 smv+100 \\\n",
+ "0 ... 0.030120 3.795747 1.823258 0.039069 0.035748 \n",
+ "1 ... 0.004710 4.267368 1.538839 0.011142 0.008922 \n",
+ "2 ... 0.016734 4.314736 2.091532 0.017572 0.015034 \n",
+ "3 ... 0.021428 3.532815 2.210218 0.021395 0.017602 \n",
+ "4 ... 0.011415 4.770863 1.538375 0.016755 0.011784 \n",
+ "\n",
+ " clarity+100+100 wig+1000 nqc+1000 smv+1000 clarity+1000+100 \n",
+ "0 3.731287 0.972405 0.019728 0.010170 3.389977 \n",
+ "1 4.106367 0.695940 0.009800 0.006195 3.815700 \n",
+ "2 4.305586 0.928886 0.012062 0.007967 4.294343 \n",
+ "3 3.516496 0.805009 0.015333 0.010520 3.508216 \n",
+ "4 4.750485 0.881017 0.009732 0.005666 4.640350 \n",
+ "\n",
+ "[5 rows x 34 columns]"
+ ]
+ },
+ "execution_count": 6,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "qpp_predictions(pt_dataset.get_topics('query').head())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:44:26.628811Z",
+ "start_time": "2024-02-27T09:44:26.070498Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 424
+ },
+ "id": "MSlPf_B5H1vT",
+ "outputId": "5b91b276-3690-4149-83d7-3e4f74e76a06"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " name | \n",
+ " qid | \n",
+ " measure | \n",
+ " value | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 70 | \n",
+ " BM25 | \n",
+ " 100653 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.428160 | \n",
+ "
\n",
+ " \n",
+ " 85 | \n",
+ " BM25 | \n",
+ " 1015624 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.580570 | \n",
+ "
\n",
+ " \n",
+ " 43 | \n",
+ " BM25 | \n",
+ " 1017690 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.754357 | \n",
+ "
\n",
+ " \n",
+ " 116 | \n",
+ " BM25 | \n",
+ " 1035857 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.652168 | \n",
+ "
\n",
+ " \n",
+ " 122 | \n",
+ " BM25 | \n",
+ " 103830 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.530858 | \n",
+ "
\n",
+ " \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " 19 | \n",
+ " BM25 | \n",
+ " 851124 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.545156 | \n",
+ "
\n",
+ " \n",
+ " 95 | \n",
+ " BM25 | \n",
+ " 896725 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.360260 | \n",
+ "
\n",
+ " \n",
+ " 104 | \n",
+ " BM25 | \n",
+ " 922849 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.498273 | \n",
+ "
\n",
+ " \n",
+ " 178 | \n",
+ " BM25 | \n",
+ " 949154 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.664721 | \n",
+ "
\n",
+ " \n",
+ " 142 | \n",
+ " BM25 | \n",
+ " 953489 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.195793 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
200 rows × 4 columns
\n",
+ "
"
+ ],
+ "text/plain": [
+ " name qid measure value\n",
+ "70 BM25 100653 ndcg_cut.10 0.428160\n",
+ "85 BM25 1015624 ndcg_cut.10 0.580570\n",
+ "43 BM25 1017690 ndcg_cut.10 0.754357\n",
+ "116 BM25 1035857 ndcg_cut.10 0.652168\n",
+ "122 BM25 103830 ndcg_cut.10 0.530858\n",
+ ".. ... ... ... ...\n",
+ "19 BM25 851124 ndcg_cut.10 0.545156\n",
+ "95 BM25 896725 ndcg_cut.10 0.360260\n",
+ "104 BM25 922849 ndcg_cut.10 0.498273\n",
+ "178 BM25 949154 ndcg_cut.10 0.664721\n",
+ "142 BM25 953489 ndcg_cut.10 0.195793\n",
+ "\n",
+ "[200 rows x 4 columns]"
+ ]
+ },
+ "execution_count": 7,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "df_eval = pt.Experiment([bm25], pt_dataset.get_topics('query'), pt_dataset.get_qrels(), names=['BM25'],\n",
+ " eval_metrics=['ndcg_cut.10'], perquery=True)\n",
+ "df_eval"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:44:26.680325Z",
+ "start_time": "2024-02-27T09:44:26.630771Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 461
+ },
+ "id": "5iDu8XkXJJhR",
+ "outputId": "1288cef2-3f77-444c-943f-8256cb144968"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " name | \n",
+ " qid | \n",
+ " measure | \n",
+ " value | \n",
+ " max-idf | \n",
+ " avg-idf | \n",
+ " scq | \n",
+ " max-scq | \n",
+ " avg-scq | \n",
+ " var | \n",
+ " ... | \n",
+ " smv+50 | \n",
+ " clarity+50+100 | \n",
+ " wig+100 | \n",
+ " nqc+100 | \n",
+ " smv+100 | \n",
+ " clarity+100+100 | \n",
+ " wig+1000 | \n",
+ " nqc+1000 | \n",
+ " smv+1000 | \n",
+ " clarity+1000+100 | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 70 | \n",
+ " BM25 | \n",
+ " 100653 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.428160 | \n",
+ " 12.908343 | \n",
+ " 9.051803 | \n",
+ " 598.373506 | \n",
+ " 50.637838 | \n",
+ " 37.398344 | \n",
+ " 23.617904 | \n",
+ " ... | \n",
+ " 0.014273 | \n",
+ " 3.544082 | \n",
+ " 1.951133 | \n",
+ " 0.016508 | \n",
+ " 0.009057 | \n",
+ " 3.544058 | \n",
+ " 0.745923 | \n",
+ " 0.009943 | \n",
+ " 0.006240 | \n",
+ " 3.544040 | \n",
+ "
\n",
+ " \n",
+ " 85 | \n",
+ " BM25 | \n",
+ " 1015624 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.580570 | \n",
+ " 12.908343 | \n",
+ " 8.369403 | \n",
+ " 280.338094 | \n",
+ " 50.578374 | \n",
+ " 40.048299 | \n",
+ " 9.523332 | \n",
+ " ... | \n",
+ " 0.011114 | \n",
+ " 3.999061 | \n",
+ " 1.929926 | \n",
+ " 0.018038 | \n",
+ " 0.013737 | \n",
+ " 3.935562 | \n",
+ " 1.029481 | \n",
+ " 0.012135 | \n",
+ " 0.007093 | \n",
+ " 3.716341 | \n",
+ "
\n",
+ " \n",
+ " 43 | \n",
+ " BM25 | \n",
+ " 1017690 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.754357 | \n",
+ " 12.215196 | \n",
+ " 7.912693 | \n",
+ " 283.592656 | \n",
+ " 49.248037 | \n",
+ " 40.513237 | \n",
