--- a
+++ b/EvaluationBaseline/baseline_results_analysis.ipynb
@@ -0,0 +1,1315 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Results evaluation from baseline"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 60,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import re"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 98,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[\"        full evaluation metrix: {'overall': {'acc': 0.7999, 'strict': {'precision': 0.5863842662632375, 'recall': 0.6398151205018158, 'f_score': 0.6119355857278181}, 'relax': {'precision': 0.7310136157337368, 'recall': 0.7976229778804886, 'f_score': 0.7628670666245659}}, 'category': {'strict': {'condition': {'precision': 0.652555910543131, 'recall': 0.7502295684113865, 'f_score': 0.6979923109782143}, 'procedure': {'precision': 0.4088888888888889, 'recall': 0.5993485342019544, 'f_score': 0.4861294583883752}, 'temporal': {'precision': 0.48656716417910445, 'recall': 0.5361842105263158, 'f_score': 0.5101721439749609}, 'measurement': {'precision': 0.5589225589225589, 'recall': 0.6287878787878788, 'f_score': 0.5918003565062389}, 'value': {'precision': 0.6620111731843575, 'recall': 0.7247706422018348, 'f_score': 0.691970802919708}, 'person': {'precision': 0.773972602739726, 'recall': 0.8129496402877698, 'f_score': 0.7929824561403509}, 'drug': {'precision': 0.5916870415647921, 'recall': 0.7289156626506024, 'f_score': 0.6531713900134952}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'observation': {'precision': 0.5161290322580645, 'recall': 0.09937888198757763, 'f_score': 0.16666666666666666}}, 'relax': {'condition': {'precision': 0.7931309904153354, 'recall': 0.9118457300275482, 'f_score': 0.8483554036736437}, 'procedure': {'precision': 0.5555555555555556, 'recall': 0.8143322475570033, 'f_score': 0.6605019815059445}, 'temporal': {'precision': 0.6716417910447762, 'recall': 0.7401315789473685, 'f_score': 0.7042253521126761}, 'measurement': {'precision': 0.734006734006734, 'recall': 0.8257575757575758, 'f_score': 0.7771836007130125}, 'value': {'precision': 0.8044692737430168, 'recall': 0.8807339449541285, 'f_score': 0.8408759124087591}, 'person': {'precision': 0.7808219178082192, 'recall': 0.8201438848920863, 'f_score': 0.7999999999999999}, 'drug': {'precision': 0.7310513447432763, 'recall': 0.9006024096385542, 'f_score': 0.8070175438596492}, 'pregnancy_considerations': {'precision': 0.2962962962962963, 'recall': 0.5714285714285714, 'f_score': 0.3902439024390244}, 'observation': {'precision': 0.6774193548387096, 'recall': 0.13043478260869565, 'f_score': 0.21875}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8083, 'strict': {'precision': 0.6034431582071832, 'recall': 0.6711786068009244, 'f_score': 0.6355110972178806}, 'relax': {'precision': 0.7441377263282873, 'recall': 0.8276658963354242, 'f_score': 0.7836824007502344}}, 'category': {'strict': {'condition': {'precision': 0.6642394822006472, 'recall': 0.7539026629935721, 'f_score': 0.7062365591397849}, 'temporal': {'precision': 0.504, 'recall': 0.6217105263157895, 'f_score': 0.5567010309278352}, 'procedure': {'precision': 0.5079787234042553, 'recall': 0.6221498371335505, 'f_score': 0.5592972181551976}, 'measurement': {'precision': 0.6063829787234043, 'recall': 0.6477272727272727, 'f_score': 0.6263736263736264}, 'value': {'precision': 0.6912181303116147, 'recall': 0.746177370030581, 'f_score': 0.7176470588235294}, 'observation': {'precision': 0.44, 'recall': 0.2732919254658385, 'f_score': 0.3371647509578544}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'person': {'precision': 0.7531645569620253, 'recall': 0.8561151079136691, 'f_score': 0.8013468013468013}, 'drug': {'precision': 0.6275, 'recall': 0.7560240963855421, 'f_score': 0.6857923497267758}, 'device': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'mood': {'precision': 0.375, 'recall': 0.046153846153846156, 'f_score': 0.08219178082191782}}, 'relax': {'condition': {'precision': 0.8074433656957929, 'recall': 0.9164370982552801, 'f_score': 0.858494623655914}, 'temporal': {'precision': 0.664, 'recall': 0.819078947368421, 'f_score': 0.7334315169366716}, 'procedure': {'precision': 0.6648936170212766, 'recall': 0.8143322475570033, 'f_score': 0.7320644216691069}, 'measurement': {'precision': 0.7801418439716312, 'recall': 0.8333333333333334, 'f_score': 0.8058608058608059}, 'value': {'precision': 0.8186968838526912, 'recall': 0.8837920489296636, 'f_score': 0.85}, 'observation': {'precision': 0.57, 'recall': 0.35403726708074534, 'f_score': 0.43678160919540227}, 'pregnancy_considerations': {'precision': 0.16666666666666666, 'recall': 0.8571428571428571, 'f_score': 0.27906976744186046}, 'person': {'precision': 0.759493670886076, 'recall': 0.8633093525179856, 'f_score': 0.8080808080808081}, 'drug': {'precision': 0.7475, 'recall': 0.9006024096385542, 'f_score': 0.8169398907103825}, 'device': {'precision': 1.0, 'recall': 0.3333333333333333, 'f_score': 0.5}, 'mood': {'precision': 0.5, 'recall': 0.06153846153846154, 'f_score': 0.10958904109589042}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8188, 'strict': {'precision': 0.623020706455542, 'recall': 0.6754704522944867, 'f_score': 0.6481862822746713}, 'relax': {'precision': 0.7633982947624848, 'recall': 0.8276658963354242, 'f_score': 0.7942341200696975}}, 'category': {'strict': {'condition': {'precision': 0.6878661087866109, 'recall': 0.7548209366391184, 'f_score': 0.7197898423817863}, 'procedure': {'precision': 0.5170603674540682, 'recall': 0.6416938110749185, 'f_score': 0.5726744186046512}, 'temporal': {'precision': 0.5373134328358209, 'recall': 0.5921052631578947, 'f_score': 0.5633802816901408}, 'measurement': {'precision': 0.6227758007117438, 'recall': 0.6628787878787878, 'f_score': 0.6422018348623854}, 'value': {'precision': 0.7025495750708215, 'recall': 0.7584097859327217, 'f_score': 0.7294117647058823}, 'observation': {'precision': 0.4, 'recall': 0.2484472049689441, 'f_score': 0.3065134099616858}, 'person': {'precision': 0.7828947368421053, 'recall': 0.8561151079136691, 'f_score': 0.8178694158075602}, 'drug': {'precision': 0.661498708010336, 'recall': 0.7710843373493976, 'f_score': 0.7121001390820584}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.15384615384615385, 'recall': 0.07407407407407407, 'f_score': 0.1}, 'mood': {'precision': 0.3888888888888889, 'recall': 0.1076923076923077, 'f_score': 0.16867469879518074}}, 'relax': {'condition': {'precision': 0.8384937238493724, 'recall': 0.9201101928374655, 'f_score': 0.8774080560420315}, 'procedure': {'precision': 0.6456692913385826, 'recall': 0.8013029315960912, 'f_score': 0.7151162790697674}, 'temporal': {'precision': 0.7104477611940299, 'recall': 0.7828947368421053, 'f_score': 0.7449139280125195}, 'measurement': {'precision': 0.8113879003558719, 'recall': 0.8636363636363636, 'f_score': 0.8366972477064221}, 'value': {'precision': 0.8243626062322946, 'recall': 0.8899082568807339, 'f_score': 0.8558823529411764}, 'observation': {'precision': 0.52, 'recall': 0.32298136645962733, 'f_score': 0.3984674329501916}, 'person': {'precision': 0.7894736842105263, 'recall': 0.8633093525179856, 'f_score': 0.8247422680412372}, 'drug': {'precision': 0.7777777777777778, 'recall': 0.9066265060240963, 'f_score': 0.8372739916550764}, 'pregnancy_considerations': {'precision': 0.13043478260869565, 'recall': 0.6428571428571429, 'f_score': 0.21686746987951808}, 'device': {'precision': 0.8461538461538461, 'recall': 0.4074074074074074, 'f_score': 0.5499999999999999}, 'mood': {'precision': 0.5, 'recall': 0.13846153846153847, 'f_score': 0.2168674698795181}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.7836, 'strict': {'precision': 0.5396489104116223, 'recall': 0.5886431165401123, 'f_score': 0.5630822674877625}, 'relax': {'precision': 0.7091404358353511, 'recall': 0.7735226147243315, 'f_score': 0.739933680720038}}, 'category': {'strict': {'condition': {'precision': 0.6074544012688342, 'recall': 0.7033976124885216, 'f_score': 0.6519148936170213}, 'procedure': {'precision': 0.3376865671641791, 'recall': 0.5895765472312704, 'f_score': 0.4294187425860024}, 'temporal': {'precision': 0.4316109422492401, 'recall': 0.46710526315789475, 'f_score': 0.4486571879936809}, 'measurement': {'precision': 0.5102739726027398, 'recall': 0.5643939393939394, 'f_score': 0.5359712230215827}, 'value': {'precision': 0.6069364161849711, 'recall': 0.6422018348623854, 'f_score': 0.6240713224368498}, 'person': {'precision': 0.781021897810219, 'recall': 0.7697841726618705, 'f_score': 0.7753623188405797}, 'drug': {'precision': 0.5905511811023622, 'recall': 0.677710843373494, 'f_score': 0.6311360448807855}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'observation': {'precision': 0.42857142857142855, 'recall': 0.018633540372670808, 'f_score': 0.03571428571428571}}, 'relax': {'condition': {'precision': 0.7787470261697066, 'recall': 0.9017447199265382, 'f_score': 0.8357446808510639}, 'procedure': {'precision': 0.48134328358208955, 'recall': 0.8403908794788274, 'f_score': 0.6120996441281139}, 'temporal': {'precision': 0.6534954407294833, 'recall': 0.7072368421052632, 'f_score': 0.679304897314376}, 'measurement': {'precision': 0.7397260273972602, 'recall': 0.8181818181818182, 'f_score': 0.776978417266187}, 'value': {'precision': 0.7947976878612717, 'recall': 0.8409785932721713, 'f_score': 0.8172362555720654}, 'person': {'precision': 0.7883211678832117, 'recall': 0.7769784172661871, 'f_score': 0.782608695652174}, 'drug': {'precision': 0.7322834645669292, 'recall': 0.8403614457831325, 'f_score': 0.782608695652174}, 'pregnancy_considerations': {'precision': 0.4, 'recall': 0.42857142857142855, 'f_score': 0.4137931034482759}, 'observation': {'precision': 0.5714285714285714, 'recall': 0.024844720496894408, 'f_score': 0.047619047619047616}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8075, 'strict': {'precision': 0.6046163069544365, 'recall': 0.6658963354242324, 'f_score': 0.6337784760408484}, 'relax': {'precision': 0.7470023980815348, 'recall': 0.8227137669197755, 'f_score': 0.7830322073841319}}, 'category': {'strict': {'condition': {'precision': 0.6650124069478908, 'recall': 0.7382920110192838, 'f_score': 0.6997389033942558}, 'temporal': {'precision': 0.4962025316455696, 'recall': 0.6447368421052632, 'f_score': 0.5608011444921315}, 'procedure': {'precision': 0.5041322314049587, 'recall': 0.5960912052117264, 'f_score': 0.5462686567164179}, 'measurement': {'precision': 0.6057347670250897, 'recall': 0.6401515151515151, 'f_score': 0.6224677716390423}, 'value': {'precision': 0.7046783625730995, 'recall': 0.7370030581039755, 'f_score': 0.7204783258594918}, 'observation': {'precision': 0.4423076923076923, 'recall': 0.2857142857142857, 'f_score': 0.3471698113207547}, 'person': {'precision': 0.7692307692307693, 'recall': 0.8633093525179856, 'f_score': 0.8135593220338984}, 'drug': {'precision': 0.6292682926829268, 'recall': 0.7771084337349398, 'f_score': 0.6954177897574124}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'mood': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}}, 'relax': {'condition': {'precision': 0.8221670802315963, 'recall': 0.9127640036730946, 'f_score': 0.8651000870322019}, 'temporal': {'precision': 0.640506329113924, 'recall': 0.8322368421052632, 'f_score': 0.7238912732474964}, 'procedure': {'precision': 0.6721763085399449, 'recall': 0.7947882736156352, 'f_score': 0.7283582089552239}, 'measurement': {'precision': 0.7813620071684588, 'recall': 0.8257575757575758, 'f_score': 0.8029465930018418}, 'value': {'precision': 0.827485380116959, 'recall': 0.8654434250764526, 'f_score': 0.8460388639760836}, 'observation': {'precision': 0.5576923076923077, 'recall': 0.36024844720496896, 'f_score': 0.4377358490566038}, 'person': {'precision': 0.7756410256410257, 'recall': 0.8705035971223022, 'f_score': 0.8203389830508475}, 'drug': {'precision': 0.7414634146341463, 'recall': 0.9156626506024096, 'f_score': 0.8194070080862533}, 'pregnancy_considerations': {'precision': 0.1323529411764706, 'recall': 0.6428571428571429, 'f_score': 0.21951219512195122}, 'device': {'precision': 0.8888888888888888, 'recall': 0.2962962962962963, 'f_score': 0.4444444444444444}, 'mood': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8123, 'strict': {'precision': 0.6153166421207659, 'recall': 0.6896665566193463, 'f_score': 0.650373599003736}, 'relax': {'precision': 0.7508100147275405, 'recall': 0.8415318586992406, 'f_score': 0.7935865504358655}}, 'category': {'strict': {'condition': {'precision': 0.6830065359477124, 'recall': 0.7676767676767676, 'f_score': 0.7228707306528318}, 'procedure': {'precision': 0.5201005025125628, 'recall': 0.6742671009771987, 'f_score': 0.5872340425531914}, 'temporal': {'precision': 0.5565749235474006, 'recall': 0.5986842105263158, 'f_score': 0.5768621236133122}, 'measurement': {'precision': 0.6330935251798561, 'recall': 0.6666666666666666, 'f_score': 0.6494464944649446}, 'value': {'precision': 0.7151162790697675, 'recall': 0.7522935779816514, 'f_score': 0.7332339791356185}, 'observation': {'precision': 0.3223684210526316, 'recall': 0.30434782608695654, 'f_score': 0.3130990415335463}, 'person': {'precision': 0.8650793650793651, 'recall': 0.7841726618705036, 'f_score': 0.8226415094339623}, 'drug': {'precision': 0.673469387755102, 'recall': 0.7951807228915663, 'f_score': 0.729281767955801}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.2962962962962963, 'recall': 0.2962962962962963, 'f_score': 0.2962962962962963}, 'mood': {'precision': 0.34285714285714286, 'recall': 0.18461538461538463, 'f_score': 0.24}}, 'relax': {'condition': {'precision': 0.8202614379084967, 'recall': 0.9219467401285583, 'f_score': 0.8681366191093817}, 'procedure': {'precision': 0.6457286432160804, 'recall': 0.8371335504885994, 'f_score': 0.7290780141843972}, 'temporal': {'precision': 0.7400611620795107, 'recall': 0.7960526315789473, 'f_score': 0.7670364500792394}, 