+ " 8.418207 | \n",
+ " ... | \n",
+ " 0.013975 | \n",
+ " 3.511534 | \n",
+ " 1.571078 | \n",
+ " 0.018584 | \n",
+ " 0.015681 | \n",
+ " 3.456468 | \n",
+ " 0.938830 | \n",
+ " 0.009677 | \n",
+ " 0.004896 | \n",
+ " 3.187760 | \n",
+ "
\n",
+ " \n",
+ " 116 | \n",
+ " BM25 | \n",
+ " 1035857 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.652168 | \n",
+ " 10.135754 | \n",
+ " 7.745829 | \n",
+ " 367.762077 | \n",
+ " 50.625796 | \n",
+ " 45.970260 | \n",
+ " 15.929586 | \n",
+ " ... | \n",
+ " 0.004655 | \n",
+ " 3.839171 | \n",
+ " 1.919674 | \n",
+ " 0.006738 | \n",
+ " 0.004749 | \n",
+ " 3.618240 | \n",
+ " 1.054783 | \n",
+ " 0.012100 | \n",
+ " 0.009408 | \n",
+ " 3.312848 | \n",
+ "
\n",
+ " \n",
+ " 122 | \n",
+ " BM25 | \n",
+ " 103830 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.530858 | \n",
+ " 12.908343 | \n",
+ " 7.721879 | \n",
+ " 277.966310 | \n",
+ " 50.531838 | \n",
+ " 39.709473 | \n",
+ " 11.214659 | \n",
+ " ... | \n",
+ " 0.005703 | \n",
+ " 3.557757 | \n",
+ " 1.829664 | \n",
+ " 0.015672 | \n",
+ " 0.013502 | \n",
+ " 3.459321 | \n",
+ " 0.914128 | \n",
+ " 0.012781 | \n",
+ " 0.007959 | \n",
+ " 3.326919 | \n",
+ "
\n",
+ " \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " 19 | \n",
+ " BM25 | \n",
+ " 851124 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.545156 | \n",
+ " 12.908343 | \n",
+ " 9.099939 | \n",
+ " 306.580831 | \n",
+ " 49.457465 | \n",
+ " 34.064537 | \n",
+ " 8.378009 | \n",
+ " ... | \n",
+ " 0.007420 | \n",
+ " 4.803370 | \n",
+ " 1.619716 | \n",
+ " 0.010545 | \n",
+ " 0.007952 | \n",
+ " 4.641726 | \n",
+ " 0.926463 | \n",
+ " 0.007980 | \n",
+ " 0.004847 | \n",
+ " 4.121996 | \n",
+ "
\n",
+ " \n",
+ " 95 | \n",
+ " BM25 | \n",
+ " 896725 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.360260 | \n",
+ " 10.962433 | \n",
+ " 7.438826 | \n",
+ " 467.079609 | \n",
+ " 48.917291 | \n",
+ " 42.461783 | \n",
+ " 12.903404 | \n",
+ " ... | \n",
+ " 0.004789 | \n",
+ " 3.113233 | \n",
+ " 1.477184 | \n",
+ " 0.008821 | \n",
+ " 0.007782 | \n",
+ " 3.153661 | \n",
+ " 0.824822 | \n",
+ " 0.007862 | \n",
+ " 0.005510 | \n",
+ " 2.968856 | \n",
+ "
\n",
+ " \n",
+ " 104 | \n",
+ " BM25 | \n",
+ " 922849 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.498273 | \n",
+ " 12.908343 | \n",
+ " 7.112997 | \n",
+ " 274.841073 | \n",
+ " 46.525209 | \n",
+ " 34.355134 | \n",
+ " 6.537749 | \n",
+ " ... | \n",
+ " 0.005467 | \n",
+ " 3.781467 | \n",
+ " 1.575306 | \n",
+ " 0.021174 | \n",
+ " 0.019889 | \n",
+ " 3.714051 | \n",
+ " 0.716917 | \n",
+ " 0.012303 | \n",
+ " 0.006179 | \n",
+ " 3.498253 | \n",
+ "
\n",
+ " \n",
+ " 178 | \n",
+ " BM25 | \n",
+ " 949154 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.664721 | \n",
+ " 12.908343 | \n",
+ " 8.687071 | \n",
+ " 248.388020 | \n",
+ " 48.224169 | \n",
+ " 31.048502 | \n",
+ " 12.579528 | \n",
+ " ... | \n",
+ " 0.017332 | \n",
+ " 3.884780 | \n",
+ " 1.356908 | \n",
+ " 0.018738 | \n",
+ " 0.015678 | \n",
+ " 3.836012 | \n",
+ " 0.721010 | \n",
+ " 0.011075 | \n",
+ " 0.006777 | \n",
+ " 3.563908 | \n",
+ "
\n",
+ " \n",
+ " 142 | \n",
+ " BM25 | \n",
+ " 953489 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.195793 | \n",
+ " 12.908343 | \n",
+ " 8.647971 | \n",
+ " 529.897082 | \n",
+ " 47.823053 | \n",
+ " 35.326472 | \n",
+ " 20.710882 | \n",
+ " ... | \n",
+ " 0.012084 | \n",
+ " 3.654480 | \n",
+ " 1.778057 | \n",
+ " 0.012714 | \n",
+ " 0.010522 | \n",
+ " 3.645432 | \n",
+ " 0.669576 | \n",
+ " 0.009267 | \n",
+ " 0.005655 | \n",
+ " 3.632626 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
200 rows × 36 columns
\n",
+ "
"
+ ],
+ "text/plain": [
+ " name qid measure value max-idf avg-idf scq \\\n",
+ "70 BM25 100653 ndcg_cut.10 0.428160 12.908343 9.051803 598.373506 \n",
+ "85 BM25 1015624 ndcg_cut.10 0.580570 12.908343 8.369403 280.338094 \n",
+ "43 BM25 1017690 ndcg_cut.10 0.754357 12.215196 7.912693 283.592656 \n",
+ "116 BM25 1035857 ndcg_cut.10 0.652168 10.135754 7.745829 367.762077 \n",
+ "122 BM25 103830 ndcg_cut.10 0.530858 12.908343 7.721879 277.966310 \n",
+ ".. ... ... ... ... ... ... ... \n",
+ "19 BM25 851124 ndcg_cut.10 0.545156 12.908343 9.099939 306.580831 \n",
+ "95 BM25 896725 ndcg_cut.10 0.360260 10.962433 7.438826 467.079609 \n",
+ "104 BM25 922849 ndcg_cut.10 0.498273 12.908343 7.112997 274.841073 \n",
+ "178 BM25 949154 ndcg_cut.10 0.664721 12.908343 8.687071 248.388020 \n",
+ "142 BM25 953489 ndcg_cut.10 0.195793 12.908343 8.647971 529.897082 \n",
+ "\n",
+ " max-scq avg-scq var ... smv+50 clarity+50+100 wig+100 \\\n",
+ "70 50.637838 37.398344 23.617904 ... 0.014273 3.544082 1.951133 \n",
+ "85 50.578374 40.048299 9.523332 ... 0.011114 3.999061 1.929926 \n",
+ "43 49.248037 40.513237 8.418207 ... 0.013975 3.511534 1.571078 \n",
+ "116 50.625796 45.970260 15.929586 ... 0.004655 3.839171 1.919674 \n",
+ "122 50.531838 39.709473 11.214659 ... 0.005703 3.557757 1.829664 \n",
+ ".. ... ... ... ... ... ... ... \n",
+ "19 49.457465 34.064537 8.378009 ... 0.007420 4.803370 1.619716 \n",
+ "95 48.917291 42.461783 12.903404 ... 0.004789 3.113233 1.477184 \n",
+ "104 46.525209 34.355134 6.537749 ... 0.005467 3.781467 1.575306 \n",
+ "178 48.224169 31.048502 12.579528 ... 0.017332 3.884780 1.356908 \n",
+ "142 47.823053 35.326472 20.710882 ... 0.012084 3.654480 1.778057 \n",
+ "\n",
+ " nqc+100 smv+100 clarity+100+100 wig+1000 nqc+1000 smv+1000 \\\n",
+ "70 0.016508 0.009057 3.544058 0.745923 0.009943 0.006240 \n",
+ "85 0.018038 0.013737 3.935562 1.029481 0.012135 0.007093 \n",
+ "43 0.018584 0.015681 3.456468 0.938830 0.009677 0.004896 \n",
+ "116 0.006738 0.004749 3.618240 1.054783 0.012100 0.009408 \n",
+ "122 0.015672 0.013502 3.459321 0.914128 0.012781 0.007959 \n",
+ ".. ... ... ... ... ... ... \n",
+ "19 0.010545 0.007952 4.641726 0.926463 0.007980 0.004847 \n",
+ "95 0.008821 0.007782 3.153661 0.824822 0.007862 0.005510 \n",
+ "104 0.021174 0.019889 3.714051 0.716917 0.012303 0.006179 \n",
+ "178 0.018738 0.015678 3.836012 0.721010 0.011075 0.006777 \n",
+ "142 0.012714 0.010522 3.645432 0.669576 0.009267 0.005655 \n",
+ "\n",
+ " clarity+1000+100 \n",
+ "70 3.544040 \n",
+ "85 3.716341 \n",
+ "43 3.187760 \n",
+ "116 3.312848 \n",
+ "122 3.326919 \n",
+ ".. ... \n",
+ "19 4.121996 \n",
+ "95 2.968856 \n",
+ "104 3.498253 \n",
+ "178 3.563908 \n",
+ "142 3.632626 \n",
+ "\n",
+ "[200 rows x 36 columns]"
+ ]
+ },
+ "execution_count": 8,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "# now with all qpp predictions\n",
+ "qpp_predictions(df_eval)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Gtq7M4GvJRwi"
+ },
+ "source": [
+ "Here we now can do some evaluations, e.g., measureing correlations or similar."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "KkTNdTzvJQS4"
+ },
+ "source": [
+ "# Now, prepare this evaluation slightly bigger, on multiple datasets\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:45:40.444234Z",
+ "start_time": "2024-02-27T09:44:26.683458Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "it34eGzzJffr",
+ "outputId": "39734b7d-60df-4f36-dad9-cc9da8b573c4"
+ },
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ " 17%|█████████████████████████████▋ | 4/23 [00:10<00:53, 2.81s/it]/home/sh3/S3806763/miniconda3/envs/tira-qpp/lib/python3.10/site-packages/pyterrier/pipelines.py:129: UserWarning: 73 topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.\n",
+ " warn(f'{backfill_count} topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.')\n",
+ " 22%|█████████████████████████████████████▏ | 5/23 [00:13<00:54, 3.03s/it]/home/sh3/S3806763/miniconda3/envs/tira-qpp/lib/python3.10/site-packages/pyterrier/pipelines.py:129: UserWarning: 1 topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.\n",
+ " warn(f'{backfill_count} topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.')\n",
+ " 57%|████████████████████████████████████████████████████████████████████████████████████████████████ | 13/23 [00:42<00:33, 3.37s/it]/home/sh3/S3806763/miniconda3/envs/tira-qpp/lib/python3.10/site-packages/pyterrier/pipelines.py:129: UserWarning: 1 topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.\n",
+ " warn(f'{backfill_count} topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.')\n",
+ " 65%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████▊ | 15/23 [00:49<00:25, 3.23s/it]/home/sh3/S3806763/miniconda3/envs/tira-qpp/lib/python3.10/site-packages/pyterrier/pipelines.py:129: UserWarning: 1 topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.\n",
+ " warn(f'{backfill_count} topic(s) not found in qrels. Scores for these topics are given as NaN and should not contribute to averages.')\n",
+ "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 23/23 [01:13<00:00, 3.21s/it]\n"
+ ]
+ }
+ ],
+ "source": [
+ "datasets = [\n",
+ " 'antique-test-20230107-training', 'argsme-touche-2021-task-1-20230209-training',\n",
+ " 'argsme-touche-2020-task-1-20230209-training',\n",
+ " 'cord19-fulltext-trec-covid-20230107-training', 'cranfield-20230107-training',\n",
+ " 'disks45-nocr-trec-robust-2004-20230209-training',\n",
+ " 'disks45-nocr-trec7-20230209-training', 'disks45-nocr-trec8-20230209-training',\n",
+ " 'gov-trec-web-2002-20230209-training',\n",
+ " 'gov-trec-web-2003-20230209-training', 'gov-trec-web-2004-20230209-training', 'gov2-trec-tb-2006-20230209-training',\n",
+ " 'gov2-trec-tb-2005-20230209-training', 'gov2-trec-tb-2004-20230209-training',\n",
+ " 'medline-2004-trec-genomics-2004-20230107-training',\n",
+ " 'medline-2004-trec-genomics-2005-20230107-training', 'medline-2017-trec-pm-2017-20230211-training',\n",
+ " 'medline-2017-trec-pm-2018-20230211-training', 'msmarco-passage-trec-dl-2019-judged-20230107-training',\n",
+ " 'msmarco-passage-trec-dl-2020-judged-20230107-training',\n",
+ " 'nfcorpus-test-20230107-training', 'vaswani-20230107-training',\n",
+ " 'wapo-v2-trec-core-2018-20230107-training'\n",
+ "]\n",
+ "df_eval = []\n",
+ "\n",
+ "for dataset in tqdm(datasets):\n",
+ " pt_dataset = pt.get_dataset(f\"irds:ir-benchmarks/{dataset}\")\n",
+ " bm25 = tira.pt.from_submission('ir-benchmarks/tira-ir-starter/BM25 Re-Rank (tira-ir-starter-pyterrier)', dataset)\n",
+ " df = pt.Experiment([bm25], pt_dataset.get_topics('query'), pt_dataset.get_qrels(), names=['BM25'],\n",
+ " eval_metrics=['ndcg_cut.10', 'map'], perquery=True)\n",
+ "\n",
+ " qpp_predictions = tira.pt.transform_queries('ir-benchmarks/qpptk/all-predictors', dataset)\n",
+ " df['dataset'] = dataset\n",
+ " df_eval += [qpp_predictions(df)]\n",
+ "df_eval = pd.concat(df_eval)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:45:40.473898Z",
+ "start_time": "2024-02-27T09:45:40.446329Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 1000