'measurement': {'precision': 0.8129496402877698, 'recall': 0.8560606060606061, 'f_score': 0.8339483394833948}, 'value': {'precision': 0.8401162790697675, 'recall': 0.8837920489296636, 'f_score': 0.8614008941877794}, 'observation': {'precision': 0.4473684210526316, 'recall': 0.422360248447205, 'f_score': 0.43450479233226846}, 'person': {'precision': 0.8809523809523809, 'recall': 0.7985611510791367, 'f_score': 0.8377358490566038}, 'drug': {'precision': 0.7806122448979592, 'recall': 0.9216867469879518, 'f_score': 0.845303867403315}, 'pregnancy_considerations': {'precision': 0.14130434782608695, 'recall': 0.9285714285714286, 'f_score': 0.2452830188679245}, 'device': {'precision': 0.5185185185185185, 'recall': 0.5185185185185185, 'f_score': 0.5185185185185185}, 'mood': {'precision': 0.5428571428571428, 'recall': 0.2923076923076923, 'f_score': 0.38000000000000006}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8182, 'strict': {'precision': 0.6247777119146414, 'recall': 0.695939253879168, 'f_score': 0.6584413556145556}, 'relax': {'precision': 0.7593360995850622, 'recall': 0.845823704192803, 'f_score': 0.8002498828674058}}, 'category': {'strict': {'condition': {'precision': 0.6767036450079239, 'recall': 0.7842056932966024, 'f_score': 0.7264993619736283}, 'drug': {'precision': 0.6795865633074936, 'recall': 0.7921686746987951, 'f_score': 0.7315716272600834}, 'procedure': {'precision': 0.5284090909090909, 'recall': 0.6058631921824105, 'f_score': 0.5644916540212442}, 'temporal': {'precision': 0.5582089552238806, 'recall': 0.6151315789473685, 'f_score': 0.5852895148669797}, 'measurement': {'precision': 0.6202090592334495, 'recall': 0.6742424242424242, 'f_score': 0.6460980036297641}, 'value': {'precision': 0.6902173913043478, 'recall': 0.7767584097859327, 'f_score': 0.730935251798561}, 'observation': {'precision': 0.2975206611570248, 'recall': 0.2236024844720497, 'f_score': 0.2553191489361702}, 'mood': {'precision': 0.33962264150943394, 'recall': 0.27692307692307694, 'f_score': 0.30508474576271183}, 'person': {'precision': 0.7577639751552795, 'recall': 0.8776978417266187, 'f_score': 0.8133333333333332}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.45454545454545453, 'recall': 0.37037037037037035, 'f_score': 0.40816326530612246}}, 'relax': {'condition': {'precision': 0.803486529318542, 'recall': 0.931129476584022, 'f_score': 0.8626116546150574}, 'drug': {'precision': 0.7855297157622739, 'recall': 0.9156626506024096, 'f_score': 0.8456189151599444}, 'procedure': {'precision': 0.6988636363636364, 'recall': 0.8013029315960912, 'f_score': 0.7465857359635812}, 'temporal': {'precision': 0.7134328358208956, 'recall': 0.7861842105263158, 'f_score': 0.7480438184663537}, 'measurement': {'precision': 0.7979094076655052, 'recall': 0.8674242424242424, 'f_score': 0.8312159709618874}, 'value': {'precision': 0.8070652173913043, 'recall': 0.908256880733945, 'f_score': 0.8546762589928057}, 'observation': {'precision': 0.47107438016528924, 'recall': 0.35403726708074534, 'f_score': 0.40425531914893614}, 'mood': {'precision': 0.5471698113207547, 'recall': 0.4461538461538462, 'f_score': 0.4915254237288135}, 'person': {'precision': 0.7701863354037267, 'recall': 0.8920863309352518, 'f_score': 0.8266666666666668}, 'pregnancy_considerations': {'precision': 0.3076923076923077, 'recall': 0.5714285714285714, 'f_score': 0.4}, 'device': {'precision': 0.6818181818181818, 'recall': 0.5555555555555556, 'f_score': 0.6122448979591836}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8189, 'strict': {'precision': 0.6260355029585799, 'recall': 0.698580389567514, 'f_score': 0.6603214229989077}, 'relax': {'precision': 0.7615384615384615, 'recall': 0.8497854077253219, 'f_score': 0.8032454361054767}}, 'category': {'strict': {'condition': {'precision': 0.6889250814332247, 'recall': 0.7768595041322314, 'f_score': 0.7302546396201985}, 'procedure': {'precision': 0.5439560439560439, 'recall': 0.6449511400651465, 'f_score': 0.5901639344262295}, 'temporal': {'precision': 0.5696969696969697, 'recall': 0.618421052631579, 'f_score': 0.5930599369085174}, 'measurement': {'precision': 0.5811688311688312, 'recall': 0.678030303030303, 'f_score': 0.6258741258741259}, 'value': {'precision': 0.6843501326259946, 'recall': 0.7889908256880734, 'f_score': 0.7329545454545454}, 'observation': {'precision': 0.31297709923664124, 'recall': 0.2546583850931677, 'f_score': 0.2808219178082192}, 'person': {'precision': 0.7880794701986755, 'recall': 0.8561151079136691, 'f_score': 0.8206896551724139}, 'mood': {'precision': 0.3137254901960784, 'recall': 0.24615384615384617, 'f_score': 0.27586206896551724}, 'drug': {'precision': 0.6921052631578948, 'recall': 0.7921686746987951, 'f_score': 0.7387640449438202}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.26666666666666666, 'recall': 0.2962962962962963, 'f_score': 0.28070175438596495}}, 'relax': {'condition': {'precision': 0.8249185667752443, 'recall': 0.9302112029384757, 'f_score': 0.8744065602071645}, 'procedure': {'precision': 0.6978021978021978, 'recall': 0.8273615635179153, 'f_score': 0.7570789865871832}, 'temporal': {'precision': 0.7242424242424242, 'recall': 0.7861842105263158, 'f_score': 0.7539432176656152}, 'measurement': {'precision': 0.7662337662337663, 'recall': 0.8939393939393939, 'f_score': 0.8251748251748252}, 'value': {'precision': 0.7931034482758621, 'recall': 0.9143730886850153, 'f_score': 0.8494318181818183}, 'observation': {'precision': 0.4580152671755725, 'recall': 0.37267080745341613, 'f_score': 0.410958904109589}, 'person': {'precision': 0.8013245033112583, 'recall': 0.8705035971223022, 'f_score': 0.8344827586206897}, 'mood': {'precision': 0.49019607843137253, 'recall': 0.38461538461538464, 'f_score': 0.43103448275862066}, 'drug': {'precision': 0.7921052631578948, 'recall': 0.9066265060240963, 'f_score': 0.8455056179775281}, 'pregnancy_considerations': {'precision': 0.26666666666666666, 'recall': 0.5714285714285714, 'f_score': 0.36363636363636365}, 'device': {'precision': 0.6, 'recall': 0.6666666666666666, 'f_score': 0.631578947368421}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8155, 'strict': {'precision': 0.6268788682581786, 'recall': 0.7022119511389898, 'f_score': 0.6624104640298973}, 'relax': {'precision': 0.7606837606837606, 'recall': 0.8520964014526247, 'f_score': 0.8037994394269697}}, 'category': {'strict': {'condition': {'precision': 0.6882690730106645, 'recall': 0.7704315886134068, 'f_score': 0.7270363951473136}, 'temporal': {'precision': 0.5223097112860893, 'recall': 0.6546052631578947, 'f_score': 0.5810218978102191}, 'procedure': {'precision': 0.5333333333333333, 'recall': 0.6775244299674267, 'f_score': 0.5968436154949784}, 'measurement': {'precision': 0.6423357664233577, 'recall': 0.6666666666666666, 'f_score': 0.654275092936803}, 'value': {'precision': 0.7126760563380282, 'recall': 0.7737003058103975, 'f_score': 0.7419354838709676}, 'observation': {'precision': 0.3333333333333333, 'recall': 0.2732919254658385, 'f_score': 0.3003412969283276}, 'person': {'precision': 0.7894736842105263, 'recall': 0.8633093525179856, 'f_score': 0.8247422680412372}, 'mood': {'precision': 0.25925925925925924, 'recall': 0.2153846153846154, 'f_score': 0.23529411764705882}, 'drug': {'precision': 0.6902887139107612, 'recall': 0.7921686746987951, 'f_score': 0.7377279102384292}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.34375, 'recall': 0.4074074074074074, 'f_score': 0.37288135593220334}}, 'relax': {'condition': {'precision': 0.8252666119770303, 'recall': 0.9237832874196511, 'f_score': 0.8717504332755632}, 'temporal': {'precision': 0.6745406824146981, 'recall': 0.8453947368421053, 'f_score': 0.7503649635036497}, 'procedure': {'precision': 0.6666666666666666, 'recall': 0.8469055374592834, 'f_score': 0.7460545193687231}, 'measurement': {'precision': 0.8321167883211679, 'recall': 0.8636363636363636, 'f_score': 0.8475836431226766}, 'value': {'precision': 0.8225352112676056, 'recall': 0.8929663608562691, 'f_score': 0.8563049853372435}, 'observation': {'precision': 0.45454545454545453, 'recall': 0.37267080745341613, 'f_score': 0.40955631399317405}, 'person': {'precision': 0.8026315789473685, 'recall': 0.8776978417266187, 'f_score': 0.8384879725085911}, 'mood': {'precision': 0.5185185185185185, 'recall': 0.4307692307692308, 'f_score': 0.47058823529411764}, 'drug': {'precision': 0.7952755905511811, 'recall': 0.9126506024096386, 'f_score': 0.849929873772791}, 'pregnancy_considerations': {'precision': 0.2608695652173913, 'recall': 0.42857142857142855, 'f_score': 0.3243243243243243}, 'device': {'precision': 0.59375, 'recall': 0.7037037037037037, 'f_score': 0.6440677966101694}}}}\"]"
+      ]
+     },
+     "execution_count": 98,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# bert model case\n",
+    "with open('../data/chia_bert_log 2.txt') as fd:\n",
+    "    log = fd.read()\n",
+    "\n",
+    "lines = log.split('\\n')\n",
+    "eval_lines = [line for line in lines if 'full evaluation metrix' in line]\n",
+    "\n",
+    "eval_lines"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 99,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "\"        full evaluation metrix: {'overall': {'acc': 0.8155, 'strict': {'precision': 0.6268788682581786, 'recall': 0.7022119511389898, 'f_score': 0.6624104640298973}, 'relax': {'precision': 0.7606837606837606, 'recall': 0.8520964014526247, 'f_score': 0.8037994394269697}}, 'category': {'strict': {'condition': {'precision': 0.6882690730106645, 'recall': 0.7704315886134068, 'f_score': 0.7270363951473136}, 'temporal': {'precision': 0.5223097112860893, 'recall': 0.6546052631578947, 'f_score': 0.5810218978102191}, 'procedure': {'precision': 0.5333333333333333, 'recall': 0.6775244299674267, 'f_score': 0.5968436154949784}, 'measurement': {'precision': 0.6423357664233577, 'recall': 0.6666666666666666, 'f_score': 0.654275092936803}, 'value': {'precision': 0.7126760563380282, 'recall': 0.7737003058103975, 'f_score': 0.7419354838709676}, 'observation': {'precision': 0.3333333333333333, 'recall': 0.2732919254658385, 'f_score': 0.3003412969283276}, 'person': {'precision': 0.7894736842105263, 'recall': 0.8633093525179856, 'f_score': 0.8247422680412372}, 'mood': {'precision': 0.25925925925925924, 'recall': 0.2153846153846154, 'f_score': 0.23529411764705882}, 'drug': {'precision': 0.6902887139107612, 'recall': 0.7921686746987951, 'f_score': 0.7377279102384292}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.34375, 'recall': 0.4074074074074074, 'f_score': 0.37288135593220334}}, 'relax': {'condition': {'precision': 0.8252666119770303, 'recall': 0.9237832874196511, 'f_score': 0.8717504332755632}, 'temporal': {'precision': 0.6745406824146981, 'recall': 0.8453947368421053, 'f_score': 0.7503649635036497}, 'procedure': {'precision': 0.6666666666666666, 'recall': 0.8469055374592834, 'f_score': 0.7460545193687231}, 'measurement': {'precision': 0.8321167883211679, 'recall': 0.8636363636363636, 'f_score': 0.8475836431226766}, 'value': {'precision': 0.8225352112676056, 'recall': 0.8929663608562691, 'f_score': 0.8563049853372435}, 'observation': {'precision': 0.45454545454545453, 'recall': 0.37267080745341613, 'f_score': 0.40955631399317405}, 'person': {'precision': 0.8026315789473685, 'recall': 0.8776978417266187, 'f_score': 0.8384879725085911}, 'mood': {'precision': 0.5185185185185185, 'recall': 0.4307692307692308, 'f_score': 0.47058823529411764}, 'drug': {'precision': 0.7952755905511811, 'recall': 0.9126506024096386, 'f_score': 0.849929873772791}, 'pregnancy_considerations': {'precision': 0.2608695652173913, 'recall': 0.42857142857142855, 'f_score': 0.3243243243243243}, 'device': {'precision': 0.59375, 'recall': 0.7037037037037037, 'f_score': 0.6440677966101694}}}}\""
+      ]
+     },
+     "execution_count": 99,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "last_eval = eval_lines[-1]\n",
+    "last_eval"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 100,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "\"{'overall': {'acc': 0.8155, 'strict': {'precision': 0.6268788682581786, 'recall': 0.7022119511389898, 'f_score': 0.6624104640298973}, 'relax': {'precision': 0.7606837606837606, 'recall': 0.8520964014526247, 'f_score': 0.8037994394269697}}, 'category': {'strict': {'condition': {'precision': 0.6882690730106645, 'recall': 0.7704315886134068, 'f_score': 0.7270363951473136}, 'temporal': {'precision': 0.5223097112860893, 'recall': 0.6546052631578947, 'f_score': 0.5810218978102191}, 'procedure': {'precision': 0.5333333333333333, 'recall': 0.6775244299674267, 'f_score': 0.5968436154949784}, 'measurement': {'precision': 0.6423357664233577, 'recall': 0.6666666666666666, 'f_score': 0.654275092936803}, 'value': {'precision': 0.7126760563380282, 'recall': 0.7737003058103975, 'f_score': 0.7419354838709676}, 'observation': {'precision': 0.3333333333333333, 'recall': 0.2732919254658385, 'f_score': 0.3003412969283276}, 'person': {'precision': 0.7894736842105263, 'recall': 0.8633093525179856, 'f_score': 0.8247422680412372}, 'mood': {'precision': 0.25925925925925924, 'recall': 0.2153846153846154, 'f_score': 0.23529411764705882}, 'drug': {'precision': 0.6902887139107612, 'recall': 0.7921686746987951, 'f_score': 0.7377279102384292}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.34375, 'recall': 0.4074074074074074, 'f_score': 0.37288135593220334}}, 'relax': {'condition': {'precision': 0.8252666119770303, 'recall': 0.9237832874196511, 'f_score': 0.8717504332755632}, 'temporal': {'precision': 0.6745406824146981, 'recall': 0.8453947368421053, 'f_score': 0.7503649635036497}, 'procedure': {'precision': 0.6666666666666666, 'recall': 0.8469055374592834, 'f_score': 0.7460545193687231}, 'measurement': {'precision': 0.8321167883211679, 'recall': 0.8636363636363636, 'f_score': 0.8475836431226766}, 'value': {'precision': 0.8225352112676056, 'recall': 0.8929663608562691, 'f_score': 0.8563049853372435}, 'observation': {'precision': 0.45454545454545453, 'recall': 0.37267080745341613, 'f_score': 0.40955631399317405}, 'person': {'precision': 0.8026315789473685, 'recall': 0.8776978417266187, 'f_score': 0.8384879725085911}, 'mood': {'precision': 0.5185185185185185, 'recall': 0.4307692307692308, 'f_score': 0.47058823529411764}, 'drug': {'precision': 0.7952755905511811, 'recall': 0.9126506024096386, 'f_score': 0.849929873772791}, 'pregnancy_considerations': {'precision': 0.2608695652173913, 'recall': 0.42857142857142855, 'f_score': 0.3243243243243243}, 'device': {'precision': 0.59375, 'recall': 0.7037037037037037, 'f_score': 0.6440677966101694}}}}\""
+      ]
+     },
+     "execution_count": 100,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "eval = last_eval.split('full evaluation metrix: ')[1]\n",
+    "eval"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 101,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import json\n",
+    "with open('../data/chia_bert_eval.json', 'w') as fd:\n",
+    "    json.dump(eval, fd, indent=4)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 137,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "with open('../data/chia_bert_eval.json', 'r') as fd:\n",
+    "    eval_dict = json.load(fd)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 138,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'overall': {'acc': 0.8155, 'strict': {'precision': 0.6268788682581786, 'recall': 0.7022119511389898, 'f_score': 0.6624104640298973}, 'relax': {'precision': 0.7606837606837606, 'recall': 0.8520964014526247, 'f_score': 0.8037994394269697}}, 'category': {'strict': {'condition': {'precision': 0.6882690730106645, 'recall': 0.7704315886134068, 'f_score': 0.7270363951473136}, 'temporal': {'precision': 0.5223097112860893, 'recall': 0.6546052631578947, 'f_score': 0.5810218978102191}, 'procedure': {'precision': 0.5333333333333333, 'recall': 0.6775244299674267, 'f_score': 0.5968436154949784}, 'measurement': {'precision': 0.6423357664233577, 'recall': 0.6666666666666666, 'f_score': 0.654275092936803}, 'value': {'precision': 0.7126760563380282, 'recall': 0.7737003058103975, 'f_score': 0.7419354838709676}, 'observation': {'precision': 0.3333333333333333, 'recall': 0.2732919254658385, 'f_score': 0.3003412969283276}, 'person': {'precision': 0.7894736842105263, 'recall': 0.8633093525179856, 'f_score': 0.8247422680412372}, 'mood': {'precision': 0.25925925925925924, 'recall': 0.2153846153846154, 'f_score': 0.23529411764705882}, 'drug': {'precision': 0.6902887139107612, 'recall': 0.7921686746987951, 'f_score': 0.7377279102384292}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.34375, 'recall': 0.4074074074074074, 'f_score': 0.37288135593220334}}, 'relax': {'condition': {'precision': 0.8252666119770303, 'recall': 0.9237832874196511, 'f_score': 0.8717504332755632}, 'temporal': {'precision': 0.6745406824146981, 'recall': 0.8453947368421053, 'f_score': 0.7503649635036497}, 'procedure': {'precision': 0.6666666666666666, 'recall': 0.8469055374592834, 'f_score': 0.7460545193687231}, 'measurement': {'precision': 0.8321167883211679, 'recall': 0.8636363636363636, 'f_score': 0.8475836431226766}, 'value': {'precision': 0.8225352112676056, 'recall': 0.8929663608562691, 'f_score': 0.8563049853372435}, 'observation': {'precision': 0.45454545454545453, 'recall': 0.37267080745341613, 'f_score': 0.40955631399317405}, 'person': {'precision': 0.8026315789473685, 'recall': 0.8776978417266187, 'f_score': 0.8384879725085911}, 'mood': {'precision': 0.5185185185185185, 'recall': 0.4307692307692308, 'f_score': 0.47058823529411764}, 'drug': {'precision': 0.7952755905511811, 'recall': 0.9126506024096386, 'f_score': 0.849929873772791}, 'pregnancy_considerations': {'precision': 0.2608695652173913, 'recall': 0.42857142857142855, 'f_score': 0.3243243243243243}, 'device': {'precision': 0.59375, 'recall': 0.7037037037037037, 'f_score': 0.6440677966101694}}}}\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(eval_dict)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 108,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "eval_df = pd.DataFrame(columns=['type', 'prec_strict', 'rec_strict', 'f1_strict', 'prec_relaxed', 'rec_relaxed', 'f1_relaxed'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 109,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "['Overall',\n",
+       " 'condition',\n",
+       " 'temporal',\n",
+       " 'procedure',\n",
+       " 'measurement',\n",
+       " 'value',\n",
+       " 'observation',\n",
+       " 'person',\n",
+       " 'mood',\n",
+       " 'drug',\n",
+       " 'pregnancy_considerations',\n",
+       " 'device']"
+      ]
+     },
+     "execution_count": 109,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "types = ['Overall']\n",
+    "for keys in eval_dict['category']['strict']:\n",
+    "    types.append(keys)\n",
+    "types"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 110,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "eval_df['type'] = types"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 111,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pre_strict = [eval_dict['overall']['strict']['precision']]\n",
+    "for keys in eval_dict['category']['strict']:\n",
+    "    pre_strict.append(eval_dict['category']['strict'][keys]['precision'])\n",
+    "eval_df['prec_strict'] = pre_strict"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 112,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "rec_strict = [eval_dict['overall']['strict']['recall']]\n",
+    "for keys in eval_dict['category']['strict']:\n",
+    "    rec_strict.append(eval_dict['category']['strict'][keys]['recall'])\n",
+    "eval_df['rec_strict'] = rec_strict"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 113,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "f1_strict = [eval_dict['overall']['strict']['f_score']]\n",
+    "for keys in eval_dict['category']['strict']:\n",
+    "    f1_strict.append(eval_dict['category']['strict'][keys]['f_score'])\n",
+    "eval_df['f1_strict'] = f1_strict"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 114,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pre_relaxed = [eval_dict['overall']['relax']['precision']]\n",
+    "for keys in eval_dict['category']['relax']:\n",
+    "    pre_relaxed.append(eval_dict['category']['relax'][keys]['precision'])\n",
+    "eval_df['prec_relaxed'] = pre_relaxed"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 115,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "rec_relaxed = [eval_dict['overall']['relax']['recall']]\n",
+    "for keys in eval_dict['category']['relax']:\n",
+    "    rec_relaxed.append(eval_dict['category']['relax'][keys]['recall'])\n",
+    "eval_df['rec_relaxed'] = rec_relaxed"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 116,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "f1_relaxed = [eval_dict['overall']['relax']['f_score']]\n",
+    "for keys in eval_dict['category']['relax']:\n",
+    "    f1_relaxed.append(eval_dict['category']['relax'][keys]['f_score'])\n",
+    "eval_df['f1_relaxed'] = f1_relaxed"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 117,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>type</th>\n",
+       "      <th>prec_strict</th>\n",
+       "      <th>rec_strict</th>\n",
+       "      <th>f1_strict</th>\n",
+       "      <th>prec_relaxed</th>\n",
+       "      <th>rec_relaxed</th>\n",
+       "      <th>f1_relaxed</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Overall</td>\n",
+       "      <td>0.626879</td>\n",
+       "      <td>0.702212</td>\n",
+       "      <td>0.662410</td>\n",
+       "      <td>0.760684</td>\n",
+       "      <td>0.852096</td>\n",
+       "      <td>0.803799</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>condition</td>\n",
+       "      <td>0.688269</td>\n",
+       "      <td>0.770432</td>\n",
+       "      <td>0.727036</td>\n",
+       "      <td>0.825267</td>\n",
+       "      <td>0.923783</td>\n",
+       "      <td>0.871750</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>temporal</td>\n",
+       "      <td>0.522310</td>\n",
+       "      <td>0.654605</td>\n",
+       "      <td>0.581022</td>\n",
+       "      <td>0.674541</td>\n",
+       "      <td>0.845395</td>\n",
+       "      <td>0.750365</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>procedure</td>\n",
+       "      <td>0.533333</td>\n",
+       "      <td>0.677524</td>\n",
+       "      <td>0.596844</td>\n",
+       "      <td>0.666667</td>\n",
+       "      <td>0.846906</td>\n",
+       "      <td>0.746055</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>measurement</td>\n",
+       "      <td>0.642336</td>\n",
+       "      <td>0.666667</td>\n",
+       "      <td>0.654275</td>\n",
+       "      <td>0.832117</td>\n",
+       "      <td>0.863636</td>\n",
+       "      <td>0.847584</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>value</td>\n",
+       "      <td>0.712676</td>\n",
+       "      <td>0.773700</td>\n",
+       "      <td>0.741935</td>\n",
+       "      <td>0.822535</td>\n",
+       "      <td>0.892966</td>\n",
+       "      <td>0.856305</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>observation</td>\n",
+       "      <td>0.333333</td>\n",
+       "      <td>0.273292</td>\n",
+       "      <td>0.300341</td>\n",
+       "      <td>0.454545</td>\n",
+       "      <td>0.372671</td>\n",
+       "      <td>0.409556</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>person</td>\n",
+       "      <td>0.789474</td>\n",
+       "      <td>0.863309</td>\n",
+       "      <td>0.824742</td>\n",
+       "      <td>0.802632</td>\n",
+       "      <td>0.877698</td>\n",
+       "      <td>0.838488</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>mood</td>\n",
+       "      <td>0.259259</td>\n",
+       "      <td>0.215385</td>\n",
+       "      <td>0.235294</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.430769</td>\n",
+       "      <td>0.470588</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>drug</td>\n",
+       "      <td>0.690289</td>\n",
+       "      <td>0.792169</td>\n",
+       "      <td>0.737728</td>\n",
+       "      <td>0.795276</td>\n",
+       "      <td>0.912651</td>\n",
+       "      <td>0.849930</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>pregnancy_considerations</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.260870</td>\n",
+       "      <td>0.428571</td>\n",
+       "      <td>0.324324</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>device</td>\n",
+       "      <td>0.343750</td>\n",
+       "      <td>0.407407</td>\n",
+       "      <td>0.372881</td>\n",
+       "      <td>0.593750</td>\n",
+       "      <td>0.703704</td>\n",
+       "      <td>0.644068</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                        type  prec_strict  rec_strict  f1_strict  \\\n",
+       "0                    Overall     0.626879    0.702212   0.662410   \n",
+       "1                  condition     0.688269    0.770432   0.727036   \n",
+       "2                   temporal     0.522310    0.654605   0.581022   \n",
+       "3                  procedure     0.533333    0.677524   0.596844   \n",
+       "4                measurement     0.642336    0.666667   0.654275   \n",
+       "5                      value     0.712676    0.773700   0.741935   \n",
+       "6                observation     0.333333    0.273292   0.300341   \n",
+       "7                     person     0.789474    0.863309   0.824742   \n",
+       "8                       mood     0.259259    0.215385   0.235294   \n",
+       "9                       drug     0.690289    0.792169   0.737728   \n",
+       "10  pregnancy_considerations     0.000000    0.000000   0.000000   \n",
+       "11                    device     0.343750    0.407407   0.372881   \n",
+       "\n",
+       "    prec_relaxed  rec_relaxed  f1_relaxed  \n",
+       "0       0.760684     0.852096    0.803799  \n",
+       "1       0.825267     0.923783    0.871750  \n",
+       "2       0.674541     0.845395    0.750365  \n",
+       "3       0.666667     0.846906    0.746055  \n",
+       "4       0.832117     0.863636    0.847584  \n",
+       "5       0.822535     0.892966    0.856305  \n",
+       "6       0.454545     0.372671    0.409556  \n",
+       "7       0.802632     0.877698    0.838488  \n",
+       "8       0.518519     0.430769    0.470588  \n",
+       "9       0.795276     0.912651    0.849930  \n",
+       "10      0.260870     0.428571    0.324324  \n",
+       "11      0.593750     0.703704    0.644068  "
+      ]
+     },
+     "execution_count": 117,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "eval_df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 118,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>type</th>\n",
+       "      <th>prec_strict</th>\n",
+       "      <th>rec_strict</th>\n",
+       "      <th>f1_strict</th>\n",
+       "      <th>prec_relaxed</th>\n",
+       "      <th>rec_relaxed</th>\n",
+       "      <th>f1_relaxed</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>person</td>\n",
+       "      <td>0.789474</td>\n",
+       "      <td>0.863309</td>\n",
+       "      <td>0.824742</td>\n",
+       "      <td>0.802632</td>\n",
+       "      <td>0.877698</td>\n",
+       "      <td>0.838488</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>value</td>\n",
+       "      <td>0.712676</td>\n",
+       "      <td>0.773700</td>\n",
+       "      <td>0.741935</td>\n",
+       "      <td>0.822535</td>\n",
+       "      <td>0.892966</td>\n",
+       "      <td>0.856305</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>drug</td>\n",
+       "      <td>0.690289</td>\n",
+       "      <td>0.792169</td>\n",