+ },
+ "id": "2iEy57VaEtMI",
+ "outputId": "08f3713e-f4ce-417d-f8ad-814a3a10e940"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " name | \n",
+ " qid | \n",
+ " measure | \n",
+ " value | \n",
+ " dataset | \n",
+ " max-idf | \n",
+ " avg-idf | \n",
+ " scq | \n",
+ " max-scq | \n",
+ " avg-scq | \n",
+ " ... | \n",
+ " smv+50 | \n",
+ " clarity+50+100 | \n",
+ " wig+100 | \n",
+ " nqc+100 | \n",
+ " smv+100 | \n",
+ " clarity+100+100 | \n",
+ " wig+1000 | \n",
+ " nqc+1000 | \n",
+ " smv+1000 | \n",
+ " clarity+1000+100 | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 140 | \n",
+ " BM25 | \n",
+ " 100653 | \n",
+ " map | \n",
+ " 0.571908 | \n",
+ " antique-test-20230107-training | \n",
+ " 12.908343 | \n",
+ " 9.051803 | \n",
+ " 598.373506 | \n",
+ " 50.637838 | \n",
+ " 37.398344 | \n",
+ " ... | \n",
+ " 0.014273 | \n",
+ " 3.544082 | \n",
+ " 1.951133 | \n",
+ " 0.016508 | \n",
+ " 0.009057 | \n",
+ " 3.544058 | \n",
+ " 0.745923 | \n",
+ " 0.009943 | \n",
+ " 0.006240 | \n",
+ " 3.544040 | \n",
+ "
\n",
+ " \n",
+ " 141 | \n",
+ " BM25 | \n",
+ " 100653 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.428160 | \n",
+ " antique-test-20230107-training | \n",
+ " 12.908343 | \n",
+ " 9.051803 | \n",
+ " 598.373506 | \n",
+ " 50.637838 | \n",
+ " 37.398344 | \n",
+ " ... | \n",
+ " 0.014273 | \n",
+ " 3.544082 | \n",
+ " 1.951133 | \n",
+ " 0.016508 | \n",
+ " 0.009057 | \n",
+ " 3.544058 | \n",
+ " 0.745923 | \n",
+ " 0.009943 | \n",
+ " 0.006240 | \n",
+ " 3.544040 | \n",
+ "
\n",
+ " \n",
+ " 170 | \n",
+ " BM25 | \n",
+ " 1015624 | \n",
+ " map | \n",
+ " 0.667378 | \n",
+ " antique-test-20230107-training | \n",
+ " 12.908343 | \n",
+ " 8.369403 | \n",
+ " 280.338094 | \n",
+ " 50.578374 | \n",
+ " 40.048299 | \n",
+ " ... | \n",
+ " 0.011114 | \n",
+ " 3.999061 | \n",
+ " 1.929926 | \n",
+ " 0.018038 | \n",
+ " 0.013737 | \n",
+ " 3.935562 | \n",
+ " 1.029481 | \n",
+ " 0.012135 | \n",
+ " 0.007093 | \n",
+ " 3.716341 | \n",
+ "
\n",
+ " \n",
+ " 171 | \n",
+ " BM25 | \n",
+ " 1015624 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.580570 | \n",
+ " antique-test-20230107-training | \n",
+ " 12.908343 | \n",
+ " 8.369403 | \n",
+ " 280.338094 | \n",
+ " 50.578374 | \n",
+ " 40.048299 | \n",
+ " ... | \n",
+ " 0.011114 | \n",
+ " 3.999061 | \n",
+ " 1.929926 | \n",
+ " 0.018038 | \n",
+ " 0.013737 | \n",
+ " 3.935562 | \n",
+ " 1.029481 | \n",
+ " 0.012135 | \n",
+ " 0.007093 | \n",
+ " 3.716341 | \n",
+ "
\n",
+ " \n",
+ " 86 | \n",
+ " BM25 | \n",
+ " 1017690 | \n",
+ " map | \n",
+ " 0.606246 | \n",
+ " antique-test-20230107-training | \n",
+ " 12.215196 | \n",
+ " 7.912693 | \n",
+ " 283.592656 | \n",
+ " 49.248037 | \n",
+ " 40.513237 | \n",
+ " ... | \n",
+ " 0.013975 | \n",
+ " 3.511534 | \n",
+ " 1.571078 | \n",
+ " 0.018584 | \n",
+ " 0.015681 | \n",
+ " 3.456468 | \n",
+ " 0.938830 | \n",
+ " 0.009677 | \n",
+ " 0.004896 | \n",
+ " 3.187760 | \n",
+ "
\n",
+ " \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " 95 | \n",
+ " BM25 | \n",
+ " 823 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.779171 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 9.253328 | \n",
+ " 6.472552 | \n",
+ " 133.850744 | \n",
+ " 53.319835 | \n",
+ " 44.616915 | \n",
+ " ... | \n",
+ " 0.018770 | \n",
+ " 6.755578 | \n",
+ " 4.113982 | \n",
+ " 0.024965 | \n",
+ " 0.018158 | \n",
+ " 6.492418 | \n",
+ " 1.785332 | \n",
+ " 0.057001 | \n",
+ " 0.048890 | \n",
+ " 6.120643 | \n",
+ "
\n",
+ " \n",
+ " 96 | \n",
+ " BM25 | \n",
+ " 824 | \n",
+ " map | \n",
+ " 0.314803 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 6.043616 | \n",
+ " 2.944800 | \n",
+ " 132.174833 | \n",
+ " 54.350762 | \n",
+ " 33.043708 | \n",
+ " ... | \n",
+ " 0.035296 | \n",
+ " 5.879690 | \n",
+ " 4.295972 | \n",
+ " 0.046101 | \n",
+ " 0.036535 | \n",
+ " 5.865900 | \n",
+ " 2.664010 | \n",
+ " 0.045072 | \n",
+ " 0.033262 | \n",
+ " 5.825797 | \n",
+ "
\n",
+ " \n",
+ " 97 | \n",
+ " BM25 | \n",
+ " 824 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.691973 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 6.043616 | \n",
+ " 2.944800 | \n",
+ " 132.174833 | \n",
+ " 54.350762 | \n",
+ " 33.043708 | \n",
+ " ... | \n",
+ " 0.035296 | \n",
+ " 5.879690 | \n",
+ " 4.295972 | \n",
+ " 0.046101 | \n",
+ " 0.036535 | \n",
+ " 5.865900 | \n",
+ " 2.664010 | \n",
+ " 0.045072 | \n",
+ " 0.033262 | \n",
+ " 5.825797 | \n",
+ "
\n",
+ " \n",
+ " 98 | \n",
+ " BM25 | \n",
+ " 825 | \n",
+ " map | \n",
+ " 0.302538 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 7.867033 | \n",
+ " 5.313623 | \n",
+ " 182.362550 | \n",
+ " 58.058460 | \n",
+ " 45.590637 | \n",
+ " ... | \n",
+ " 0.020023 | \n",
+ " 6.314339 | \n",
+ " 3.748762 | \n",
+ " 0.027191 | \n",
+ " 0.021909 | \n",
+ " 6.201889 | \n",
+ " 2.259479 | \n",
+ " 0.032763 | \n",
+ " 0.025157 | \n",
+ " 5.856790 | \n",
+ "
\n",
+ " \n",
+ " 99 | \n",
+ " BM25 | \n",
+ " 825 | \n",
+ " ndcg_cut.10 | \n",
+ " 0.634952 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 7.867033 | \n",
+ " 5.313623 | \n",
+ " 182.362550 | \n",
+ " 58.058460 | \n",
+ " 45.590637 | \n",
+ " ... | \n",
+ " 0.020023 | \n",
+ " 6.314339 | \n",
+ " 3.748762 | \n",
+ " 0.027191 | \n",
+ " 0.021909 | \n",
+ " 6.201889 | \n",
+ " 2.259479 | \n",
+ " 0.032763 | \n",
+ " 0.025157 | \n",