+       "      <td>0.737728</td>\n",
+       "      <td>0.795276</td>\n",
+       "      <td>0.912651</td>\n",
+       "      <td>0.849930</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>condition</td>\n",
+       "      <td>0.688269</td>\n",
+       "      <td>0.770432</td>\n",
+       "      <td>0.727036</td>\n",
+       "      <td>0.825267</td>\n",
+       "      <td>0.923783</td>\n",
+       "      <td>0.871750</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>measurement</td>\n",
+       "      <td>0.642336</td>\n",
+       "      <td>0.666667</td>\n",
+       "      <td>0.654275</td>\n",
+       "      <td>0.832117</td>\n",
+       "      <td>0.863636</td>\n",
+       "      <td>0.847584</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>procedure</td>\n",
+       "      <td>0.533333</td>\n",
+       "      <td>0.677524</td>\n",
+       "      <td>0.596844</td>\n",
+       "      <td>0.666667</td>\n",
+       "      <td>0.846906</td>\n",
+       "      <td>0.746055</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>temporal</td>\n",
+       "      <td>0.522310</td>\n",
+       "      <td>0.654605</td>\n",
+       "      <td>0.581022</td>\n",
+       "      <td>0.674541</td>\n",
+       "      <td>0.845395</td>\n",
+       "      <td>0.750365</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>device</td>\n",
+       "      <td>0.343750</td>\n",
+       "      <td>0.407407</td>\n",
+       "      <td>0.372881</td>\n",
+       "      <td>0.593750</td>\n",
+       "      <td>0.703704</td>\n",
+       "      <td>0.644068</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>observation</td>\n",
+       "      <td>0.333333</td>\n",
+       "      <td>0.273292</td>\n",
+       "      <td>0.300341</td>\n",
+       "      <td>0.454545</td>\n",
+       "      <td>0.372671</td>\n",
+       "      <td>0.409556</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>mood</td>\n",
+       "      <td>0.259259</td>\n",
+       "      <td>0.215385</td>\n",
+       "      <td>0.235294</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.430769</td>\n",
+       "      <td>0.470588</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>pregnancy_considerations</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.260870</td>\n",
+       "      <td>0.428571</td>\n",
+       "      <td>0.324324</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                        type  prec_strict  rec_strict  f1_strict  \\\n",
+       "0                     person     0.789474    0.863309   0.824742   \n",
+       "1                      value     0.712676    0.773700   0.741935   \n",
+       "2                       drug     0.690289    0.792169   0.737728   \n",
+       "3                  condition     0.688269    0.770432   0.727036   \n",
+       "4                measurement     0.642336    0.666667   0.654275   \n",
+       "5                  procedure     0.533333    0.677524   0.596844   \n",
+       "6                   temporal     0.522310    0.654605   0.581022   \n",
+       "7                     device     0.343750    0.407407   0.372881   \n",
+       "8                observation     0.333333    0.273292   0.300341   \n",
+       "9                       mood     0.259259    0.215385   0.235294   \n",
+       "10  pregnancy_considerations     0.000000    0.000000   0.000000   \n",
+       "\n",
+       "    prec_relaxed  rec_relaxed  f1_relaxed  \n",
+       "0       0.802632     0.877698    0.838488  \n",
+       "1       0.822535     0.892966    0.856305  \n",
+       "2       0.795276     0.912651    0.849930  \n",
+       "3       0.825267     0.923783    0.871750  \n",
+       "4       0.832117     0.863636    0.847584  \n",
+       "5       0.666667     0.846906    0.746055  \n",
+       "6       0.674541     0.845395    0.750365  \n",
+       "7       0.593750     0.703704    0.644068  \n",
+       "8       0.454545     0.372671    0.409556  \n",
+       "9       0.518519     0.430769    0.470588  \n",
+       "10      0.260870     0.428571    0.324324  "
+      ]
+     },
+     "execution_count": 118,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "eval_df[1:].sort_values(by=['f1_strict'], ascending=False, ignore_index=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 119,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "import matplotlib.pyplot as plt\n",
+    "eval_df.plot(x='type', y=['f1_strict', 'f1_relaxed'], title= f'Strict and Relax F1_Score for Bert base model on Chia', kind=\"bar\", rot=90)\n",
+    "plt.xlabel(\"\")\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 139,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[\"        full evaluation metrix: {'overall': {'acc': 0.7955, 'strict': {'precision': 0.5658914728682171, 'recall': 0.6266094420600858, 'f_score': 0.594704684317719}, 'relax': {'precision': 0.7185450208706022, 'recall': 0.7956421261142291, 'f_score': 0.7551308162306125}}, 'category': {'strict': {'condition': {'precision': 0.5995575221238938, 'recall': 0.7465564738292011, 'f_score': 0.6650306748466257}, 'drug': {'precision': 0.5701559020044543, 'recall': 0.7710843373493976, 'f_score': 0.6555697823303457}, 'procedure': {'precision': 0.4103194103194103, 'recall': 0.5439739413680782, 'f_score': 0.4677871148459384}, 'temporal': {'precision': 0.47619047619047616, 'recall': 0.4605263157894737, 'f_score': 0.46822742474916385}, 'measurement': {'precision': 0.5236486486486487, 'recall': 0.5871212121212122, 'f_score': 0.5535714285714286}, 'value': {'precision': 0.6628895184135978, 'recall': 0.7155963302752294, 'f_score': 0.6882352941176472}, 'observation': {'precision': 0.5652173913043478, 'recall': 0.2422360248447205, 'f_score': 0.3391304347826087}, 'person': {'precision': 0.7642276422764228, 'recall': 0.6762589928057554, 'f_score': 0.7175572519083969}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}}, 'relax': {'condition': {'precision': 0.7463126843657817, 'recall': 0.9292929292929293, 'f_score': 0.8278118609406953}, 'drug': {'precision': 0.6859688195991092, 'recall': 0.927710843373494, 'f_score': 0.7887323943661971}, 'procedure': {'precision': 0.5847665847665847, 'recall': 0.7752442996742671, 'f_score': 0.6666666666666666}, 'temporal': {'precision': 0.6768707482993197, 'recall': 0.6546052631578947, 'f_score': 0.6655518394648829}, 'measurement': {'precision': 0.75, 'recall': 0.8409090909090909, 'f_score': 0.7928571428571429}, 'value': {'precision': 0.8101983002832861, 'recall': 0.8746177370030581, 'f_score': 0.8411764705882353}, 'observation': {'precision': 0.6811594202898551, 'recall': 0.2919254658385093, 'f_score': 0.408695652173913}, 'person': {'precision': 0.7723577235772358, 'recall': 0.6834532374100719, 'f_score': 0.7251908396946565}, 'pregnancy_considerations': {'precision': 0.42857142857142855, 'recall': 0.21428571428571427, 'f_score': 0.2857142857142857}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.7878, 'strict': {'precision': 0.5638719915366305, 'recall': 0.703862660944206, 'f_score': 0.6261380323054332}, 'relax': {'precision': 0.6857974080930971, 'recall': 0.8560581049851436, 'f_score': 0.7615271659324522}}, 'category': {'strict': {'condition': {'precision': 0.5894085281980743, 'recall': 0.7869605142332415, 'f_score': 0.6740070782540307}, 'temporal': {'precision': 0.4962962962962963, 'recall': 0.6611842105263158, 'f_score': 0.5669957686882934}, 'procedure': {'precision': 0.4050387596899225, 'recall': 0.6807817589576547, 'f_score': 0.5078979343863912}, 'measurement': {'precision': 0.584717607973422, 'recall': 0.6666666666666666, 'f_score': 0.6230088495575221}, 'value': {'precision': 0.6963788300835655, 'recall': 0.764525993883792, 'f_score': 0.7288629737609329}, 'observation': {'precision': 0.366412213740458, 'recall': 0.2981366459627329, 'f_score': 0.3287671232876712}, 'person': {'precision': 0.7349397590361446, 'recall': 0.8776978417266187, 'f_score': 0.7999999999999999}, 'drug': {'precision': 0.6397058823529411, 'recall': 0.786144578313253, 'f_score': 0.7054054054054053}, 'device': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'mood': {'precision': 0.25, 'recall': 0.12307692307692308, 'f_score': 0.16494845360824742}}, 'relax': {'condition': {'precision': 0.7070151306740028, 'recall': 0.9439853076216712, 'f_score': 0.8084939048368069}, 'temporal': {'precision': 0.6320987654320988, 'recall': 0.8421052631578947, 'f_score': 0.7221438645980254}, 'procedure': {'precision': 0.5193798449612403, 'recall': 0.8729641693811075, 'f_score': 0.6512758201701094}, 'measurement': {'precision': 0.7574750830564784, 'recall': 0.8636363636363636, 'f_score': 0.8070796460176991}, 'value': {'precision': 0.8189415041782729, 'recall': 0.8990825688073395, 'f_score': 0.8571428571428572}, 'observation': {'precision': 0.48091603053435117, 'recall': 0.391304347826087, 'f_score': 0.43150684931506855}, 'person': {'precision': 0.7409638554216867, 'recall': 0.8848920863309353, 'f_score': 0.8065573770491803}, 'drug': {'precision': 0.7573529411764706, 'recall': 0.9307228915662651, 'f_score': 0.8351351351351352}, 'device': {'precision': 0.7777777777777778, 'recall': 0.25925925925925924, 'f_score': 0.3888888888888889}, 'mood': {'precision': 0.53125, 'recall': 0.26153846153846155, 'f_score': 0.3505154639175258}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.814, 'strict': {'precision': 0.6220472440944882, 'recall': 0.7041928029052492, 'f_score': 0.6605760297305667}, 'relax': {'precision': 0.7430737824438611, 'recall': 0.8412017167381974, 'f_score': 0.7890987921957261}}, 'category': {'strict': {'condition': {'precision': 0.6640986132511556, 'recall': 0.7915518824609734, 'f_score': 0.7222454964390448}, 'temporal': {'precision': 0.5188916876574308, 'recall': 0.6776315789473685, 'f_score': 0.587731811697575}, 'procedure': {'precision': 0.4678111587982833, 'recall': 0.7100977198697068, 'f_score': 0.5640362225097024}, 'measurement': {'precision': 0.6589147286821705, 'recall': 0.6439393939393939, 'f_score': 0.6513409961685823}, 'value': {'precision': 0.7697160883280757, 'recall': 0.746177370030581, 'f_score': 0.7577639751552795}, 'observation': {'precision': 0.5757575757575758, 'recall': 0.2360248447204969, 'f_score': 0.33480176211453744}, 'person': {'precision': 0.8297872340425532, 'recall': 0.841726618705036, 'f_score': 0.8357142857142857}, 'drug': {'precision': 0.7280701754385965, 'recall': 0.75, 'f_score': 0.7388724035608308}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.3939393939393939, 'recall': 0.48148148148148145, 'f_score': 0.43333333333333324}, 'mood': {'precision': 0.2807017543859649, 'recall': 0.24615384615384617, 'f_score': 0.26229508196721313}}, 'relax': {'condition': {'precision': 0.7850539291217258, 'recall': 0.9357208448117539, 'f_score': 0.8537913699204022}, 'temporal': {'precision': 0.6599496221662469, 'recall': 0.8618421052631579, 'f_score': 0.7475035663338089}, 'procedure': {'precision': 0.5772532188841202, 'recall': 0.8762214983713354, 'f_score': 0.6959896507115135}, 'measurement': {'precision': 0.8294573643410853, 'recall': 0.8106060606060606, 'f_score': 0.8199233716475095}, 'value': {'precision': 0.886435331230284, 'recall': 0.8593272171253823, 'f_score': 0.8726708074534162}, 'observation': {'precision': 0.6818181818181818, 'recall': 0.2795031055900621, 'f_score': 0.3964757709251101}, 'person': {'precision': 0.8368794326241135, 'recall': 0.8489208633093526, 'f_score': 0.8428571428571429}, 'drug': {'precision': 0.8245614035087719, 'recall': 0.8493975903614458, 'f_score': 0.8367952522255192}, 'pregnancy_considerations': {'precision': 0.16666666666666666, 'recall': 0.6428571428571429, 'f_score': 0.2647058823529412}, 'device': {'precision': 0.5454545454545454, 'recall': 0.6666666666666666, 'f_score': 0.6}, 'mood': {'precision': 0.543859649122807, 'recall': 0.47692307692307695, 'f_score': 0.5081967213114754}}}}\",\n",
+       " \"        full evaluation metrix: {'overall': {'acc': 0.8285, 'strict': {'precision': 0.6570648878107944, 'recall': 0.7154176295807197, 'f_score': 0.6850007902639481}, 'relax': {'precision': 0.7786537295330503, 'recall': 0.8478045559590623, 'f_score': 0.8117591275486012}}, 'category': {'strict': {'condition': {'precision': 0.7032418952618454, 'recall': 0.7768595041322314, 'f_score': 0.7382198952879582}, 'temporal': {'precision': 0.5580110497237569, 'recall': 0.6644736842105263, 'f_score': 0.6066066066066066}, 'procedure': {'precision': 0.5608465608465608, 'recall': 0.6905537459283387, 'f_score': 0.618978102189781}, 'measurement': {'precision': 0.6762589928057554, 'recall': 0.7121212121212122, 'f_score': 0.6937269372693727}, 'value': {'precision': 0.7159420289855073, 'recall': 0.7553516819571865, 'f_score': 0.7351190476190476}, 'observation': {'precision': 0.5280898876404494, 'recall': 0.2919254658385093, 'f_score': 0.37599999999999995}, 'person': {'precision': 0.7469135802469136, 'recall': 0.8705035971223022, 'f_score': 0.8039867109634551}, 'drug': {'precision': 0.7176165803108808, 'recall': 0.8343373493975904, 'f_score': 0.7715877437325905}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.5185185185185185, 'recall': 0.5185185185185185, 'f_score': 0.5185185185185185}, 'mood': {'precision': 0.24074074074074073, 'recall': 0.2, 'f_score': 0.2184873949579832}}, 'relax': {'condition': {'precision': 0.8337489609310058, 'recall': 0.921028466483012, 'f_score': 0.87521815008726}, 'temporal': {'precision': 0.6933701657458563, 'recall': 0.8256578947368421, 'f_score': 0.7537537537537538}, 'procedure': {'precision': 0.6772486772486772, 'recall': 0.8338762214983714, 'f_score': 0.7474452554744526}, 'measurement': {'precision': 0.8057553956834532, 'recall': 0.8484848484848485, 'f_score': 0.8265682656826568}, 'value': {'precision': 0.8521739130434782, 'recall': 0.8990825688073395, 'f_score': 0.8750000000000001}, 'observation': {'precision': 0.6067415730337079, 'recall': 0.33540372670807456, 'f_score': 0.432}, 'person': {'precision': 0.7592592592592593, 'recall': 0.8848920863309353, 'f_score': 0.8172757475083057}, 'drug': {'precision': 0.8031088082901554, 'recall': 0.9337349397590361, 'f_score': 0.8635097493036212}, 'pregnancy_considerations': {'precision': 0.42857142857142855, 'recall': 0.42857142857142855, 'f_score': 0.42857142857142855}, 'device': {'precision': 0.7037037037037037, 'recall': 0.7037037037037037, 'f_score': 0.7037037037037037}, 'mood': {'precision': 0.5185185185185185, 'recall': 0.4307692307692308, 'f_score': 0.47058823529411764}}}}\"]"
+      ]
+     },
+     "execution_count": 139,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Roberta model case\n",
+    "with open('../data/chia_roberta_log.txt') as fd:\n",
+    "    log = fd.read()\n",
+    "\n",
+    "lines = log.split('\\n')\n",
+    "eval_lines = [line for line in lines if 'full evaluation metrix' in line]\n",
+    "\n",
+    "eval_lines"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 140,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "\"        full evaluation metrix: {'overall': {'acc': 0.8285, 'strict': {'precision': 0.6570648878107944, 'recall': 0.7154176295807197, 'f_score': 0.6850007902639481}, 'relax': {'precision': 0.7786537295330503, 'recall': 0.8478045559590623, 'f_score': 0.8117591275486012}}, 'category': {'strict': {'condition': {'precision': 0.7032418952618454, 'recall': 0.7768595041322314, 'f_score': 0.7382198952879582}, 'temporal': {'precision': 0.5580110497237569, 'recall': 0.6644736842105263, 'f_score': 0.6066066066066066}, 'procedure': {'precision': 0.5608465608465608, 'recall': 0.6905537459283387, 'f_score': 0.618978102189781}, 'measurement': {'precision': 0.6762589928057554, 'recall': 0.7121212121212122, 'f_score': 0.6937269372693727}, 'value': {'precision': 0.7159420289855073, 'recall': 0.7553516819571865, 'f_score': 0.7351190476190476}, 'observation': {'precision': 0.5280898876404494, 'recall': 0.2919254658385093, 'f_score': 0.37599999999999995}, 'person': {'precision': 0.7469135802469136, 'recall': 0.8705035971223022, 'f_score': 0.8039867109634551}, 'drug': {'precision': 0.7176165803108808, 'recall': 0.8343373493975904, 'f_score': 0.7715877437325905}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.5185185185185185, 'recall': 0.5185185185185185, 'f_score': 0.5185185185185185}, 'mood': {'precision': 0.24074074074074073, 'recall': 0.2, 'f_score': 0.2184873949579832}}, 'relax': {'condition': {'precision': 0.8337489609310058, 'recall': 0.921028466483012, 'f_score': 0.87521815008726}, 'temporal': {'precision': 0.6933701657458563, 'recall': 0.8256578947368421, 'f_score': 0.7537537537537538}, 'procedure': {'precision': 0.6772486772486772, 'recall': 0.8338762214983714, 'f_score': 0.7474452554744526}, 'measurement': {'precision': 0.8057553956834532, 'recall': 0.8484848484848485, 'f_score': 0.8265682656826568}, 'value': {'precision': 0.8521739130434782, 'recall': 0.8990825688073395, 'f_score': 0.8750000000000001}, 'observation': {'precision': 0.6067415730337079, 'recall': 0.33540372670807456, 'f_score': 0.432}, 'person': {'precision': 0.7592592592592593, 'recall': 0.8848920863309353, 'f_score': 0.8172757475083057}, 'drug': {'precision': 0.8031088082901554, 'recall': 0.9337349397590361, 'f_score': 0.8635097493036212}, 'pregnancy_considerations': {'precision': 0.42857142857142855, 'recall': 0.42857142857142855, 'f_score': 0.42857142857142855}, 'device': {'precision': 0.7037037037037037, 'recall': 0.7037037037037037, 'f_score': 0.7037037037037037}, 'mood': {'precision': 0.5185185185185185, 'recall': 0.4307692307692308, 'f_score': 0.47058823529411764}}}}\""
+      ]
+     },
+     "execution_count": 140,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "last_eval = eval_lines[-1]\n",
+    "last_eval"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 141,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "\"{'overall': {'acc': 0.8285, 'strict': {'precision': 0.6570648878107944, 'recall': 0.7154176295807197, 'f_score': 0.6850007902639481}, 'relax': {'precision': 0.7786537295330503, 'recall': 0.8478045559590623, 'f_score': 0.8117591275486012}}, 'category': {'strict': {'condition': {'precision': 0.7032418952618454, 'recall': 0.7768595041322314, 'f_score': 0.7382198952879582}, 'temporal': {'precision': 0.5580110497237569, 'recall': 0.6644736842105263, 'f_score': 0.6066066066066066}, 'procedure': {'precision': 0.5608465608465608, 'recall': 0.6905537459283387, 'f_score': 0.618978102189781}, 'measurement': {'precision': 0.6762589928057554, 'recall': 0.7121212121212122, 'f_score': 0.6937269372693727}, 'value': {'precision': 0.7159420289855073, 'recall': 0.7553516819571865, 'f_score': 0.7351190476190476}, 'observation': {'precision': 0.5280898876404494, 'recall': 0.2919254658385093, 'f_score': 0.37599999999999995}, 'person': {'precision': 0.7469135802469136, 'recall': 0.8705035971223022, 'f_score': 0.8039867109634551}, 'drug': {'precision': 0.7176165803108808, 'recall': 0.8343373493975904, 'f_score': 0.7715877437325905}, 'pregnancy_considerations': {'precision': 0.0, 'recall': 0.0, 'f_score': 0.0}, 'device': {'precision': 0.5185185185185185, 'recall': 0.5185185185185185, 'f_score': 0.5185185185185185}, 'mood': {'precision': 0.24074074074074073, 'recall': 0.2, 'f_score': 0.2184873949579832}}, 'relax': {'condition': {'precision': 0.8337489609310058, 'recall': 0.921028466483012, 'f_score': 0.87521815008726}, 'temporal': {'precision': 0.6933701657458563, 'recall': 0.8256578947368421, 'f_score': 0.7537537537537538}, 'procedure': {'precision': 0.6772486772486772, 'recall': 0.8338762214983714, 'f_score': 0.7474452554744526}, 'measurement': {'precision': 0.8057553956834532, 'recall': 0.8484848484848485, 'f_score': 0.8265682656826568}, 'value': {'precision': 0.8521739130434782, 'recall': 0.8990825688073395, 'f_score': 0.8750000000000001}, 'observation': {'precision': 0.6067415730337079, 'recall': 0.33540372670807456, 'f_score': 0.432}, 'person': {'precision': 0.7592592592592593, 'recall': 0.8848920863309353, 'f_score': 0.8172757475083057}, 'drug': {'precision': 0.8031088082901554, 'recall': 0.9337349397590361, 'f_score': 0.8635097493036212}, 'pregnancy_considerations': {'precision': 0.42857142857142855, 'recall': 0.42857142857142855, 'f_score': 0.42857142857142855}, 'device': {'precision': 0.7037037037037037, 'recall': 0.7037037037037037, 'f_score': 0.7037037037037037}, 'mood': {'precision': 0.5185185185185185, 'recall': 0.4307692307692308, 'f_score': 0.47058823529411764}}}}\""
+      ]
+     },
+     "execution_count": 141,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "eval_roberta = last_eval.split('full evaluation metrix: ')[1]\n",
+    "eval_roberta"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 155,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import json\n",
+    "with open('../data/chia_roberta_eval.json', 'w') as fd:\n",
+    "    json.dump(eval_roberta, fd, indent=4)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 156,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "with open('../data/chia_roberta_eval.json', 'r') as fd:\n",
+    "    eval_dict_roberta = json.load(fd)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 157,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "eval_df_roberta = pd.DataFrame(columns=['type', 'prec_strict', 'rec_strict', 'f1_strict', 'prec_relaxed', 'rec_relaxed', 'f1_relaxed'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 158,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "['Overall',\n",
+       " 'condition',\n",
+       " 'temporal',\n",
+       " 'procedure',\n",
+       " 'measurement',\n",
+       " 'value',\n",
+       " 'observation',\n",
+       " 'person',\n",
+       " 'drug',\n",
+       " 'pregnancy_considerations',\n",
+       " 'device',\n",
+       " 'mood']"
+      ]
+     },
+     "execution_count": 158,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "types_roberta = ['Overall']\n",
+    "for keys in eval_dict_roberta['category']['strict']:\n",
+    "    types_roberta.append(keys)\n",
+    "types_roberta"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 166,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "eval_df_roberta['type'] = types_roberta"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 159,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pre_strict = [eval_dict_roberta['overall']['strict']['precision']]\n",
+    "for keys in eval_dict_roberta['category']['strict']:\n",
+    "    pre_strict.append(eval_dict_roberta['category']['strict'][keys]['precision'])\n",
+    "eval_df_roberta['prec_strict'] = pre_strict"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 160,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "rec_strict = [eval_dict_roberta['overall']['strict']['recall']]\n",
+    "for keys in eval_dict_roberta['category']['strict']:\n",
+    "    rec_strict.append(eval_dict_roberta['category']['strict'][keys]['recall'])\n",
+    "eval_df_roberta['rec_strict'] = rec_strict"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 161,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "f1_strict = [eval_dict_roberta['overall']['strict']['f_score']]\n",
+    "for keys in eval_dict_roberta['category']['strict']:\n",
+    "    f1_strict.append(eval_dict_roberta['category']['strict'][keys]['f_score'])\n",
+    "eval_df_roberta['f1_strict'] = f1_strict"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 162,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pre_relaxed = [eval_dict_roberta['overall']['relax']['precision']]\n",
+    "for keys in eval_dict_roberta['category']['relax']:\n",
+    "    pre_relaxed.append(eval_dict_roberta['category']['relax'][keys]['precision'])\n",
+    "eval_df_roberta['prec_relaxed'] = pre_relaxed"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 163,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "rec_relaxed = [eval_dict_roberta['overall']['relax']['recall']]\n",
+    "for keys in eval_dict_roberta['category']['relax']:\n",
+    "    rec_relaxed.append(eval_dict_roberta['category']['relax'][keys]['recall'])\n",
+    "eval_df_roberta['rec_relaxed'] = rec_relaxed"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 164,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "f1_relaxed = [eval_dict_roberta['overall']['relax']['f_score']]\n",
+    "for keys in eval_dict_roberta['category']['relax']:\n",
+    "    f1_relaxed.append(eval_dict_roberta['category']['relax'][keys]['f_score'])\n",
+    "eval_df_roberta['f1_relaxed'] = f1_relaxed"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 167,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>type</th>\n",
+       "      <th>prec_strict</th>\n",
+       "      <th>rec_strict</th>\n",
+       "      <th>f1_strict</th>\n",
+       "      <th>prec_relaxed</th>\n",
+       "      <th>rec_relaxed</th>\n",
+       "      <th>f1_relaxed</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Overall</td>\n",
+       "      <td>0.657065</td>\n",
+       "      <td>0.715418</td>\n",
+       "      <td>0.685001</td>\n",
+       "      <td>0.778654</td>\n",
+       "      <td>0.847805</td>\n",
+       "      <td>0.811759</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>condition</td>\n",
+       "      <td>0.703242</td>\n",
+       "      <td>0.776860</td>\n",
+       "      <td>0.738220</td>\n",
+       "      <td>0.833749</td>\n",
+       "      <td>0.921028</td>\n",
+       "      <td>0.875218</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>temporal</td>\n",
+       "      <td>0.558011</td>\n",
+       "      <td>0.664474</td>\n",
+       "      <td>0.606607</td>\n",
+       "      <td>0.693370</td>\n",
+       "      <td>0.825658</td>\n",
+       "      <td>0.753754</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>procedure</td>\n",
+       "      <td>0.560847</td>\n",
+       "      <td>0.690554</td>\n",
+       "      <td>0.618978</td>\n",
+       "      <td>0.677249</td>\n",
+       "      <td>0.833876</td>\n",
+       "      <td>0.747445</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>measurement</td>\n",
+       "      <td>0.676259</td>\n",
+       "      <td>0.712121</td>\n",
+       "      <td>0.693727</td>\n",
+       "      <td>0.805755</td>\n",
+       "      <td>0.848485</td>\n",
+       "      <td>0.826568</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>value</td>\n",
+       "      <td>0.715942</td>\n",
+       "      <td>0.755352</td>\n",
+       "      <td>0.735119</td>\n",
+       "      <td>0.852174</td>\n",
+       "      <td>0.899083</td>\n",
+       "      <td>0.875000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>observation</td>\n",
+       "      <td>0.528090</td>\n",
+       "      <td>0.291925</td>\n",
+       "      <td>0.376000</td>\n",
+       "      <td>0.606742</td>\n",
+       "      <td>0.335404</td>\n",
+       "      <td>0.432000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>person</td>\n",
+       "      <td>0.746914</td>\n",
+       "      <td>0.870504</td>\n",
+       "      <td>0.803987</td>\n",
+       "      <td>0.759259</td>\n",
+       "      <td>0.884892</td>\n",
+       "      <td>0.817276</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>drug</td>\n",
+       "      <td>0.717617</td>\n",
+       "      <td>0.834337</td>\n",
+       "      <td>0.771588</td>\n",
+       "      <td>0.803109</td>\n",
+       "      <td>0.933735</td>\n",
+       "      <td>0.863510</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>pregnancy_considerations</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.428571</td>\n",
+       "      <td>0.428571</td>\n",
+       "      <td>0.428571</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>device</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.703704</td>\n",
+       "      <td>0.703704</td>\n",
+       "      <td>0.703704</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>mood</td>\n",