+ " 5.856790 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
4288 rows × 37 columns
\n",
+ "
"
+ ],
+ "text/plain": [
+ " name qid measure value \\\n",
+ "140 BM25 100653 map 0.571908 \n",
+ "141 BM25 100653 ndcg_cut.10 0.428160 \n",
+ "170 BM25 1015624 map 0.667378 \n",
+ "171 BM25 1015624 ndcg_cut.10 0.580570 \n",
+ "86 BM25 1017690 map 0.606246 \n",
+ ".. ... ... ... ... \n",
+ "95 BM25 823 ndcg_cut.10 0.779171 \n",
+ "96 BM25 824 map 0.314803 \n",
+ "97 BM25 824 ndcg_cut.10 0.691973 \n",
+ "98 BM25 825 map 0.302538 \n",
+ "99 BM25 825 ndcg_cut.10 0.634952 \n",
+ "\n",
+ " dataset max-idf avg-idf \\\n",
+ "140 antique-test-20230107-training 12.908343 9.051803 \n",
+ "141 antique-test-20230107-training 12.908343 9.051803 \n",
+ "170 antique-test-20230107-training 12.908343 8.369403 \n",
+ "171 antique-test-20230107-training 12.908343 8.369403 \n",
+ "86 antique-test-20230107-training 12.215196 7.912693 \n",
+ ".. ... ... ... \n",
+ "95 wapo-v2-trec-core-2018-20230107-training 9.253328 6.472552 \n",
+ "96 wapo-v2-trec-core-2018-20230107-training 6.043616 2.944800 \n",
+ "97 wapo-v2-trec-core-2018-20230107-training 6.043616 2.944800 \n",
+ "98 wapo-v2-trec-core-2018-20230107-training 7.867033 5.313623 \n",
+ "99 wapo-v2-trec-core-2018-20230107-training 7.867033 5.313623 \n",
+ "\n",
+ " scq max-scq avg-scq ... smv+50 clarity+50+100 \\\n",
+ "140 598.373506 50.637838 37.398344 ... 0.014273 3.544082 \n",
+ "141 598.373506 50.637838 37.398344 ... 0.014273 3.544082 \n",
+ "170 280.338094 50.578374 40.048299 ... 0.011114 3.999061 \n",
+ "171 280.338094 50.578374 40.048299 ... 0.011114 3.999061 \n",
+ "86 283.592656 49.248037 40.513237 ... 0.013975 3.511534 \n",
+ ".. ... ... ... ... ... ... \n",
+ "95 133.850744 53.319835 44.616915 ... 0.018770 6.755578 \n",
+ "96 132.174833 54.350762 33.043708 ... 0.035296 5.879690 \n",
+ "97 132.174833 54.350762 33.043708 ... 0.035296 5.879690 \n",
+ "98 182.362550 58.058460 45.590637 ... 0.020023 6.314339 \n",
+ "99 182.362550 58.058460 45.590637 ... 0.020023 6.314339 \n",
+ "\n",
+ " wig+100 nqc+100 smv+100 clarity+100+100 wig+1000 nqc+1000 \\\n",
+ "140 1.951133 0.016508 0.009057 3.544058 0.745923 0.009943 \n",
+ "141 1.951133 0.016508 0.009057 3.544058 0.745923 0.009943 \n",
+ "170 1.929926 0.018038 0.013737 3.935562 1.029481 0.012135 \n",
+ "171 1.929926 0.018038 0.013737 3.935562 1.029481 0.012135 \n",
+ "86 1.571078 0.018584 0.015681 3.456468 0.938830 0.009677 \n",
+ ".. ... ... ... ... ... ... \n",
+ "95 4.113982 0.024965 0.018158 6.492418 1.785332 0.057001 \n",
+ "96 4.295972 0.046101 0.036535 5.865900 2.664010 0.045072 \n",
+ "97 4.295972 0.046101 0.036535 5.865900 2.664010 0.045072 \n",
+ "98 3.748762 0.027191 0.021909 6.201889 2.259479 0.032763 \n",
+ "99 3.748762 0.027191 0.021909 6.201889 2.259479 0.032763 \n",
+ "\n",
+ " smv+1000 clarity+1000+100 \n",
+ "140 0.006240 3.544040 \n",
+ "141 0.006240 3.544040 \n",
+ "170 0.007093 3.716341 \n",
+ "171 0.007093 3.716341 \n",
+ "86 0.004896 3.187760 \n",
+ ".. ... ... \n",
+ "95 0.048890 6.120643 \n",
+ "96 0.033262 5.825797 \n",
+ "97 0.033262 5.825797 \n",
+ "98 0.025157 5.856790 \n",
+ "99 0.025157 5.856790 \n",
+ "\n",
+ "[4288 rows x 37 columns]"
+ ]
+ },
+ "execution_count": 10,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "df_eval"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:45:40.501808Z",
+ "start_time": "2024-02-27T09:45:40.475513Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 424
+ },
+ "id": "kaHsxq5PmMei",
+ "outputId": "80bb46e5-1115-4764-9902-dba47c05924a"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " name | \n",
+ " dataset | \n",
+ " qid | \n",
+ " ndcg@10 | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 141 | \n",
+ " BM25 | \n",
+ " antique-test-20230107-training | \n",
+ " 100653 | \n",
+ " 0.428160 | \n",
+ "
\n",
+ " \n",
+ " 171 | \n",
+ " BM25 | \n",
+ " antique-test-20230107-training | \n",
+ " 1015624 | \n",
+ " 0.580570 | \n",
+ "
\n",
+ " \n",
+ " 87 | \n",
+ " BM25 | \n",
+ " antique-test-20230107-training | \n",
+ " 1017690 | \n",
+ " 0.754357 | \n",
+ "
\n",
+ " \n",
+ " 233 | \n",
+ " BM25 | \n",
+ " antique-test-20230107-training | \n",
+ " 1035857 | \n",
+ " 0.652168 | \n",
+ "
\n",
+ " \n",
+ " 245 | \n",
+ " BM25 | \n",
+ " antique-test-20230107-training | \n",
+ " 103830 | \n",
+ " 0.530858 | \n",
+ "
\n",
+ " \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " 91 | \n",
+ " BM25 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 821 | \n",
+ " 0.447272 | \n",
+ "
\n",
+ " \n",
+ " 93 | \n",
+ " BM25 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 822 | \n",
+ " 0.521296 | \n",
+ "
\n",
+ " \n",
+ " 95 | \n",
+ " BM25 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 823 | \n",
+ " 0.779171 | \n",
+ "
\n",
+ " \n",
+ " 97 | \n",
+ " BM25 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 824 | \n",
+ " 0.691973 | \n",
+ "
\n",
+ " \n",
+ " 99 | \n",
+ " BM25 | \n",
+ " wapo-v2-trec-core-2018-20230107-training | \n",
+ " 825 | \n",
+ " 0.634952 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
2144 rows × 4 columns
\n",
+ "
"
+ ],
+ "text/plain": [
+ " name dataset qid ndcg@10\n",
+ "141 BM25 antique-test-20230107-training 100653 0.428160\n",
+ "171 BM25 antique-test-20230107-training 1015624 0.580570\n",
+ "87 BM25 antique-test-20230107-training 1017690 0.754357\n",