+       "      <td>0.240741</td>\n",
+       "      <td>0.200000</td>\n",
+       "      <td>0.218487</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.430769</td>\n",
+       "      <td>0.470588</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                        type  prec_strict  rec_strict  f1_strict  \\\n",
+       "0                    Overall     0.657065    0.715418   0.685001   \n",
+       "1                  condition     0.703242    0.776860   0.738220   \n",
+       "2                   temporal     0.558011    0.664474   0.606607   \n",
+       "3                  procedure     0.560847    0.690554   0.618978   \n",
+       "4                measurement     0.676259    0.712121   0.693727   \n",
+       "5                      value     0.715942    0.755352   0.735119   \n",
+       "6                observation     0.528090    0.291925   0.376000   \n",
+       "7                     person     0.746914    0.870504   0.803987   \n",
+       "8                       drug     0.717617    0.834337   0.771588   \n",
+       "9   pregnancy_considerations     0.000000    0.000000   0.000000   \n",
+       "10                    device     0.518519    0.518519   0.518519   \n",
+       "11                      mood     0.240741    0.200000   0.218487   \n",
+       "\n",
+       "    prec_relaxed  rec_relaxed  f1_relaxed  \n",
+       "0       0.778654     0.847805    0.811759  \n",
+       "1       0.833749     0.921028    0.875218  \n",
+       "2       0.693370     0.825658    0.753754  \n",
+       "3       0.677249     0.833876    0.747445  \n",
+       "4       0.805755     0.848485    0.826568  \n",
+       "5       0.852174     0.899083    0.875000  \n",
+       "6       0.606742     0.335404    0.432000  \n",
+       "7       0.759259     0.884892    0.817276  \n",
+       "8       0.803109     0.933735    0.863510  \n",
+       "9       0.428571     0.428571    0.428571  \n",
+       "10      0.703704     0.703704    0.703704  \n",
+       "11      0.518519     0.430769    0.470588  "
+      ]
+     },
+     "execution_count": 167,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "eval_df_roberta"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 168,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>type</th>\n",
+       "      <th>prec_strict</th>\n",
+       "      <th>rec_strict</th>\n",
+       "      <th>f1_strict</th>\n",
+       "      <th>prec_relaxed</th>\n",
+       "      <th>rec_relaxed</th>\n",
+       "      <th>f1_relaxed</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>person</td>\n",
+       "      <td>0.746914</td>\n",
+       "      <td>0.870504</td>\n",
+       "      <td>0.803987</td>\n",
+       "      <td>0.759259</td>\n",
+       "      <td>0.884892</td>\n",
+       "      <td>0.817276</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>drug</td>\n",
+       "      <td>0.717617</td>\n",
+       "      <td>0.834337</td>\n",
+       "      <td>0.771588</td>\n",
+       "      <td>0.803109</td>\n",
+       "      <td>0.933735</td>\n",
+       "      <td>0.863510</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>condition</td>\n",
+       "      <td>0.703242</td>\n",
+       "      <td>0.776860</td>\n",
+       "      <td>0.738220</td>\n",
+       "      <td>0.833749</td>\n",
+       "      <td>0.921028</td>\n",
+       "      <td>0.875218</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>value</td>\n",
+       "      <td>0.715942</td>\n",
+       "      <td>0.755352</td>\n",
+       "      <td>0.735119</td>\n",
+       "      <td>0.852174</td>\n",
+       "      <td>0.899083</td>\n",
+       "      <td>0.875000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>measurement</td>\n",
+       "      <td>0.676259</td>\n",
+       "      <td>0.712121</td>\n",
+       "      <td>0.693727</td>\n",
+       "      <td>0.805755</td>\n",
+       "      <td>0.848485</td>\n",
+       "      <td>0.826568</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>procedure</td>\n",
+       "      <td>0.560847</td>\n",
+       "      <td>0.690554</td>\n",
+       "      <td>0.618978</td>\n",
+       "      <td>0.677249</td>\n",
+       "      <td>0.833876</td>\n",
+       "      <td>0.747445</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>temporal</td>\n",
+       "      <td>0.558011</td>\n",
+       "      <td>0.664474</td>\n",
+       "      <td>0.606607</td>\n",
+       "      <td>0.693370</td>\n",
+       "      <td>0.825658</td>\n",
+       "      <td>0.753754</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>device</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.703704</td>\n",
+       "      <td>0.703704</td>\n",
+       "      <td>0.703704</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>observation</td>\n",
+       "      <td>0.528090</td>\n",
+       "      <td>0.291925</td>\n",
+       "      <td>0.376000</td>\n",
+       "      <td>0.606742</td>\n",
+       "      <td>0.335404</td>\n",
+       "      <td>0.432000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>mood</td>\n",
+       "      <td>0.240741</td>\n",
+       "      <td>0.200000</td>\n",
+       "      <td>0.218487</td>\n",
+       "      <td>0.518519</td>\n",
+       "      <td>0.430769</td>\n",
+       "      <td>0.470588</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>pregnancy_considerations</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.428571</td>\n",
+       "      <td>0.428571</td>\n",
+       "      <td>0.428571</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                        type  prec_strict  rec_strict  f1_strict  \\\n",
+       "0                     person     0.746914    0.870504   0.803987   \n",
+       "1                       drug     0.717617    0.834337   0.771588   \n",
+       "2                  condition     0.703242    0.776860   0.738220   \n",
+       "3                      value     0.715942    0.755352   0.735119   \n",
+       "4                measurement     0.676259    0.712121   0.693727   \n",
+       "5                  procedure     0.560847    0.690554   0.618978   \n",
+       "6                   temporal     0.558011    0.664474   0.606607   \n",
+       "7                     device     0.518519    0.518519   0.518519   \n",
+       "8                observation     0.528090    0.291925   0.376000   \n",
+       "9                       mood     0.240741    0.200000   0.218487   \n",
+       "10  pregnancy_considerations     0.000000    0.000000   0.000000   \n",
+       "\n",
+       "    prec_relaxed  rec_relaxed  f1_relaxed  \n",
+       "0       0.759259     0.884892    0.817276  \n",
+       "1       0.803109     0.933735    0.863510  \n",
+       "2       0.833749     0.921028    0.875218  \n",
+       "3       0.852174     0.899083    0.875000  \n",
+       "4       0.805755     0.848485    0.826568  \n",
+       "5       0.677249     0.833876    0.747445  \n",
+       "6       0.693370     0.825658    0.753754  \n",
+       "7       0.703704     0.703704    0.703704  \n",
+       "8       0.606742     0.335404    0.432000  \n",
+       "9       0.518519     0.430769    0.470588  \n",
+       "10      0.428571     0.428571    0.428571  "
+      ]
+     },
+     "execution_count": 168,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "eval_df_roberta[1:].sort_values(by=['f1_strict'], ascending=False, ignore_index=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 169,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAJZCAYAAAB7kzUWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB8BElEQVR4nO3dd1RUx98G8GepS0cREBUpggqKvWKLsWsw0ahYsWASxRIRNZrE3guosccSNJZg/8USxdhir1gxFkTBiEGxoBCUMu8fvmxcFxAQmV3yfM7Zc2T27u6z5e5+nTt3RiGEECAiIiKSRE92ACIiIvpvYzFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpCpSxcipU6fQoUMHlC1bFsbGxrC3t0f9+vURFBSktt3ixYsRGhqap/s+dOgQFAoFDh06lKfbHT9+HBMmTMDTp0/zdLvC4OzsjD59+rxzO4VCoXaxtLSEt7c3NmzY8MEfu6B99NFHGs8n83LlyhXVdt9//z0++eQTlC5dGgqF4r2yXrt2Db169YKrqyuUSiVKlCiBGjVqYPDgwUhMTCyAZ1W4wsLCUKlSJZiYmEChUODChQsf7LEy97vMi76+PmxtbeHj44OzZ8/m6z5DQ0OhUCjyffu82r17NyZMmFDg9zthwgQoFAo8evSowO+7KFMoFPl6P+7cuQOFQpHn344P7fbt2xg8eDDKly8PExMTmJqaolKlSvj+++/x119/qbb76KOPULly5Xfen6znWWSKkV27dsHb2xuJiYmYNWsWwsPDMX/+fDRo0ABhYWFq2+anGKlRowZOnDiBGjVq5Ol2x48fx8SJE7WyGMmLTp064cSJEzh+/DiWLl2KxMREdO/eHevXr5cdLc9cXV1x4sQJjUu5cuVU28ydOxcJCQlo3749jIyM8v1YERERqFmzJiIjIzFu3Djs2bMHS5cuRbt27bB37148fvy4IJ5SoXn48CF69eqFcuXKYc+ePThx4gTKly//wR932rRpOHHiBA4dOoSxY8fi+PHjaNKkCW7evPnBH/t97d69GxMnTpQdg4qgnTt3okqVKti5cye+/PJL7Ny5U/XvHTt24JNPPsnzfTo4OODEiRNo167dB0icPYNCfbQPaNasWXBxccHevXthYPDv0+ratStmzZqV7/tNTU1V9QbUq1evIKLqJHt7e9Xzr1+/Pho0aABnZ2csW7YM3bt3l5wub0xMTN75Xj5//hx6eq9r9Z9//jnfjzVv3jzo6enh0KFDsLCwULV36tQJkydPRmEuDZWcnAxTU9P3uo8bN24gNTUVPXv2RJMmTQotl7u7u+o9a9SoEaytrdG7d2+sXbtWa3/oC+L1JspOdHQ0unbtivLly+PgwYOwsrJSXffxxx9j6NCh2LZtW57v19jYWMpvXZHpGUlISECJEiXUCpFMmT8qwOvDA1evXsXhw4dVXb/Ozs4A/u0S/vnnnxEUFITSpUvD2NgYt27dyvYwzalTp+Dj4wMbGxsolUqUK1cOw4YNA/C6G3XkyJEAABcXF9Xj5XSo5+zZs+jatSucnZ1hYmICZ2dndOvWDXfv3lXbLrOr+eDBgxg4cCBKlCgBGxsbdOzYEffv31fbNjU1FaNGjULJkiVhamqKhg0b4vTp07l8ZbPm5OQEW1tb/P3332rtiYmJGDFiBFxcXGBkZITSpUtj2LBhSEpKyvH+UlJSEBQUhGrVqsHKygrFixdH/fr18b///U9tu19++QUKhQILFy5Uax8/fjz09fWxb9++93pemd78zLyPhIQEWFpawtzcPMvrFQqF2t979uxBs2bNYGVlBVNTU3h4eGD69Olq2/z666+oX78+TE1NYWFhgRYtWuDEiRNq22R24Z8/fx6dOnVCsWLFVD0/QggsXrwY1apVg4mJCYoVK4ZOnTrh9u3bOT6XPn36oGHDhgAAX19fKBQKfPTRRwWWKy9q1aoFABqfv6NHj6JZs2awsLCAqakpvL29sWvXrizv48mTJ+jbty+KFy8OMzMz+Pj4ZPka/P7772jWrBksLS1hamqKBg0aYP/+/bl6Xn369MGiRYsAqB/uvHPnDgBg0aJFaNy4Mezs7GBmZgYvLy/MmjULqampuX4tYmNj0bFjR1haWsLKygo9e/bEw4cP1bYJCwtDy5Yt4eDgABMTE3h4eGD06NEa++Xt27fRtWtXlCpVSnWou1mzZhqH4sLCwlC/fn2YmZnB3NwcrVq1QkRExDuzZn5vHThwAF988QVsbGxgaWkJPz8/JCUl4cGDB+jSpQusra3h4OCAESNGaLwWjx8/RkBAAEqXLg0jIyO4urriu+++w8uXL9W2S0xMVD2Gubk5WrdujRs3bmSZ6+bNm+jevTvs7OxgbGwMDw8P1fuWHzExMejZs6fa/QUHByMjI0O1TebhkDlz5iAkJAQuLi4wNzdH/fr1cfLkyXc+RkhICJKSkrB48WK1QiSTQqFAx44dNdrPnDmDRo0awdTUFK6urpgxY0aWud48enDr1i307dsX7u7uMDU1RenSpeHj44PLly/n8ZXJXpEpRurXr49Tp05h6NChOHXqVLY787Zt2+Dq6orq1auruuffrh7HjBmDmJgYLF26FDt27ICdnV2W97V37140atQIMTExCAkJwW+//Ybvv/9e9QXZv39/DBkyBACwdetW1ePldKjnzp07qFChAubNm4e9e/di5syZiIuLQ+3atbM8Nty/f38YGhpi/fr1mDVrFg4dOoSePXuqbfPFF19gzpw58PPzw//+9z98/vnn6NixI548eZL9C/oOz549w+PHj9W66JOTk9GkSROsXr0aQ4cOxW+//YZvvvkGoaGhaN++fY69AC9fvsTjx48xYsQIbN++HRs2bEDDhg3RsWNHrFmzRrVd165dMWDAAAQFBamO+R84cABTpkzBt99+ixYtWuQqf1pamtrlzZ2xINWvXx9xcXHo0aMHDh8+jH/++SfbbVeuXIm2bdsiIyND9dkbOnQo7t27p9pm/fr1+PTTT2FpaYkNGzZg5cqVePLkCT766CMcPXpU4z47duwINzc3bNq0CUuXLgUAfPXVVxg2bBiaN2+O7du3Y/Hixbh69Sq8vb01ftzfNHbsWNUXdOZhk8WLFxdYrryIjo4GALXP3+HDh/Hxxx/j2bNnWLlyJTZs2AALCwv4+PhoHKoFAH9/f+jp6WH9+vWYN28eTp8+jY8++kjtkOratWvRsmVLWFpaYvXq1di4cSOKFy+OVq1aaRQkWT2vsWPHolOnTgCgdkjQwcEBABAVFYXu3bvj559/xs6dO+Hv74/Zs2fjq6++yvVr0aFDB7i5uWHz5s2YMGECtm/fjlatWql9B968eRNt27bFypUrsWfPHgwbNgwbN26Ej4+P2n21bdsW586dw6xZs7Bv3z4sWbIE1atXV3tNpk2bhm7dusHT0xMbN27Ezz//jOfPn6NRo0aIjIzMVeb+/fvDysoKv/zyC77//nusX78eX3zxBdq1a4eqVati8+bN6N27N4KDg7FgwQLV7VJSUtC0aVOsWbMGw4cPx65du9CzZ0/MmjVL7YdXCIHPPvtM9R/Lbdu2oV69emjTpo1GlsjISNSuXRtXrlxBcHAwdu7ciXbt2mHo0KH56nV7+PAhvL29ER4ejsmTJ+PXX39F8+bNMWLECAwePFhj+0WLFmHfvn2YN28e1q1bh6SkJLRt2xbPnj3L8XHCw8PVeqxz48GDB+jRowd69uyJX3/9FW3atMGYMWOwdu3aHG93//592NjYYMaMGdizZw8WLVoEAwMD1K1bF9evX8/14+dIFBGPHj0SDRs2FAAEAGFoaCi8vb3F9OnTxfPnz9W2rVSpkmjSpInGfRw8eFAAEI0bN872uoMHD6raypUrJ8qVKyf++eefbHPNnj1bABDR0dH5el5paWnixYsXwszMTMyfP1/V/tNPPwkAIiAgQG37WbNmCQAiLi5OCCHEtWvXBAARGBiott26desEANG7d+93Zsh8nNTUVPHq1Stx48YN0b59e2FhYSHOnj2r2m769OlCT09PnDlzRu32mzdvFgDE7t27VW1OTk45PnZaWppITU0V/v7+onr16mrXpaSkiOrVqwsXFxcRGRkp7O3tRZMmTURaWto7n0uTJk1Un5E3Lz169Mj2NmZmZrl6nbKSkpIiPvvsM9Xj6Ovri+rVq4vvvvtOxMfHq7Z7/vy5sLS0FA0bNhQZGRlZ3ld6erooVaqU8PLyEunp6Wq3tbOzE97e3qq28ePHCwBi3Lhxavdx4sQJAUAEBwertcfGxgoTExMxatSoHJ9P5n6wadOmAs31rscLCwsTqampIjk5WRw7dkxUqFBBeHp6iidPnqi2rVevnrCzs1Pb39PS0kTlypVFmTJlVK9r5r7ToUMHtcc6duyYACCmTJkihBAiKSlJFC9eXPj4+Khtl56eLqpWrSrq1KmTq+c1aNAgkZuv2vT0dJGamirWrFkj9PX1xePHj3PcPvMxs9u3165dm+XtMjIyRGpqqjh8+LAAIC5evCiEeP0dCkDMmzcv28eMiYkRBgYGYsiQIWrtz58/FyVLlhRdunTJMXPma//27TP3kZCQELX2atWqiRo1aqj+Xrp0qQAgNm7cqLbdzJkzBQARHh4uhBDit99+EwDUvjOFEGLq1KkCgBg/fryqrVWrVqJMmTLi2bNnatsOHjxYKJVK1fsQHR0tAIiffvopx+c4evRoAUCcOnVKrX3gwIFCoVCI69evq92fl5eX2nfX6dOnBQCxYcOGHB9HqVSKevXq5bjNmzK/+97O5enpKVq1aqX6OzfPMy0tTbx69Uq4u7trfP7yq8j0jNjY2ODIkSM4c+YMZsyYgU8//RQ3btzAmDFj4OXllacR559//vk7t7lx4waioqLg7+8PpVL5PtHVvHjxAt988w3c3NxgYGAAAwMDmJubIykpCdeuXdPYvn379mp/V6lSBQBUh3UOHjwIAOjRo4fadl26dMnykFZ2Fi9eDENDQxgZGaF8+fL47bffsGHDBtSsWVO1zc6dO1G5cmVUq1ZNrdehVatWuToTadOmTWjQoAHMzc1hYGAAQ0NDrFy5UuN5GxsbY+PGjUhISECNGjUghMCGDRugr6+fq+dSrlw5nDlzRu0yefLkXL8WeWFsbIxt27YhMjISc+fORdeuXfHw4UNMnToVHh4eqv9VHD9+HImJiQgICNA4dJPp+vXruH//Pnr16qV2GMnc3Byff/45Tp48ieTkZLXbvP1Z3rlzJxQKBXr27Kn2HpUsWRJVq1bN89liBZXrXXx9fWFoaKg6TJKYmIhdu3bB2toaAJCUlIRTp06hU6dOaofE9PX10atXL9y7d0/jf3Bv7xPe3t5wcnJS7TPHjx/H48eP0bt3b41etNatW+PMmTMahzny+rwiIiLQvn172NjYQF9fH4aGhvDz80N6enq2hxTelt2+nfk8gNeHX7p3746SJUuqHidzzE/m/lW8eHGUK1cOs2fPRkhICCIiIjR6DPfu3Yu0tDT4+fmpvSZKpRJNmjTJ9efn7YGVHh4eAKAxaNLDw0PtEPWBAwdgZmam6m3KlHm2W2ZvVXbfe2+Pb0tJScH+/fvRoUMHmJqaqj2ntm3bIiUlJVeHTN504MABeHp6ok6dOhoZhRA4cOCAWnu7du3Uvrve/g4vSCVLltTIVaVKlXc+VlpaGqZNmwZPT08YGRnBwMAARkZGuHnzZpa/S/lRZAawZqpVq5bqeHJqaiq++eYbzJ07F7Nmzcr1QNbMLtScZB6TLVOmTP7DZqF79+7Yv38/xo4di9q1a8PS0hIKhQJt27bNsovfxsZG7W9jY2MAUG2bkJAA4PWH8E0GBgYat81Jly5dMHLkSKSmpuLy5csYM2YMunbtivPnz8Pd3R3A6+P3t27dgqGhYZb3kVNBuHXrVnTp0gWdO3fGyJEjUbJkSRgYGGDJkiVYtWqVxvZubm5o1KgRdu3ahYEDB+bqPcukVCpVn5HC4uHhofrCFUJg3rx5GD58OMaOHYuNGzfm6vOU+V5m9VxLlSqFjIwMPHnyRG3Q5Nvb/v333xBCwN7ePsvHcHV1zdsTK6Bc7zJz5kx8/PHHSE5ORnh4OKZPn47PPvsMp06dgrGxMZ48eQIhRLYZ3syZ6e19IrMtc7vMQ1Zv//C96fHjxzAzM8vX84qJiUGjRo1QoUIFzJ8/H87OzlAqlTh9+jQGDRqU4yG9nJ5H5r6d+TxevHiBRo0aQalUYsqUKShfvjxMTU1VY00yH0ehUGD//v2YNGkSZs2ahaCgIBQvXhw9evTA1KlTYWFhoXpNateunWWW3I61Kl68uNrfmWesZdWekpKi+jshIQElS5bUKNjt7OxgYGCges4JCQlZfse9/VolJCQgLS0NCxYsUDsc9Ka8njqdkJCgGof4puw+h+/6Ds9O2bJlVYcrcyur73xjY+N3Ptbw4cOxaNEifPPNN2jSpAmKFSsGPT099O/fP9ef03cpcsXImwwNDTF+/HjMnTtXbQ6Jd8nuf6ZvsrW1BQC14/nv69mzZ9i5cyfGjx+P0aNHq9ozx1PkR+aH78GDByhdurSqPS0tTWOnyImtra3qB7x+/frw8PBAkyZNEBgYiJ07dwIASpQoARMTkyyLh8zrs7N27Vq4uLggLCxM7fV/e1BaphUrVmDXrl2oU6cOFi5cCF9fX9StWzfXz0cmhUKBwMBATJo0SfW5zM3nKfO9jIuL07ju/v370NPTQ7FixTQe600lSpSAQqHAkSNHVF96b8qq7V0KIte7uLq6qj5/jRs3homJCb7//nssWLAAI0aMUH05ZpcB0Pz8PXjwQGPbBw8ewM3NTW37BQsWZHtc/u2iLi/Pa/v27UhKSsLWrVvh5OSkas/rvC3Z7duZ78uBAwdw//59HDp0SO0MqKymG3BycsLKlSsBvO793bhxIyZMmIBXr15h6dKlqtdk8+bNapkLi42NDU6dOgUhhNprHR8fj7S0NFU+GxsbjdcB0HzPixUrpuo9GzRoUJaP6eLikueMefkc5lerVq2wYMECnDx58oOf/bJ27Vr4+flh2rRpau2PHj1S9U6+ryJzmCarNx/4twsysyoFclcJvkv58uVRrlw5rFq1KtsfzMzHAt5d5QKvv8iEEBo/CCtWrEB6enq+cmae7bBu3Tq19o0bNyItLS1f9wm8Pr3Sz88Pu3btUp0x8cknnyAqKgo2NjaqHqo3L1n9byGTQqGAkZGR2hfMgwcPNM6mAYDLly9j6NCh8PPzw5EjR1ClShX4+vq+14DcDyW7z+X9+/eRmJio+lx6e3vDysoKS5cuzXagb4UKFVC6dGmsX79ebZukpCRs2bJFdSZLTj755BMIIfDXX39l+R55eXnl+TkWRK68GjVqFNzc3DBjxgw8f/4cZmZmqFu3LrZu3aq2r2VkZGDt2rUoU6aMxnwob+8Tx48fx927d1X7TIMGDWBtbY3IyMgsX6tatWrlag6a7L4DMj/rb+7vQggsX7489y9EFs8jc9/OfB5ZPQ4ALFu2LMf7LV++PL7//nt4eXnh/PnzAF7/ABoYGCAqKirb1+RDatasGV68eIHt27ertWcOcm/WrBkAoGnTpgA0X5u350UyNTVF06ZNERERgSpVqmT5fPLSg5yZITIyUvWavZlRoVCosr2vwMBAmJmZISAgIMvBrkKIfJ3amxWFQqHx+dm1a5fapGrvq8j0jLRq1QplypSBj48PKlasiIyMDFy4cAHBwcEwNzfH119/rdrWy8sLv/zyC8LCwlSzYubnS3jRokXw8fFBvXr1EBgYiLJlyyImJgZ79+5V7QSZ9zt//nz07t0bhoaGqFChgtqcE5ksLS3RuHFjzJ49GyVKlICzszMOHz6MlStX5rv69PDwQM+ePTFv3jwYGhqiefPmuHLlCubMmQNLS8t83WemyZMnIywsDGPHjsXvv/+OYcOGYcuWLWjcuDECAwNRpUoVZGRkICYmBuHh4QgKCsq29+KTTz7B1q1bERAQgE6dOiE2NhaTJ0+Gg4OD2sRWSUlJ6NKlC1xcXLB48WIYGRlh48aNqFGjBvr27avxJZVfhw8fVh06SU9Px927d7F582YAQJMmTVQ9Ge/y5Zdf4unTp/j8889RuXJl6Ovr488//8TcuXOhp6eHb775BsDr8RXBwcHo378/mjdvji+++AL29va4desWLl68iIULF0JPTw+zZs1Cjx498Mknn+Crr77Cy5cvMXv2bDx9+hQzZsx4Z54GDRrgyy+/RN++fXH27Fk0btwYZmZmiIuLw9GjR+Hl5YWBAwfm6bUqiFx5ZWhoiGnTpqFLly6YP38+vv/+e0yfPh0tWrRA06ZNMWLECBgZGWHx4sW4cuUKNmzYoNFrcfbsWfTv3x+dO3dGbGwsvvvuO5QuXRoBAQEAXr8nCxYsQO/evfH48WN06tQJdnZ2ePjwIS5evIiHDx9iyZIl78ya+R0wc+ZMtGnTBvr6+qhSpQpatGgBIyMjdOvWDaNGjUJKSgqWLFmS56J669atMDAwQIsWLXD16lWMHTsWVatWRZcuXQC8LnSLFSuGAQMGYPz48TA0NMS6detw8eJFtfu5dOkSBg8ejM6dO8Pd3R1GRkY4cOAALl26pOqpdXZ2xqRJk/Ddd9/h9u3baN26NYoVK4a///4bp0+fhpmZ2Qed98XPzw+LFi1C7969cefOHXh5eeHo0aOYNm0a2rZti+bNmwMAWrZsicaNG2PUqFFISkpCrVq1cOzYsSznDJo/fz4aNmyIRo0aYeDAgXB2dsbz589x69Yt7NixQ2OMx7sEBgZizZo1aNeuHSZNmgQnJyfs2rULixcvxsCBAwtskkAXFxf88ssv8PX1RbVq1TB48GBUr14dwOszhFatWgUhBDp06PDej/XJJ58gNDQUFStWRJUqVXDu3DnMnj27YIcpFMgwWC0QFhYmunfvLtzd3YW5ubkwNDQUZcuWFb169RKRkZFq2965c0e0bNlSWFhYCADCyclJCJH1mQKZsjqbRojXZye0adNGWFlZCWNjY1GuXDmN0cVjxowRpUqVEnp6elnex5vu3bsnPv/8c1GsWDFhYWEhWrduLa5cuaJx9knmqPS3z1zJKufLly9FUFCQsLOzU43APnHixDvPaMkEQAwaNCjL60aOHCkAiMOHDwshhHjx4oX4/vvvRYUKFYSRkZGwsrISXl5eIjAwUDx48EB1u6wee8aMGcLZ2VkYGxsLDw8PsXz5ctUZA5l69uwpTE1NxdWrV9Vuu2nTJgFAzJ07N8fn0qRJE1GpUqV3Pufszrp51/v3tr1794p+/foJT09PYWVlJQwMDISDg4Po2LGjOHHihMb2u3fvFk2aNBFmZmbC1NRUeHp6ipkzZ6pts337dlG3bl2hVCqFmZmZaNasmTh27JjaNpmv28OHD7PMtWrVKlG3bl1hZmYmTExMRLly5YSfn5/a2VFZyWkfKYhceXk8IYSoW7euKFasmHj69KkQQogjR46Ijz/+WPW86tWrJ3bs2KF2m8x9Jzw8XPTq1UtYW1sLExMT0bZtW3Hz5k2Nxzh8+LBo166dKF68uDA0NBSlS5cW7dq1U8uU0/N6+fKl6N+/v7C1tRUKhULt7LodO3aIqlWrCqVSKUqXLi1GjhypOhPkXZ+zzMc8d+6c8PHxEebm5sLCwkJ069ZN/P3332rbHj9+XNSvX1+YmpoKW1tb0b9/f3H+/Hm1syb+/vtv0adPH1GxYkVhZmYmzM3NRZUqVcTcuXM1zlTbvn27aNq0qbC0tBTGxsbCyclJdOrUSfz+++85Zs7ueyu71693797CzMxMrS0hIUEMGDBAODg4CAMDA+Hk5CTGjBkjUlJS1LZ7+vSp6Nevn7C2thampqaiRYsW4s8//9Q4m0aI12eQ9OvXT5QuXVoYGhoKW1tb4e3trTqzKnObN1+vnNy9e1d0795d2NjYCENDQ1GhQgUxe/ZstbPNMu9v9uzZGrfPKmN2oqKiREBAgHBzcxPGxsbCxMREeHp6iuHDh6udxZndd1/v3r1Vv4HZPc8nT54If39/YWdnJ0xNTUXDhg3FkSNHRJMmTbI8MzU/FEIU4hSQRERERG8pMmNGiIiISDcVmTEjRIVJCPHOQcX6+vp5PmuEiOi/iD0jRPmwevVqGBoa5ng5fPiw7JhERDqBY0aI8iEhIeGdEw5ld9YUERGpYzFCREREUvEwDREREUmlEwNYMzIycP/+fVhYWHBAIBERkY4QQuD58+coVapUjmsX6UQxcv/+fTg6OsqOQURERPkQGxub44ytOlGMZA4CjI2Nfe8pzImIiKhwJCYmwtHR8Z2D+XWiGMk8NGNpaclihIiISMe8a4gFB7ASERGRVCxGiIiISCqdOExDRES6Kz09HampqbJj0AdgaGgIfX39974fFiNERPRBCCHw4MEDPH36VHYU+oCsra1RsmTJ95p6g8UIERF9EJmFiJ2dHUxNTTlPVBEjhEBycjLi4+MBAA4ODvm+LxYjRERU4NLT01WFiI2Njew49IGYmJgAAOLj42FnZ5fvQzYcwEpERAUuc4yIqamp5CT0oWW+x+8zLojFCBERfTA8NFP0FcR7zGKEiIiIpGIxQkRERFJxACsRERUq59G7CvXx7sxol6fthRD46quvsHnzZjx58gQRERGoVq3ahwmXBxMmTMD27dtx4cIF2VEKHHtGiIiI3rBnzx6EhoZi586diIuLQ2JiInx8fFCqVCkoFAps3769wB7ro48+wrBhw3K17YgRI7B///5c33dBZ/2QWIwQERG9ISoqCg4ODvD29kbJkiWRlJSEqlWrYuHChVLyCCGQlpYGc3PzInuaNIsRIiKi/9enTx8MGTIEMTExUCgUcHZ2Rps2bTBlyhR07NgxX/e5ePFiuLu7Q6lUwt7eHp06dVI91uHDhzF//nwoFAooFArcuXMHhw4dgkKhwN69e1GrVi0YGxvjyJEjmDBhgsbholWrVqFSpUowNjaGg4MDBg8eDABwdnYGAHTo0EH1PLQZx4xouwlWudzu2YfNQZQpt59JgJ9L0jnz589HuXLl8OOPP+LMmTPvve7K2bNnMXToUPz888/w9vbG48ePceTIEdVj3bhxA5UrV8akSZMAALa2trhz5w4AYNSoUZgzZw5cXV1hbW2Nw4cPq933kiVLMHz4cMyYMQNt2rTBs2fPcOzYMQDAmTNnYGdnh59++gmtW7cukPVjPiQWI0RERP/PysoKFhYW0NfXR8mSJd/7/mJiYmBmZoZPPvkEFhYWcHJyQvXq1VWPZWRkBFNT0ywfa9KkSWjRokW29z1lyhQEBQXh66+/VrXVrl0bwOuiBvh33Rhtx8M0REREH0iLFi3g5OQEV1dX9OrVC+vWrUNycnKublurVq1sr4uPj8f9+/fRrFmzgooqFYsRIiKiD8TCwgLnz5/Hhg0b4ODggHHjxqFq1aq5WsnYzMws2+sy14QpKliMEBERfUAGBgZo3rw5Zs2ahUuXLuHOnTs4cOAAAMDIyAjp6el5vk8LCws4OzvneKqvoaFhvu5bBo4ZISKi7HHAMl68eIFbt26p/o6OjsaFCxdQvHhxlC1bNsfb7ty5E7dv30bjxo1RrFgx7N69GxkZGahQoQKA12e9nDp1Cnfu3IG5uTmKFy+e61wTJkzAgAEDYGdnhzZt2uD58+c4duwYhgwZorrv/fv3o0GDBjA2NkaxYsXy8ewLB4sRIiIqVHmdEVW2s2fPomnTpqq/hw8fDgDo3bs3QkNDc7yttbU1tm7digkTJiAlJQXu7u7YsGEDKlWqBOD1RGa9e/eGp6cn/vnnH0RHR+c6V+/evZGSkoK5c+dixIgRKFGihOq0YQAIDg7G8OHDsXz5cpQuXVp1lo42UgghhOwQ75KYmAgrKys8e/YMlpaWsuMULp7aS9qG/1P+b8nn+52SkoLo6Gi4uLhAqVR+gGCkLXJ6r3P7+80xI0RERCQVixEiIqJ8OnLkCMzNzbO9UO5wzAgREVE+1apVq0iuolvYWIwQERHlk4mJCdzc3GTH0Hk8TENERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFc+mISKiwpWXWV0L5PHyNhOwEAJfffUVNm/ejCdPniAiIgLVqlX7MNly4OzsjGHDhmHYsGGF/thvUigU2LZtGz777LMP9hjsGSEiInrDnj17EBoaip07dyIuLg6JiYnw8fFBqVKloFAosH37dtkRixwWI0RERG+IioqCg4MDvL29UbJkSSQlJaFq1apYuHBhgdz/q1evCuR+ihIWI0TaZIJV7i5E9EH06dMHQ4YMQUxMDBQKBZydndGmTRtMmTIFHTt2zNd9Ojs7Y8qUKejTpw+srKzwxRdfAACOHz+Oxo0bw8TEBI6Ojhg6dCiSkpKyvZ+QkBB4eXnBzMwMjo6OCAgIwIsXL1TX9+vXD1WqVMHLly8BAKmpqahZsyZ69Oih2mbHjh2oWbMmlEolXF1dMXHiRKSlpamuv3nzJho3bgylUglPT0/s27cvX885r1iMEBER/b/58+dj0qRJKFOmDOLi4nDmzJkCud/Zs2ejcuXKOHfuHMaOHYvLly+jVatW6NixIy5duoSwsDAcPXoUgwcPzvY+9PT08MMPP+DKlStYvXo1Dhw4gFGjRqmu/+GHH5CUlITRo0cDAMaOHYtHjx5h8eLFAIC9e/eiZ8+eGDp0KCIjI7Fs2TKEhoZi6tSpAICMjAx07NgR+vr6OHnyJJYuXYpvvvmmQJ7/u3AAKxER0f+zsrKChYUF9PX1UbJkyQK7348//hgjRoxQ/e3n54fu3burBqe6u7vjhx9+QJMmTbBkyRIolUqN+3hzIKuLiwsmT56MgQMHqooNc3NzrF27Fk2aNIGFhQWCg4Oxf/9+WFm97k2dOnUqRo8ejd69ewMAXF1dMXnyZIwaNQrjx4/H77//jmvXruHOnTsoU6YMAGDatGlo06ZNgb0O2WExQkQkS24PueXxbBDSPrVq1VL7+9y5c7h16xbWrVunahNCICMjA9HR0fDw8NC4j4MHD2LatGmIjIxEYmIi0tLSkJKSgqSkJJiZmQEA6tevjxEjRmDy5Mn45ptv0LhxY7XHPHPmjKonBADS09ORkpKC5ORkXLt2DWXLllUVIpn3VxhYjBAREX1gmcVCpoyMDHz11VcYOnSoxrZly5bVaLt79y7atm2LAQMGYPLkyShevDiOHj0Kf39/pKamqt3vsWPHoK+vj5s3b2o85sSJE7Mc+6JUKiGE0GhXKBS5fo7vg8UIERFRIatRowauXr2a6xV/z549i7S0NAQHB0NP7/Vwz40bN2psN3v2bFy7dg2HDx9Gq1at8NNPP6Fv376qx7x+/Xq2j+np6YmYmBjcv38fpUqVAgCcOHEiP08vz/7bxQi7SImI6B1evHiBW7duqf6Ojo7GhQsXULx48Sx7MXLjm2++Qb169TBo0CB88cUXMDMzw7Vr17Bv3z4sWLBAY/ty5cohLS0NCxYsgI+PD44dO4alS5eqbXPhwgWMGzcOmzdvRoMGDTB//nx8/fXXaNKkCVxdXTFu3Dh88skncHR0ROfOnaGnp4dLly7h8uXLmDJlCpo3b44KFSrAz88PwcHBSExMxHfffZev55dX+SpGFi9ejNmzZyMuLg6VKlXCvHnz0KhRo2y3X7duHWbNmoWbN2/CysoKrVu3xpw5c2BjY5Pv4EREpKN07D94Z8+eRdOmTVV/Dx8+HADQu3dvhIaG5us+q1SpgsOHD+O7775Do0aNIIRAuXLl4Ovrm+X21apVQ0hICGbOnIkxY8agcePGmD59Ovz8/AAAKSkp6NGjB/r06QMfHx8AgL+/P3bt2oVevXrhjz/+QKtWrbBz505MmjQJs2bNgqGhISpWrIj+/fsDeH22zrZt2+Dv7486derA2dkZP/zwA1q3bp2v55gXCpHVQaIchIWFoVevXli8eDEaNGiAZcuWYcWKFYiMjMyyQjx69CiaNGmCuXPnwsfHB3/99RcGDBgAd3d3bNu2LVePmZiYCCsrKzx79gyWlpZ5iZszXegZ0YWMVHB04f3Oyzwn/FzmrAi/3ykpKYiOjoaLi0uWZ4ZQ0ZHTe53b3+8894yEhITA399fVUnNmzcPe/fuxZIlSzB9+nSN7U+ePAlnZ2fVIB0XFxd89dVXmDVrVraP8fLlS9WkLZlPhohIVziP3pWr7e7wN5oIQB4nPXv16hXOnTuHli1bqrW3bNkSx48fz/I23t7euHfvHnbv3g0hBP7++29s3rwZ7dq1y/Zxpk+fDisrK9XF0dExLzGJiIgKxZEjR2Bubp7thXInTz0jjx49Qnp6Ouzt7dXa7e3t8eDBgyxv4+3tjXXr1sHX1xcpKSlIS0tD+/btsxygk2nMmDGqY3LA654RFiRERKRtatWqhQsXLsiOofPyNYD17fOOhRDZnoscGRmJoUOHYty4cWjVqhXi4uIwcuRIDBgwACtXrszyNsbGxjA2Ns5PNCIiokJjYmKS69NzKXt5KkZKlCgBfX19jV6Q+Ph4jd6STNOnT0eDBg0wcuRIAK9HEJuZmaFRo0aYMmUKHBwc8hmdiIi0XR7PkSAdVBDvcZ7GjBgZGaFmzZoaq/jt27cP3t7eWd4mOTlZNUFLJn19fQD8kBIRFVWGhoYAXv8GUNGW+R5nvuf5kefDNMOHD0evXr1Qq1Yt1K9fHz/++CNiYmIwYMAAAK/He/z1119Ys2YNAMDHxwdffPEFlixZojpMM2zYMNSpU0c1wxsRERUt+vr6sLa2Rnx8PADA1NS00KYWp8IhhEBycjLi4+NhbW2t6mjIjzwXI76+vkhISMCkSZMQFxeHypUrY/fu3XBycgIAxMXFISYmRrV9nz598Pz5cyxcuBBBQUGwtrbGxx9/jJkzZ+Y7NBERab/MVW8zCxIqmqytrd97heN8DWANCAhAQEBAltdlNRvdkCFDMGTIkPw8FBER6SiFQgEHBwfY2dmpLeZGRYehoeF79Yhk+m+vTUNERB+cvr5+gfxgUdGVpwGsRERERAWNxQgRERFJxWKEiIiIpOKYESKi/yAu5kfahD0jREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpeGovvb8JVnnY9tmHy0FERDqJPSNEREQkFYsRIiIikoqHaei/I7eHk3goiYioULFnhIiIiKRiMUJERERS8TCNJFykioiI6DUWI0QEgAUyEcnDwzREREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpOJCeURERIVhglUut3v2YXNoIfaMEBERkVQsRoiIiEgqFiNEREQkFceMEBUC59G7crXdHeUHDkJEpIXYM0JERERSsRghIiIiqViMEBERkVRFbsxIbo/NAzw+T0REpA3YM0JERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSGcgOQPQ+nEfvyvW2d5QfMAgREeUbe0aIiIhIKhYjREREJBWLESIiIpIqX8XI4sWL4eLiAqVSiZo1a+LIkSM5bv/y5Ut89913cHJygrGxMcqVK4dVq1blKzAREREVLXkewBoWFoZhw4Zh8eLFaNCgAZYtW4Y2bdogMjISZcuWzfI2Xbp0wd9//42VK1fCzc0N8fHxSEtLe+/w9GHldnAoB4YSEdH7yHMxEhISAn9/f/Tv3x8AMG/ePOzduxdLlizB9OnTNbbfs2cPDh8+jNu3b6N48eIAAGdn5/dLTUREREVGng7TvHr1CufOnUPLli3V2lu2bInjx49neZtff/0VtWrVwqxZs1C6dGmUL18eI0aMwD///JPt47x8+RKJiYlqFyIiIiqa8tQz8ujRI6Snp8Pe3l6t3d7eHg8ePMjyNrdv38bRo0ehVCqxbds2PHr0CAEBAXj8+HG240amT5+OiRMn5iUaERER6ah8DWBVKBRqfwshNNoyZWRkQKFQYN26dahTpw7atm2LkJAQhIaGZts7MmbMGDx79kx1iY2NzU9MIiIi0gF56hkpUaIE9PX1NXpB4uPjNXpLMjk4OKB06dKwsrJStXl4eEAIgXv37sHd3V3jNsbGxjA2Ns5LNCIiItJReeoZMTIyQs2aNbFv3z619n379sHb2zvL2zRo0AD379/HixcvVG03btyAnp4eypQpk4/IREREVJTk+TDN8OHDsWLFCqxatQrXrl1DYGAgYmJiMGDAAACvD7H4+fmptu/evTtsbGzQt29fREZG4o8//sDIkSPRr18/mJiYFNwzISIiIp2U51N7fX19kZCQgEmTJiEuLg6VK1fG7t274eTkBACIi4tDTEyMantzc3Ps27cPQ4YMQa1atWBjY4MuXbpgypQpBfcsiIiISGfla9XegIAABAQEZHldaGioRlvFihU1Du0QERERAVybhoiIiCRjMUJERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSGcgOQEREpMucR+/K1XZ3lB84iA5jzwgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikMpAdgIiIiLTEBKs8bPuswB6WPSNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpDGQHICL6IHK7FHoBLoNORPnDnhEiIiKSij0jRESk+9gTptPYM0JERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUmVr2Jk8eLFcHFxgVKpRM2aNXHkyJFc3e7YsWMwMDBAtWrV8vOwREREVATluRgJCwvDsGHD8N133yEiIgKNGjVCmzZtEBMTk+Ptnj17Bj8/PzRr1izfYYmIiKjoyXMxEhISAn9/f/Tv3x8eHh6YN28eHB0dsWTJkhxv99VXX6F79+6oX7/+Ox/j5cuXSExMVLsQERFR0ZSnYuTVq1c4d+4cWrZsqdbesmVLHD9+PNvb/fTTT4iKisL48eNz9TjTp0+HlZWV6uLo6JiXmERERKRD8lSMPHr0COnp6bC3t1drt7e3x4MHD7K8zc2bNzF69GisW7cOBga5WwpnzJgxePbsmeoSGxubl5hERESkQ/K1UJ5CoVD7Wwih0QYA6enp6N69OyZOnIjy5cvn+v6NjY1hbGycn2hERESkY/JUjJQoUQL6+voavSDx8fEavSUA8Pz5c5w9exYREREYPHgwACAjIwNCCBgYGCA8PBwff/zxe8QnIiIiXZenwzRGRkaoWbMm9u3bp9a+b98+eHt7a2xvaWmJy5cv48KFC6rLgAEDUKFCBVy4cAF169Z9v/RERESk8/J8mGb48OHo1asXatWqhfr16+PHH39ETEwMBgwYAOD1eI+//voLa9asgZ6eHipXrqx2ezs7OyiVSo12IiIi+m/KczHi6+uLhIQETJo0CXFxcahcuTJ2794NJycnAEBcXNw75xwhIiIiypSvAawBAQEICAjI8rrQ0NAcbzthwgRMmDAhPw9LRERERRDXpiEiIiKpWIwQERGRVPk6TENEJIPz6F253vaO8gMGIaICxZ4RIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVAayAxAREdGH5Tx6V662u6P8wEGywZ4RIiIikorFCBEREUnFYoSIiIikylcxsnjxYri4uECpVKJmzZo4cuRItttu3boVLVq0gK2tLSwtLVG/fn3s3bs334GJiIioaMlzMRIWFoZhw4bhu+++Q0REBBo1aoQ2bdogJiYmy+3/+OMPtGjRArt378a5c+fQtGlT+Pj4ICIi4r3DExERke7LczESEhICf39/9O/fHx4eHpg3bx4cHR2xZMmSLLefN28eRo0ahdq1a8Pd3R3Tpk2Du7s7duzY8d7hiYiISPflqRh59eoVzp07h5YtW6q1t2zZEsePH8/VfWRkZOD58+coXrx4ttu8fPkSiYmJahciIiIqmvJUjDx69Ajp6emwt7dXa7e3t8eDBw9ydR/BwcFISkpCly5dst1m+vTpsLKyUl0cHR3zEpOIiIh0SL4GsCoUCrW/hRAabVnZsGEDJkyYgLCwMNjZ2WW73ZgxY/Ds2TPVJTY2Nj8xiYiISAfkaQbWEiVKQF9fX6MXJD4+XqO35G1hYWHw9/fHpk2b0Lx58xy3NTY2hrGxcV6iERERkY7KU8+IkZERatasiX379qm179u3D97e3tnebsOGDejTpw/Wr1+Pdu3a5S8pERERFUl5Xptm+PDh6NWrF2rVqoX69evjxx9/RExMDAYMGADg9SGWv/76C2vWrAHwuhDx8/PD/PnzUa9ePVWviomJCaysrArwqRAREZEuynMx4uvri4SEBEyaNAlxcXGoXLkydu/eDScnJwBAXFyc2pwjy5YtQ1paGgYNGoRBgwap2nv37o3Q0ND3fwZERESk0/K1am9AQAACAgKyvO7tAuPQoUP5eQgiIiL6j+DaNERERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFKxGCEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERScVihIiIiKRiMUJERERSsRghIiIiqViMEBERkVQsRoiIiEgqFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpGIxQkRERFLlqxhZvHgxXFxcoFQqUbNmTRw5ciTH7Q8fPoyaNWtCqVTC1dUVS5cuzVdYIiIiKnryXIyEhYVh2LBh+O677xAREYFGjRqhTZs2iImJyXL76OhotG3bFo0aNUJERAS+/fZbDB06FFu2bHnv8ERERKT78lyMhISEwN/fH/3794eHhwfmzZsHR0dHLFmyJMvtly5dirJly2LevHnw8PBA//790a9fP8yZM+e9wxMREZHuM8jLxq9evcK5c+cwevRotfaWLVvi+PHjWd7mxIkTaNmypVpbq1atsHLlSqSmpsLQ0FDjNi9fvsTLly9Vfz979gwAkJiY+M6MGS+T37lNpkSFyOWG737cvMptziKVESjwnHy/Cw7f74LD91uCl0XotdSFjECucmb+bgvxjvsVefDXX38JAOLYsWNq7VOnThXly5fP8jbu7u5i6tSpam3Hjh0TAMT9+/ezvM348eMFAF544YUXXnjhpQhcYmNjc6wv8tQzkkmhUKj9LYTQaHvX9lm1ZxozZgyGDx+u+jsjIwOPHz+GjY1Njo+TF4mJiXB0dERsbCwsLS0L5D4LGjMWHF3IyYwFRxdyMmPB0YWc/9WMQgg8f/4cpUqVynG7PBUjJUqUgL6+Ph48eKDWHh8fD3t7+yxvU7JkySy3NzAwgI2NTZa3MTY2hrGxsVqbtbV1XqLmmqWlpdZ+MDIxY8HRhZzMWHB0ISczFhxdyPlfzGhlZfXObfI0gNXIyAg1a9bEvn371Nr37dsHb2/vLG9Tv359je3Dw8NRq1atLMeLEBER0X9Lns+mGT58OFasWIFVq1bh2rVrCAwMRExMDAYMGADg9SEWPz8/1fYDBgzA3bt3MXz4cFy7dg2rVq3CypUrMWLEiIJ7FkRERKSz8jxmxNfXFwkJCZg0aRLi4uJQuXJl7N69G05OTgCAuLg4tTlHXFxcsHv3bgQGBmLRokUoVaoUfvjhB3z++ecF9yzywdjYGOPHj9c4HKRNmLHg6EJOZiw4upCTGQuOLuRkxpwphHjX+TZEREREHw7XpiEiIiKpWIwQERGRVCxGiIiISCoWI0RERCQVixEiIiKSisUIERERSZWvtWl0xQ8//JDrbYcOHfoBk+Tdq1evEB8fj4yMDLX2smXLSkpEH8KaNWvg6+urcV7/q1ev8Msvv6hNICjbrVu3EBUVhcaNG8PExOSda1LJwP2GtM3PP/+MpUuXIjo6GidOnICTkxPmzZsHFxcXfPrpp7LjaY0iPc+Ii4tLrrZTKBS4ffv2B06TOzdv3kS/fv1w/PhxtfbML/709HQpuapXr57rH57z589/4DS5p+1fBPr6+oiLi4OdnZ1ae0JCAuzs7KS9329n8fX1xYEDB6BQKHDz5k24urrC398f1tbWCA4Olh1Ra/ebrDx9+hSnT5/OsmjShuKzWLFiWe7rCoUCSqUSbm5u6NOnD/r27Ssh3b/++ecfCCFgamoKALh79y62bdsGT09PtGzZUmq2TEuWLMG4ceMwbNgwTJ06FVeuXIGrqytCQ0OxevVqHDx4UEquxMTEXG9bWOvoFOmekejoaNkR8qxPnz4wMDDAzp074eDgoDX/8/zss89kR8izt78IMn+QrK2tMW/ePK0oRrLrXbh3716uFpcqDIGBgTAwMEBMTAw8PDxU7b6+vggMDNSKYkRb95u37dixAz169EBSUhIsLCzUcioUCq0oRsaNG4epU6eiTZs2qFOnDoQQOHPmDPbs2YNBgwYhOjoaAwcORFpaGr744gtpOT/99FN07NgRAwYMwNOnT1G3bl0YGhri0aNHCAkJwcCBA6Vly7RgwQIsX74cn332GWbMmKFqr1WrltQlUaytrXO9jxRaIS9Iq5iamopr167JjlEkeHh4iG3btgkhhDA3NxdRUVFCCCEuX74sbGxsJCYTolq1aqJ69epCT09PeHl5ierVq6suVapUERYWFqJz585SM2ayt7cXFy5cEEKov463b98WZmZmMqOp6Mp+4+7uLr7++muRlJQkO0q2OnbsKJYsWaLRvnTpUtGxY0chhBA//PCDqFy5cmFHU2NjYyOuXLkihBBi+fLlokqVKiI9PV1s3LhRVKxYUWq2TEqlUty5c0cIob7v3LhxQyiVSmm5Dh06pLqEhoaKkiVLitGjR4v//e9/4n//+58YPXq0cHBwEKGhoYWWqUj3jAwfPjzX24aEhHzAJLnn6emJR48eyY5RJERHR6N69eoa7cbGxkhKSpKQ6F+ZPU0XLlxAq1atYG5urrrOyMgIzs7O0tdvypSUlKTqCn/To0ePtGadDV3Zb/766y8MHTo0y9dTW+zduxczZ87UaG/WrBmCgoIAAG3btsXo0aMLO5qa5ORkWFhYAHi9EnzHjh2hp6eHevXq4e7du1KzZXJxccGFCxdUa7dl+u233+Dp6SkpFdCkSRPVvydNmoSQkBB069ZN1da+fXt4eXnhxx9/RO/evQslU5EuRiIiInK1nTZ16c6cOROjRo3CtGnT4OXlBUNDQ7XrC+v4XU7S09Mxd+5cbNy4ETExMXj16pXa9Y8fP5aUTJ22fhEAwPjx4wEAzs7O8PX1hVKplJonJ40bN8aaNWswefJkAK/3l4yMDMyePRtNmzaVnO41XdhvAKBVq1Y4e/YsXF1dZUfJVvHixbFjxw4EBgaqte/YsQPFixcHANVhJpnc3Nywfft2dOjQAXv37lXljY+P15r3e+TIkRg0aBBSUlIghMDp06exYcMGTJ8+HStWrJAdDwBw4sQJLF26VKO9Vq1a6N+/f+EFKbQ+GMoVhUIhFAqF0NPTU7tktmmDsWPHCgcHBzF79myhVCrF5MmThb+/v7CxsRHz58+XHU9l1apVonTp0uKXX34RZmZmYsOGDWLKlCmqf2uTly9fitjYWHH37l21iza4evWqsLW1Fa1btxZGRkaiU6dOwsPDQ9jb24tbt27JjieE0I39RgghVqxYIcqWLSvGjx8vNm/erOoWz7xogx9//FHo6+sLHx8fMXnyZDFlyhTRvn17YWBgIFasWCGEEGLOnDmiS5cuUnNu2rRJGBoaCj09PdGiRQtV+7Rp00Tr1q0lJlP3448/irJly6o+o2XKlFG9jtqgfPnyYvjw4Rrtw4cPF+XLly+0HEX6bBpddPjw4Ryvf7N7TZZy5crhhx9+QLt27WBhYYELFy6o2k6ePIn169fLjqiyfPlyTJkyBbGxsQCA0qVLY8KECfD395ec7DVdOQvkwYMHWLJkCc6dO4eMjAzUqFEDgwYNgoODg+xoAHRjvwEAPb3sp3bSpvf72LFjWLhwIa5fvw4hBCpWrIghQ4bA29tbdjQ1Dx48QFxcHKpWrap6bU+fPg1LS0tUrFhRcjp1jx49QkZGhsaZc7Lt3r0bn3/+OcqVK4d69eoBAE6ePImoqChs2bIFbdu2LZQc/6li5MyZM9i0aVOWhxa2bt0qKZXuMTMzw7Vr11C2bFk4ODhg165dqFGjBm7fvo3q1avj2bNnsiMiLS0N69atQ6tWrVCyZEmt/SJo0KABDAwMMHr06CzPAqlataqkZET0vqKjo5GWlgZ3d3e19ps3b8LQ0BDOzs5ygr3l3r17WLx4Mf78808IIeDp6YkBAwbA0dGx0DIU6TEjb8qcQKply5bYt28fWrZsiZs3b+LBgwfo0KGD7Hhqnj59ipUrV+LatWtQKBTw9PREv379tOZUzzJlyiAuLg5ly5aFm5sbwsPDUaNGDZw5c0ZrBjQaGBhg4MCBuHbtGgCgRIkSkhNl7cKFCzh37pzW/S/uTX/88UeO1zdu3LiQkuRM2/cbXRETE5Pj9doygVxSUhJmzJiB/fv3ZzlnizbMHdWnTx/069dPoxg5deoUVqxYgUOHDskJ9pYyZcpg2rRpUjP8Z3pGqlSpgq+++gqDBg2ChYUFLl68CBcXF3z11VdwcHDAxIkTZUcEAJw9exatWrWCiYmJ6hz/s2fP4p9//lH96Ms2evRoWFpa4ttvv8XmzZvRrVs3ODs7IyYmBoGBgWrn08vUtGlTfP3111o9R0rt2rUxd+5cNGzYUHaUbGV1aOHNHhxtOLSgC/tNpsOHD2POnDmqosnDwwMjR45Eo0aNZEcD8Pr9zmlQvza83wDQrVs3HD58GL169cqyV/Hrr7+WlOxflpaWOH/+PNzc3NTab926hVq1auHp06dygr1FKwr5QhudIpmpqamIjo4WQrw+P/3SpUtCCCEiIyNFyZIlJSZT17BhQ9GnTx+RmpqqaktNTRW9e/cWjRo1kpgseydPnhTBwcFaMwAv08aNG4Wrq6tYsGCBOH78uLh48aLaRRvs379f1K9fXxw8eFA8evRIPHv2TO2iDZ4+fap2efjwoQgPDxd169YVv//+u+x4Qgjd2W9+/vlnYWBgILp06SLmz58v5s2bJ7p06SIMDQ3FunXrZMcTQghx4cIFtcuZM2fEjz/+KCpWrCi2bNkiO56KlZWVOHr0qOwYObK0tBTnz5/XaD979qwwNzeXkEjTmTNnRPHixUXp0qVFhw4dxGeffSbKlCkjbGxsxLlz5wotx3+mGClTpoyqAKlSpYpYv369EEKI48ePC0tLS5nR1CiVyiwnb7p69aowMTGRkEjdq1evRJ8+fVST92izzNHrb1607QwLXTkLJCuHDx8WNWrUkB1DCKH9+02mihUripCQEI324OBgrZmoKzs7d+4UTZo0kR1DxdnZWURGRsqOkaN27dqJzp07i7S0NFVbWlqa+Pzzz7XmjB9tKeT/M2NGGjVqhH379sHLywtdunTB119/jQMHDmDfvn1o1qyZ7HgqlpaWiImJ0RhDEBsbK/28fgAwNDTEtm3bMHbsWNlR3kkXlgOQtTZFQbC1tcX169dlxwCg/ftNptu3b8PHx0ejvX379vj2228lJMq98uXL48yZM7JjqEyePBnjxo3D6tWrtXYSuVmzZqFx48aoUKGC6jDckSNHkJiYiAMHDkhO99rZs2exfPlyGBj8Ww4YGBhg1KhRqFWrVqHl+M8UIwsXLkRKSgoAYMyYMTA0NMTRo0fRsWNHrfph9fX1hb+/P+bMmQNvb28oFAocPXoUI0eOVJshT6YOHTpg+/bteZrhVoa3JzvTRtpyymlOLl26pPa3EAJxcXGYMWOG1pztowv7DQA4Ojpi//79GmMI9u/fX6hnLuTk7UXUMt/vCRMmaAzElCk4OBhRUVGwt7eHs7OzxkR32rBgp6enJy5duoSFCxfi4sWLMDExgZ+fHwYPHqyaQE42bSnk/xPFSFpaGnbs2IFWrVoBeD1Aa9SoURg1apTkZJrmzJmjWjArLS0NwOveiIEDB2rNwFA3NzdMnjwZx48fR82aNWFmZqZ2/dChQyUlU7dmzZocr9eGRcmA1/9TWrZsGW7fvo1NmzahdOnS+Pnnn+Hi4qIVA1urVasGhUIB8dZY93r16mHVqlWSUqnThf0GAIKCgjB06FBcuHBBrWgKDQ3F/PnzZccDkPUiakIIODo64pdffpGUSpM2D0x/U6lSpaSfqZITbSnk/zNn05iamuLatWs68b9l4PW6C1FRURBCwM3NTau6IV1cXLK9TqFQaMUpdcDrpdDflJqaiuTkZBgZGcHU1FQrpq3fsmULevXqhR49euDnn39GZGQkXF1dsXjxYuzcuRO7d++WHVFjnQ89PT3Y2tpq5RT22rzfZNq2bRuCg4NVp51nnk2jDatIA5oTyGW+325ubmpd+ZS1S5cuoXLlytDT09PoVXxblSpVCilV9l69eoWRI0di6dKlWRbyhTVdw3+mGNGF0zzpw7t58yYGDhyIkSNHqnrKZKpevToCAwPh5+enOuXc1dUVFy5cQOvWrfHgwQPZEek/JDU1FV9++SXGjh2r1evnvOncuXNqp6RmtThmYdLT08ODBw9gZ2enOk06q59ZbZpxF5BfyP9nytyAgAAEBQXh3r17WR5akFmhduzYEaGhobC0tETHjh1z3FbbZorN3Mm0abHBnLi7u2PGjBno2bMn/vzzT9lxcP369SwnDbO0tJQ6B8EPP/yQ621lHZbTxf0mNjYWCoUCZcqUAfB66vL169fD09MTX375peR0ujVAPT4+Hl27dsWhQ4dgbW0NIQSePXuGpk2b4pdffoGtra2UXNHR0arH1oVB9JlMTU1RrFgxKBQKKT2K/5lixNfXF4D6F2dmxSq7QrWyslL9mFtaWurED/uaNWswe/Zs3Lx5E8DrkfYjR45Er169JCd7N319fdy/f192DACAg4MDbt26pTEt9NGjR6X+z3Tu3Lm52k6hUEgrRnRxv+nevTu+/PJL9OrVCw8ePEDz5s1RuXJlrF27Fg8ePMC4ceNkR9SZAepDhgxBYmIirl69Cg8PDwBAZGQkevfujaFDh2LDhg1Scr05FMDW1lYrDxW+KSMjA1OmTEFwcDBevHgBALCwsEBQUBC+++67HNdTKkj/mcM0bx/3fpuujCXRBiEhIRg7diwGDx6MBg0aQAiBY8eOYdGiRZgyZYrG0uOy/Prrr2p/Z54VsHDhQjg6OuK3336TlOxfs2bNwurVq7Fq1Sq0aNECu3fvxt27dxEYGIhx48Zh8ODBsiNSASpWrBhOnjyJChUq4IcffkBYWBiOHTuG8PBwDBgwQCvGW02dOhVz5sxBs2bNtHqAupWVFX7//XfUrl1brf306dNo2bKlVsxuam5ujs8++wy9evVCixYtCu2HPS/GjBmDlStXYuLEiWrf5xMmTMAXX3yBqVOnFk6QQpvRhHKladOm4smTJxrtz549E02bNi38QFlwdnYWq1ev1mgPDQ0Vzs7OEhJlLasJz+zt7UW3bt3E/fv3ZcdT+fbbb4WJiYkqp1KpFN9//73sWDpFF/YbIYQwMzNTzQTt4+MjZsyYIYQQ4u7du0KpVEpM9i9nZ+dsLy4uLrLjqZibm4uIiAiN9vPnzwsLC4vCD5SFLVu2iE6dOgkTExNhb28vhg4dKk6fPi07lhoHB4csZ8/evn27KFWqVKHl+M/0jADAzz//jKVLlyI6OhonTpyAk5MT5s2bBxcXF60Zyf7m4Kc3xcfHo3Tp0khNTZWU7F9KpRJXrlzRmCvh5s2b8PLyUs3nQrmXnJyMyMhIZGRkwNPTE+bm5rIjqbl37x5+/fXXLFe8DgkJkZTqX7qw3wBA3bp10bRpU7Rr1w4tW7bEyZMnUbVqVZw8eRKdOnXCvXv3ZEfUGZ9++imePn2KDRs2oFSpUgCAv/76Cz169ECxYsWwbds2yQn/9fz5c2zevBkbNmzAwYMH4eLigp49e2rFYTmlUolLly6hfPnyau3Xr19HtWrV8M8//xRKjv/MmJElS5Zg3LhxGDZsGKZOnaoaI2JtbY158+ZJL0bePAUsMjJS7SyK9PR07NmzB6VLl5YRTYObmxs2btyoMWNkWFiYVk2KpEtMTU0LdbbDvNi/fz/at28PFxcXXL9+HZUrV8adO3cghJC+AJ0u7TcAMHPmTHTo0AGzZ89G7969VZPG/frrr6hTp460XLkdH6JQKBAcHPyB0+TOwoUL8emnn8LZ2RmOjo5QKBSIiYmBl5cX1q5dKzueGgsLC/Tt2xd9+/ZFZGQkevTogYkTJ2pFMVK1alUsXLhQY9D6woULC3VSw/9Mz4inpyemTZuGzz77TO0UyitXruCjjz7Co0ePpOZ7c6XMrN4SExMTLFiwAP369SvsaBq2bNkCX19fNG/eHA0aNFBNkrN//35s3LgRHTp0kJYtL4PutOF/9CkpKViwYAEOHjyY5TLo2jCLZJ06ddC6dWtMmjRJte/Y2dmhR48eaN26NQYOHCgtmy7tN0IIxMTEoFixYkhPT1ebB+fOnTswNTXV6NkpLE2bNlX7+9y5c0hPT0eFChUAADdu3IC+vj5q1qypNdOYZ9q3bx/+/PNPCCHg6emJ5s2by46kISUlBb/++ivWr1+PPXv2wM7ODt26dcPMmTNlR8Phw4fRrl07lC1bFvXr14dCocDx48cRExOD3377rdBWk/7PFCMmJib4888/4eTkpFaM3Lx5E1WqVCm0rqjs3L17F0IIuLq64vTp02qnpRkZGcHOzg76+voSE6o7d+4c5s6di2vXrqm+BIKCgqSf469rX6rdu3fHvn370KlTJ9jb22ucETJ+/HhJyf5lYWGBCxcuoFy5cihWrBiOHj2KSpUq4eLFi/j0009x584dadl0ab/JyMiAUqnE1atXtboHMSQkBIcOHcLq1atVBdOTJ0/Qt29fNGrUCEFBQZIT6o7w8HCsW7cO27dvh76+Pjp16oQePXpo3TIQf/31F5YsWaL2fR4QEKA6/FUY/jOHaVxcXHDhwgWNs2Z+++03eHp6Skr1r8xcb//PWFvVrFlT67pCAfWF50JCQmBhYZHtl6o22LVrF3bv3o0GDRrIjpItMzMzvHz5EsDrqa2joqJQqVIlAJDeo6hL+42enh7c3d2RkJCg1cVIcHAwwsPD1XpuihUrhilTpqBly5ZSi5EffvgBX375JZRK5TvnwtGGs34+++wztGvXDqtXr0a7du001s/RFjY2Nmjfvj3q1aun2pfOnj0L4PUijoXhP1OMjBw5EoMGDUJKSgqEEDh9+jQ2bNiA6dOnY8WKFVKz/frrr2jTpg0MDQ01Tkd9W2F9MN4lPT0d27ZtU8186OHhgU8//VSrpovW5i/VTKVLl9aqVWWzUq9ePRw7dgyenp5o164dgoKCcPnyZWzduhX16tWTHU9NZGRkloNstWW/mTVrFkaOHIklS5agcuXKsuNkKTExEX///beq4MwUHx+P58+fS0r12ty5c9GjRw8olcoc58KROf/Nmx48eABLS0vZMXK0Z88e+Pn5ISEhQeNQZ6HOwVVo5+1ogR9//FGULVtWdQplmTJlxIoVK2THEgqFQvz999+qf2d30dPTk5z0tcuXLwtXV1dhamoqqlevLqpXry7MzMyEs7OzuHTpkux4Kubm5mL//v0a7fv37xfm5uYSEmnavXu3aN26tbhz547sKNmKiooSFy9eFEIIkZSUJAYOHCi8vLxEhw4dtCZ3VFSUqFKlimo/eXOf0Zb9RgghrK2thZGRkdDT0xNKpVIUK1ZM7aINevXqJcqWLSs2bdokYmNjRWxsrNi0aZNwdnYWfn5+suPpnFu3bonvvvtOdO3aVfU9/9tvv4krV65ITvZauXLlREBAgHjw4IHUHP+pYiTTw4cPVR8Kyru6desKHx8f8fjxY1Xb48ePRfv27UW9evUkJlOnC1+q8fHx4qOPPhJ6enrC3NxcK3+c+vTpI37//XeRkZEhO0q2PvnkE/Hpp5+K+Ph4YW5uLiIjI8WRI0dEnTp1xB9//CE7nkpoaGiOF22QWXAaGxurijkjIyMxcOBA8eLFC9nxVCZOnCiSkpI02pOTk8XEiRMlJNJ06NAhYWJiIpo3by6MjIxEVFSUEEKImTNnis8//1xyutcsLCzErVu3ZMf478wzMnHiRPTs2RPlypWTHUXnmZiY4OzZsxrduFeuXEHt2rWlDwbOlJycjBEjRmDVqlWqeSYMDAzg7++P2bNna8wsKUPz5s0RExMDf3//LAew9u7dW1Kyf7Vv3x7h4eGwsbFB165d0atXL1SrVk12LDUlSpTAgQMHUKVKFVhZWeH06dOoUKECDhw4gKCgIERERMiOqHOSkpLUFk7Thv3lTfr6+oiLi9M4AykhIQF2dnZasQhd/fr10blzZwwfPlztxIkzZ87gs88+w19//SU7Ivr164cGDRrA399fag7tOcD/gW3ZsgWTJk1C7dq10bNnT/j6+kpbSOlturAo2ZsqVKiQ7THltydCk8nU1BSLFy/G7NmztfZL9fjx4zhx4kShns+fV7/++iuePn2KjRs3Yv369Zg3bx4qVKiAnj17onv37hrr6siQnp6umiiuRIkSuH//PipUqAAnJydcv35dcjp1UVFR+OmnnxAVFYX58+fDzs4Oe/bsgaOjo8Y+JZOZmZlWLHGfHfH/64q97eLFiyhevLiERJouX76M9evXa7Tb2toiISFBQiJNCxcuROfOnXHkyBF4eXlpDLIttN8cqf0yhezKlStizJgxwsXFRRgaGoo2bdqIdevWZdnVV5jennLZzMxMKBQKVVe9QqEQZmZmWjMV865du0SlSpU0Dn94eXmJXbt2iWfPnqku2uDmzZtiz549Ijk5WQghtOpwQ/Xq1cWJEydkx8iT2NhYMWvWLFGxYkWhr68vO44QQoiGDRuKbdu2CSGE6Natm2jdurU4evSo8PPzE5UqVZIb7g260G2v7aytrUWxYsWEnp6e6t+ZF0tLS6GnpycCAgJkxxRCCFG6dGlx7NgxIcTrMWyZ7/fWrVuFq6urzGgqy5cvF/r6+sLc3Fw4OTlJm/7/P1WMvOno0aMiICBA2Nraas06BkIIsW7dOtGgQQPx559/qtr+/PNP0ahRI7F27VqJyf719qDarAYMasOA20ePHomPP/5YlSXzi6Bfv35i+PDhUrNl2rt3r/D29hYHDx4Ujx49UivktKWYe9OrV6/Etm3bxOeffy6USmWhrl2Rkz179ogtW7YIIV4PZvXw8BAKhUKUKFEiy0HMstSrV08EBwcLIdR/nE6fPq01r6W2Cw0NFT/99JNQKBRi/vz5amNu1q9fL44fPy47osrIkSNFw4YNRVxcnLCwsBA3b94UR48eFa6urmLChAmy4wkhhLC3txdTp04V6enpUnP8Z8aMvO3ChQtYu3YtfvnlFyQkJGjNOIdy5cph8+bNGpOHnTt3Dp06dUJ0dLSkZP86fPhwrreVObmPn58f4uPjsWLFCnh4eKiO14aHhyMwMBBXr16Vli1T5iqeb3c3i//vgtaG497A6/lb1q9fjy1btiA9PR0dO3ZEjx498PHHH2vlSqQA8PjxYxQrVizLrnxZzM3NcfnyZbi4uKiNIbhz5w4qVqzIdZ3y4PDhw/D29tbauTsAIDU1FX369MEvv/wCIQQMDAyQlpaGHj16IDQ0VCsm5CtevDjOnDkjfTzlf2bMCABER0dj/fr1WLduHW7cuIHGjRtjwoQJ6Ny5s+xoKnFxcVku6pWeno6///5bQiJN2jZ7YHbCw8Oxd+9elClTRq3d3d0dd+/elZRK3ZuTtGmrMmXKICEhAa1atcKyZcvg4+MDpVIpO5aa1atXo1OnTmrjgbRl3MCbrK2tERcXBxcXF7X2iIgIrVpDRxe8+T30zz//aHxvasP8HoaGhli3bh0mT56M8+fPIyMjA9WrV9eqSe969+6NsLAwjbXGCtt/phipX78+Tp06hSpVqqBv377o3r27Vu78zZo1wxdffIGVK1eiZs2aUCgUOHv2LL766iutWnMhJSUFly5dynI9FW2ZYCopKQmmpqYa7Y8ePYKxsbGERJp0obAbN24cOnfurDZ5nLYZMWIEAgIC4OPjg549e6J169ZaNQFfpu7du+Obb77Bpk2boFAokJGRgWPHjmHEiBHw8/OTHU+nJCcnY9SoUdi4cWOWg0Fl9Sq+a32skydPqv6tDetjpaenY9asWdi7dy+qVKmi0dNUWBn/M4dpvv32W/To0UN1+qSNjY3sSFl6+PAhevfujT179qg+FGlpaWjVqhVCQ0OlLaT1pswZ+7KaClybDi20a9cONWrUwOTJk2FhYYFLly7ByckJXbt2RUZGBjZv3iw7IgDgyJEjWLZsGW7fvo1NmzahdOnS+Pnnn+Hi4oKGDRvKjqcT0tLSsGfPHmzYsAH/+9//YGJigs6dO6Nnz57w9vaWHU9FF7rtdcWgQYNw8OBBTJo0CX5+fli0aBH++usvLFu2DDNmzECPHj2k5NK19bHezvsmhUJReBkljlcpNE+ePBEDBw4UNjY2qgGWNjY2YtCgQeLJkyey42Xp+vXr4n//+5/Yvn27uH79uuw4arRlxr53uXr1qrC1tRWtW7cWRkZGolOnTsLDw0PY29trxSQ/QgixefNmYWJiIvr37y+MjY1VAxoXLVok2rRpIzmdbkpKShJr164Vbdu2FUZGRlpz1sKboqKiRFhYmNi4caO4efOm7Dg6ydHRURw8eFAIIVSDQ4UQYs2aNVqz7wQHB2c5QeSnn34q5syZIzGZ9inyxUhCQoIoX768MDMzE19++aWYO3euCAkJEV988YUwMzMTFStWVPug0Ltpy4x9uREXFyfGjRsn2rVrJ9q0aSO+++47cf/+fdmxVKpVqyZWr14thFA/uyIiIkLY29vLjKbTHj58KBYsWCAqVaok/ayut61YsUJUqlRJGBkZCSMjI1GpUiWxfPly2bF0jpmZmWo5gtKlS4tTp04JIYS4ffu2MDMzkxlNpVSpUllO+3758mXh4OAgIZH20r6DqgVs0qRJMDIyQlRUFOzt7TWua9myJSZNmpTjokuFKT09HaGhodi/f3+W4zG0oVuvU6dOOHTokPTR17lRsmRJTJw4UXaMbF2/fh2NGzfWaLe0tMTTp08LP5AOS05OxrZt27Bu3Tr8/vvvcHR0RLdu3bBp0ybZ0VTGjh2LuXPnYsiQIahfvz4A4MSJEwgMDMSdO3cwZcoUyQl1R+ZZSE5OTvD09MTGjRtRp04d7NixA9bW1rLjAdDuRQe1TZEvRrZv345ly5ZpFCLA6x+qWbNmYcCAAVpTjHz99dcIDQ1Fu3btULlyZa06LTGT1szY9w4//fQTzM3NNc6W2rRpE5KTk7ViqnUHBwfcunVLYxbTo0ePwtXVVU4oHdStWzfs2LEDpqam6Ny5Mw4dOqRVY0UyLVmyBMuXL0e3bt1Ube3bt0eVKlUwZMgQFiN50LdvX1y8eBFNmjTBmDFj0K5dOyxYsABpaWlaMTAUADp06IC+ffsiODhYtcL1yZMnMXLkSHTs2FFyOi0ju2vmQzMyMhKxsbHZXh8bGyuMjY0LMVHObGxsxK5du2THyJG2zNj3LuXLlxcHDhzQaD906JAoX768hESaZs6cKTw9PcXJkyeFhYWFOHLkiFi7dq2wtbUVCxYskB1PZ3Tr1k3s3LlTpKamyo6SI2tra3Hjxg2N9uvXrwsrK6vCD1SE3L17V2zZskVcuHBBdhQVXVl0UBsU+WKkVKlS4siRI9le/8cff2jVzIcODg5aN2D1bdoyY9+7GBsbi+joaI326OhooVQqCz9QNr799lthYmKimsVWqVSK77//XnYsnfHq1Svx0Ucfaf1+I4QQgwcPFoGBgRrtQUFBWjOFuS7QpfdcCCFevHghLl68KC5cuMAiJBtF/tRef39/3Lp1C/v27YORkZHadS9fvkSrVq1Qrlw5rFy5UlJCdcHBwbh9+zYWLlyolYdoAO2Zse9dypYti4ULF2rMe/K///0PgwYNwr179yQley09PR1Hjx6Fl5cXlEolIiMjkZGRAU9PT9Wib5Q7tra2OH78uFZNJpWVIUOGYM2aNXB0dFTrto+NjYWfn5/aIU9tOdSgrXTlPafcKfLFyL1791CrVi0YGxtj0KBBqFixIgAgMjISixcvxsuXL3H27Fk4OjpKTvpahw4dcPDgQRQvXhyVKlXSGI+xdetWScn+FRgYCFtbW+kz9r1L5oRIP/30k2qQ6OHDh9GvXz906tQJc+bMkZwQUCqVuHbtmsaMnJQ3QUFBMDQ0xIwZM2RHyVFOczq8qVDnd9BRuvKeU+4U+QGsZcqUwYkTJxAQEIAxY8Ygs/ZSKBRo0aIFFi5cqDWFCPB6uugOHTrIjpEjbZmx712mTJmCu3fvolmzZqrZODMyMuDn54dp06ZJTveal5cXbt++zWLkPb169QorVqzAvn37UKtWLbVp4QHt+UzqwvT/ukJX3nPKnSLfM/KmJ0+e4ObNmwAANzc3rVy7QhdozYx9uXTjxg1cvHgRJiYm8PLygpOTk+xIKuHh4fjmm28wefJk1KxZU+MLVRvW19AFuvaZpPfH97xo+U8VI7rk4cOHuH79OhQKBcqXLw9bW1vZkXTamz1i2uTNFW/fzCa0bNVeIqIPqcgfptE1SUlJqkFumROe6evrw8/PDwsWLMhy4TdZbt26haioKDRu3BgmJiaqH1BtsmbNGsyePVvVI1a+fHmMHDkSvXr1kpzsNXbbFyxd+ExSweJ7XkTIOIWHsvfll18KV1dXsXv3bvHs2TPx7NkzsWvXLlGuXDkxYMAA2fGEEEI8evRIfPzxx0KhUAg9PT3VFOb9+vUTw4cPl5zuX8HBwcLU1FSMGjVKtc7PyJEjhampqQgJCZEdjwqQrnwmqeDwPS9aeJhGy5QoUQKbN2/GRx99pNZ+8OBBdOnSBQ8fPpQT7A1+fn6Ij4/HihUr4OHhgYsXL8LV1RXh4eEIDAzE1atXZUcEALi4uGDixIkaS7OvXr0aEyZMQHR0tKRk//rjjz9yvD6rqeJJk658Jqng8D0vWniYRsskJydnOXW9nZ0dkpOTJSTSFB4ejr1796JMmTJq7e7u7rh7966kVJri4uKynBLc29sbcXFxEhJpervoBNTHjnDMSO7oymeSCg7f86JF792bUGGqX78+xo8fj5SUFFXbP//8g4kTJ6oW1pItKSkpy7Erjx49grGxsYREWXNzc8PGjRs12sPCwrRmoqQnT56oXeLj47Fnzx7Url0b4eHhsuPpDF35TFLB4XtetLBnRMvMmzcPbdq0QZkyZVC1alUoFApcuHABxsbGWvPj1LhxY6xZswaTJ08G8Pp/8hkZGZg9e3auJ3UqDBMnToSvry/++OMPNGjQAAqFAkePHsX+/fuzLFJksLKy0mhr0aIFjI2NERgYiHPnzklIpXt05TNJBYfvedHCMSNa6J9//sHatWvx559/QggBT09P9OjRAyYmJrKjAXg9e+1HH32EmjVr4sCBA2jfvj2uXr2Kx48f49ixY1o1Tfz58+cREhKCa9euqV7LoKAgVK9eXXa0HF27dg21a9fGixcvZEfRCbr0maSCwfe8aGExomWmT58Oe3t79OvXT6191apVePjwIb755htJyf4VExMDAwMDLFu2DOfOnUNGRgZq1KiBQYMGITU1FWXLlpUdEampqfjyyy8xduxYuLq6yo6TrUuXLqn9LYRAXFwcZsyYgdTUVBw7dkxSMt3z4MEDLFmyROMz6eDgIDsafSB8z4sOFiNaxtnZGevXr9cYeHnq1Cl07dpVK84A0dfXR1xcHOzs7NTaExISYGdnpzWDLq2trXH+/HmtLkb09PSgUCjw9m5Yr149rFq1SrWWEhFRUcYxI1rmwYMHWVb1tra2WnMGSHb164sXL6BUKgs5TfY6dOiA7du3Y/jw4bKjZOvt4lJPTw+2trZa9Trqgj179sDc3BwNGzYEACxatAjLly+Hp6cnFi1ahGLFiklOSAXh7Z7EnFSpUuUDJqGCxmJEyzg