+ "233 BM25 antique-test-20230107-training 1035857 0.652168\n",
+ "245 BM25 antique-test-20230107-training 103830 0.530858\n",
+ ".. ... ... ... ...\n",
+ "91 BM25 wapo-v2-trec-core-2018-20230107-training 821 0.447272\n",
+ "93 BM25 wapo-v2-trec-core-2018-20230107-training 822 0.521296\n",
+ "95 BM25 wapo-v2-trec-core-2018-20230107-training 823 0.779171\n",
+ "97 BM25 wapo-v2-trec-core-2018-20230107-training 824 0.691973\n",
+ "99 BM25 wapo-v2-trec-core-2018-20230107-training 825 0.634952\n",
+ "\n",
+ "[2144 rows x 4 columns]"
+ ]
+ },
+ "execution_count": 11,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "ap_df = df_eval.loc[df_eval['measure'] == 'map', ['name', 'dataset', 'qid', 'value']].rename({'value': 'map'}, axis=1)\n",
+ "ndcg_10_df = df_eval.loc[df_eval['measure'] == 'ndcg_cut.10', ['name', 'dataset', 'qid', 'value']].rename(\n",
+ " {'value': 'ndcg@10'}, axis=1)\n",
+ "ndcg_10_df"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:45:40.505660Z",
+ "start_time": "2024-02-27T09:45:40.503241Z"
+ },
+ "id": "a2FNhMMWn3sp"
+ },
+ "outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T09:45:40.541559Z",
+ "start_time": "2024-02-27T09:45:40.507058Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 444
+ },
+ "id": "1eVIWHT6kbqU",
+ "outputId": "c4f5abb4-8b89-41bd-9193-17aee9c3bcc1"
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " wig+5 | \n",
+ " nqc+5 | \n",
+ " smv+5 | \n",
+ " clarity+5+100 | \n",
+ " wig+10 | \n",
+ " nqc+10 | \n",
+ " smv+10 | \n",
+ " clarity+10+100 | \n",
+ " wig+20 | \n",
+ " nqc+20 | \n",
+ " ... | \n",
+ " smv+50 | \n",
+ " clarity+50+100 | \n",
+ " wig+100 | \n",
+ " nqc+100 | \n",
+ " smv+100 | \n",
+ " clarity+100+100 | \n",
+ " wig+1000 | \n",
+ " nqc+1000 | \n",
+ " smv+1000 | \n",
+ " clarity+1000+100 | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 140 | \n",
+ " 5.114781 | \n",
+ " 0.000341 | \n",
+ " 0.000256 | \n",
+ " 3.856976 | \n",
+ " 4.320086 | \n",
+ " 0.021442 | \n",
+ " 0.019348 | \n",
+ " 3.544212 | \n",
+ " 3.190317 | \n",
+ " 0.025751 | \n",
+ " ... | \n",
+ " 0.014273 | \n",
+ " 3.544082 | \n",
+ " 1.951133 | \n",
+ " 0.016508 | \n",
+ " 0.009057 | \n",
+ " 3.544058 | \n",
+ " 0.745923 | \n",
+ " 0.009943 | \n",
+ " 0.006240 | \n",
+ " 3.544040 | \n",
+ "
\n",
+ " \n",
+ " 141 | \n",
+ " 5.114781 | \n",
+ " 0.000341 | \n",
+ " 0.000256 | \n",
+ " 3.856976 | \n",
+ " 4.320086 | \n",
+ " 0.021442 | \n",
+ " 0.019348 | \n",
+ " 3.544212 | \n",
+ " 3.190317 | \n",
+ " 0.025751 | \n",
+ " ... | \n",
+ " 0.014273 | \n",
+ " 3.544082 | \n",
+ " 1.951133 | \n",
+ " 0.016508 | \n",
+ " 0.009057 | \n",
+ " 3.544058 | \n",
+ " 0.745923 | \n",
+ " 0.009943 | \n",
+ " 0.006240 | \n",
+ " 3.544040 | \n",
+ "
\n",
+ " \n",
+ " 170 | \n",
+ " 3.266062 | \n",
+ " 0.002495 | \n",
+ " 0.002357 | \n",
+ " 4.257047 | \n",
+ " 3.025123 | \n",
+ " 0.009375 | \n",
+ " 0.008483 | \n",
+ " 4.113981 | \n",
+ " 2.743009 | \n",
+ " 0.011908 | \n",
+ " ... | \n",
+ " 0.011114 | \n",
+ " 3.999061 | \n",
+ " 1.929926 | \n",
+ " 0.018038 | \n",
+ " 0.013737 | \n",
+ " 3.935562 | \n",
+ " 1.029481 | \n",
+ " 0.012135 | \n",
+ " 0.007093 | \n",
+ " 3.716341 | \n",
+ "
\n",
+ " \n",
+ " 171 | \n",
+ " 3.266062 | \n",
+ " 0.002495 | \n",
+ " 0.002357 | \n",
+ " 4.257047 | \n",
+ " 3.025123 | \n",
+ " 0.009375 | \n",
+ " 0.008483 | \n",
+ " 4.113981 | \n",
+ " 2.743009 | \n",
+ " 0.011908 | \n",
+ " ... | \n",
+ " 0.011114 | \n",
+ " 3.999061 | \n",
+ " 1.929926 | \n",
+ " 0.018038 | \n",
+ " 0.013737 | \n",
+ " 3.935562 | \n",
+ " 1.029481 | \n",
+ " 0.012135 | \n",
+ " 0.007093 | \n",
+ " 3.716341 | \n",
+ "
\n",
+ " \n",
+ " 86 | \n",
+ " 2.891790 | \n",
+ " 0.009229 | \n",
+ " 0.008948 | \n",
+ " 4.269795 | \n",
+ " 2.691789 | \n",
+ " 0.009796 | \n",
+ " 0.008685 | \n",
+ " 3.880512 | \n",
+ " 2.422779 | \n",
+ " 0.011821 | \n",
+ " ... | \n",
+ " 0.013975 | \n",
+ " 3.511534 | \n",
+ " 1.571078 | \n",
+ " 0.018584 | \n",
+ " 0.015681 | \n",
+ " 3.456468 | \n",
+ " 0.938830 | \n",
+ " 0.009677 | \n",
+ " 0.004896 | \n",
+ " 3.187760 | \n",
+ "
\n",
+ " \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " 95 | \n",
+ " 5.529098 | \n",
+ " 0.013639 | \n",
+ " 0.010667 | \n",
+ " 8.384763 | \n",
+ " 5.237965 | \n",
+ " 0.017567 | \n",
+ " 0.014289 | \n",
+ " 7.664940 | \n",
+ " 4.947712 | \n",
+ " 0.019486 | \n",
+ " ... | \n",
+ " 0.018770 | \n",
+ " 6.755578 | \n",
+ " 4.113982 | \n",
+ " 0.024965 | \n",
+ " 0.018158 | \n",
+ " 6.492418 | \n",
+ " 1.785332 | \n",
+ " 0.057001 | \n",
+ " 0.048890 | \n",
+ " 6.120643 | \n",
+ "
\n",
+ " \n",
+ " 96 | \n",
+ " 6.094802 | \n",
+ " 0.014414 | \n",
+ " 0.012791 | \n",
+ " 5.968235 | \n",
+ " 5.895828 | \n",
+ " 0.016362 | \n",
+ " 0.013797 | \n",
+ " 6.011216 | \n",
+ " 5.565914 | \n",
+ " 0.025530 | \n",
+ " ... | \n",
+ " 0.035296 | \n",
+ " 5.879690 | \n",
+ " 4.295972 | \n",
+ " 0.046101 | \n",
+ " 0.036535 | \n",
+ " 5.865900 | \n",
+ " 2.664010 | \n",
+ " 0.045072 | \n",
+ " 0.033262 | \n",
+ " 5.825797 | \n",
+ "
\n",
+ " \n",
+ " 97 | \n",
+ " 6.094802 | \n",
+ " 0.014414 | \n",
+ " 0.012791 | \n",
+ " 5.968235 | \n",
+ " 5.895828 | \n",
+ " 0.016362 | \n",
+ " 0.013797 | \n",
+ " 6.011216 | \n",
+ " 5.565914 | \n",
+ " 0.025530 | \n",
+ " ... | \n",
+ " 0.035296 | \n",
+ " 5.879690 | \n",
+ " 4.295972 | \n",
+ " 0.046101 | \n",
+ " 0.036535 | \n",
+ " 5.865900 | \n",
+ " 2.664010 | \n",
+ " 0.045072 | \n",
+ " 0.033262 | \n",
+ " 5.825797 | \n",
+ "
\n",
+ " \n",
+ " 98 | \n",
+ " 5.064507 | \n",
+ " 0.003857 | \n",
+ " 0.003020 | \n",
+ " 7.272382 | \n",
+ " 4.962457 | \n",
+ " 0.006068 | \n",
+ " 0.005158 | \n",
+ " 6.847458 | \n",
+ " 4.699434 | \n",
+ " 0.015013 | \n",
+ " ... | \n",
+ " 0.020023 | \n",
+ " 6.314339 | \n",
+ " 3.748762 | \n",
+ " 0.027191 | \n",
+ " 0.021909 | \n",
+ " 6.201889 | \n",
+ " 2.259479 | \n",
+ " 0.032763 | \n",
+ " 0.025157 | \n",
+ " 5.856790 | \n",
+ "
\n",
+ " \n",
+ " 99 | \n",
+ " 5.064507 | \n",
+ " 0.003857 | \n",
+ " 0.003020 | \n",
+ " 7.272382 | \n",
+ " 4.962457 | \n",
+ " 0.006068 | \n",
+ " 0.005158 | \n",
+ " 6.847458 | \n",
+ " 4.699434 | \n",
+ " 0.015013 | \n",
+ " ... | \n",
+ " 0.020023 | \n",
+ " 6.314339 | \n",
+ " 3.748762 | \n",
+ " 0.027191 | \n",
+ " 0.021909 | \n",
+ " 6.201889 | \n",
+ " 2.259479 | \n",
+ " 0.032763 | \n",
+ " 0.025157 | \n",
+ " 5.856790 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
4288 rows × 24 columns
\n",
+ "
"
+ ],
+ "text/plain": [
+ " wig+5 nqc+5 smv+5 clarity+5+100 wig+10 nqc+10 \\\n",
+ "140 5.114781 0.000341 0.000256 3.856976 4.320086 0.021442 \n",
+ "141 5.114781 0.000341 0.000256 3.856976 4.320086 0.021442 \n",
+ "170 3.266062 0.002495 0.002357 4.257047 3.025123 0.009375 \n",
+ "171 3.266062 0.002495 0.002357 4.257047 3.025123 0.009375 \n",
+ "86 2.891790 0.009229 0.008948 4.269795 2.691789 0.009796 \n",
+ ".. ... ... ... ... ... ... \n",
+ "95 5.529098 0.013639 0.010667 8.384763 5.237965 0.017567 \n",
+ "96 6.094802 0.014414 0.012791 5.968235 5.895828 0.016362 \n",
+ "97 6.094802 0.014414 0.012791 5.968235 5.895828 0.016362 \n",
+ "98 5.064507 0.003857 0.003020 7.272382 4.962457 0.006068 \n",
+ "99 5.064507 0.003857 0.003020 7.272382 4.962457 0.006068 \n",
+ "\n",
+ " smv+10 clarity+10+100 wig+20 nqc+20 ... smv+50 \\\n",
+ "140 0.019348 3.544212 3.190317 0.025751 ... 0.014273 \n",
+ "141 0.019348 3.544212 3.190317 0.025751 ... 0.014273 \n",
+ "170 0.008483 4.113981 2.743009 0.011908 ... 0.011114 \n",
+ "171 0.008483 4.113981 2.743009 0.011908 ... 0.011114 \n",
+ "86 0.008685 3.880512 2.422779 0.011821 ... 0.013975 \n",
+ ".. ... ... ... ... ... ... \n",
+ "95 0.014289 7.664940 4.947712 0.019486 ... 0.018770 \n",
+ "96 0.013797 6.011216 5.565914 0.025530 ... 0.035296 \n",
+ "97 0.013797 6.011216 5.565914 0.025530 ... 0.035296 \n",
+ "98 0.005158 6.847458 4.699434 0.015013 ... 0.020023 \n",
+ "99 0.005158 6.847458 4.699434 0.015013 ... 0.020023 \n",
+ "\n",
+ " clarity+50+100 wig+100 nqc+100 smv+100 clarity+100+100 wig+1000 \\\n",
+ "140 3.544082 1.951133 0.016508 0.009057 3.544058 0.745923 \n",
+ "141 3.544082 1.951133 0.016508 0.009057 3.544058 0.745923 \n",
+ "170 3.999061 1.929926 0.018038 0.013737 3.935562 1.029481 \n",
+ "171 3.999061 1.929926 0.018038 0.013737 3.935562 1.029481 \n",
+ "86 3.511534 1.571078 0.018584 0.015681 3.456468 0.938830 \n",
+ ".. ... ... ... ... ... ... \n",
+ "95 6.755578 4.113982 0.024965 0.018158 6.492418 1.785332 \n",
+ "96 5.879690 4.295972 0.046101 0.036535 5.865900 2.664010 \n",
+ "97 5.879690 4.295972 0.046101 0.036535 5.865900 2.664010 \n",
+ "98 6.314339 3.748762 0.027191 0.021909 6.201889 2.259479 \n",
+ "99 6.314339 3.748762 0.027191 0.021909 6.201889 2.259479 \n",
+ "\n",
+ " nqc+1000 smv+1000 clarity+1000+100 \n",
+ "140 0.009943 0.006240 3.544040 \n",
+ "141 0.009943 0.006240 3.544040 \n",
+ "170 0.012135 0.007093 3.716341 \n",
+ "171 0.012135 0.007093 3.716341 \n",
+ "86 0.009677 0.004896 3.187760 \n",
+ ".. ... ... ... \n",
+ "95 0.057001 0.048890 6.120643 \n",
+ "96 0.045072 0.033262 5.825797 \n",
+ "97 0.045072 0.033262 5.825797 \n",
+ "98 0.032763 0.025157 5.856790 \n",
+ "99 0.032763 0.025157 5.856790 \n",
+ "\n",
+ "[4288 rows x 24 columns]"
+ ]
+ },
+ "execution_count": 12,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "pre_ret_predictors = df_eval.columns[5:13]\n",
+ "post_ret_predictors = df_eval.columns[13:].str.split('+').str[0].unique()\n",
+ "post_ret_predictors\n",
+ "\n",
+ "pre_qpp_df = df_eval.iloc[:, 5:13]\n",
+ "post_qpp_df = df_eval.iloc[:, 13:]\n",
+ "post_qpp_df"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 57,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T11:08:27.690453Z",
+ "start_time": "2024-02-27T11:08:27.673798Z"
+ },
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 721
+ },
+ "id": "rrGwcml4npj8",
+ "outputId": "fc29895b-22e4-4f4a-9147-0d16e6e8e3bf"
+ },
+ "outputs": [],
+ "source": [
+ "import seaborn as sns\n",
+ "import matplotlib.pyplot as plt\n",
+ "import numpy as np\n",
+ "from scipy.stats import rankdata\n",
+ "import numpy.typing as npt\n",
+ "\n",
+ "PlotNames = {'scq': 'SCQ', 'avg-scq': 'AvgSCQ', 'max-scq': 'MaxSCQ', 'var': 'SumVAR', 'avg-var': 'AvgVAR',\n",
+ " 'max-var': 'MaxVAR', 'max-idf': 'MaxIDF', 'avg-idf': 'AvgIDF', 'clarity': 'Clarity', 'smv': 'SMV',\n",
+ " 'nqc': 'NQC', 'wig': 'WIG', 'qf': 'QF', 'uef-clarity': 'UEF(Clarity)', 'uef-smv': 'UEF(SMV)',\n",
+ " 'uef-nqc': 'UEF(NQC)', 'uef-wig': 'UEF(WIG)', 'uef-qf': 'UEF(QF)', 'sd': 'SD', 'mean': 'Mean',\n",
+ " 'entropy': 'Entropy', 'kl': 'KL', 'skew': 'Skew', 'kurtosis': 'Kurtosis', 'ap@1000': 'AP',\n",
+ " 'ndcg@10': 'nDCG@10', 'title': 'Title Queries', 'vars': 'Query Variants', 'lovins': r'\\texttt{lovins}',\n",
+ " 'porter': r'\\texttt{porter}', 'nostem': r'\\texttt{nostem}', 'indri': r'\\texttt{indri}',\n",
+ " 'nostop': r'\\texttt{nostop}', 'lingpipe': r'\\texttt{lingpipe}', 'atire': r'\\texttt{atire}',\n",
+ " 'zettair': r'\\texttt{zettair}'}\n",
+ "\n",
+ "plt.rcParams.update(plt.rcParamsDefault)\n",
+ "paper_fmt = {\n",
+ " \"font.family\": [\"Linux Libertine O\"],\n",
+ " \"font.serif\": [\"Linux Libertine O\"],\n",
+ " \"font.sans-serif\": \"Linux Biolinum\",\n",
+ " 'font.size': 12,\n",
+ " 'pdf.fonttype': 42,\n",
+ " # 'figure.facecolor': (0.98, 0.98, 0.98),\n",
+ " # 'figure.facecolor':'#212121',\n",
+ " # 'text.color': '#23373b',\n",
+ " # 'axes.labelcolor': 'white',\n",
+ " # 'xtick.color': 'white',\n",
+ " # 'ytick.color': 'white',\n",
+ " # 'axes.titlecolor':'white',\n",
+ " 'figure.dpi': 300,\n",
+ " 'savefig.dpi': 300,\n",
+ " 'figure.figsize': (8, 2),\n",
+ " 'legend.borderaxespad': 0.5,\n",
+ " 'legend.fontsize': 'small',\n",
+ " 'legend.title_fontsize': 'small',\n",
+ " # 'legend.facecolor':'white',\n",
+ " # \"axes.labelpad\": 20.0\n",
+ " \"axes.labelpad\": 5.0,\n",
+ " # 'axes.titlesize':'medium',\n",
+ " # 'axes.labelsize':'medium'\n",
+ "}\n",
+ "plt.rcParams.update(paper_fmt)\n",
+ "\n",
+ "\n",
+ "def calc_sare(x: npt.ArrayLike, y: npt.ArrayLike) -> npt.NDArray[np.float64]:\n",
+ " \"\"\"\n",
+ " Calculate the scaled Absolute Rank Error (sARE) between two arrays.\n",
+ " \"\"\"\n",
+ " assert len(x) == len(y), f'Lengths of a and b must be equal: {len(x)} != {len(y)}'\n",
+ " return abs(rankdata(x, method='average') - rankdata(y, method='average')) / len(x)\n",
+ "\n",
+ "\n",
+ "def calc_smare(x: npt.ArrayLike, y: npt.ArrayLike) -> float:\n",
+ " \"\"\"\n",
+ " Calculate the scaled Mean Absolute Rank Error (sMARE) between two arrays.\n",
+ " \"\"\"\n",
+ " return calc_sare(x, y).mean()\n",
+ "\n",
+ "\n",
+ "def calc_correlations(df, corr_method='pearson'):\n",
+ " if corr_method == 'smare':\n",
+ " corr_func = calc_smare\n",
+ " else:\n",
+ " corr_func = corr_method\n",
+ " _df = df.groupby(['measure', 'dataset']).corr(method=corr_func, numeric_only=True).drop(columns='value')\n",
+ " _df = _df.loc[_df.index.get_level_values(2) == 'value'].droplevel(2).stack().reset_index().rename(\n",
+ " columns={'level_2': 'qpp+params', 0: corr_method})\n",
+ " return _df.assign(QPP=_df['qpp+params'].str.split('+').str[0])\n",
+ "\n",
+ "\n",
+ "def plot_pre_qpp_boxplot(corr_df, qpp_methods, measure):\n",
+ " pre_order = corr_df.loc[corr_df['QPP'].isin(qpp_methods)].groupby('QPP')[measure].mean().index\n",
+ " ax = sns.boxplot(data=corr_df.set_index('QPP').loc[pre_order].rename(PlotNames), x='QPP', y=measure, hue='measure')\n",
+ " # annotate the boxplot outliers (fliers)\n",
+ " for i, artist in enumerate(ax.artists):\n",
+ " x = []\n",
+ " y = []\n",
+ " for j in range(i * 6, i * 6 + 6):\n",
+ " line = ax.lines[j]\n",
+ " x.extend(line.get_xdata())\n",
+ " y.extend(line.get_ydata())\n",
+ " y = [y[i] for i in range(len(y)) if i % 6 == 0]\n",
+ " x = [x[i] for i in range(len(x)) if i % 6 == 0]\n",
+ " for i in range(len(x)):\n",
+ " ax.text(x[i], y[i], f'{y[i]:.2f}', ha='center', va='bottom', fontsize=8, color='black')\n",
+ " plt.xticks(rotation=45)\n",
+ "\n",
+ " plt.title('Pre-Retrieval Predictors')\n",
+ " plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 55,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T11:04:25.957032Z",
+ "start_time": "2024-02-27T11:04:10.819596Z"
+ },
+ "collapsed": false
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " measure | \n",
+ " dataset | \n",
+ " qpp+params | \n",
+ " smare | \n",
+ " QPP | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 | \n",
+ " map | \n",
+ " antique-test-20230107-training | \n",
+ " max-idf | \n",
+ " 0.309125 | \n",
+ " max-idf | \n",
+ "
\n",
+ " \n",
+ " 1 | \n",
+ " map | \n",
+ " antique-test-20230107-training | \n",
+ " avg-idf | \n",
+ " 0.316800 | \n",
+ " avg-idf | \n",
+ "
\n",
+ " \n",
+ " 2 | \n",
+ " map | \n",
+ " antique-test-20230107-training | \n",
+ " scq | \n",
+ " 0.327675 | \n",
+ " scq | \n",
+ "
\n",
+ " \n",
+ " 3 | \n",
+ " map | \n",
+ " antique-test-20230107-training | \n",
+ " max-scq | \n",
+ " 0.266175 | \n",
+ " max-scq | \n",
+ "
\n",
+ " \n",
+ " 4 | \n",
+ " map | \n",
+ " antique-test-20230107-training | \n",
+ " avg-scq | \n",
+ " 0.300750 | \n",
+ " avg-scq | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
"
+ ],
+ "text/plain": [
+ " measure dataset qpp+params smare QPP\n",
+ "0 map antique-test-20230107-training max-idf 0.309125 max-idf\n",
+ "1 map antique-test-20230107-training avg-idf 0.316800 avg-idf\n",
+ "2 map antique-test-20230107-training scq 0.327675 scq\n",
+ "3 map antique-test-20230107-training max-scq 0.266175 max-scq\n",
+ "4 map antique-test-20230107-training avg-scq 0.300750 avg-scq"
+ ]
+ },
+ "execution_count": 55,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "pearson_df = calc_correlations(df_eval, 'pearson')\n",
+ "kendall_df = calc_correlations(df_eval, 'kendall')\n",
+ "smare_df = calc_correlations(df_eval, 'smare')\n",
+ "smare_df.head()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 59,
+ "metadata": {
+ "ExecuteTime": {
+ "end_time": "2024-02-27T11:09:17.256906Z",
+ "start_time": "2024-02-27T11:09:15.511670Z"
+ },
+ "collapsed": false
+ },
+ "outputs": [
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ "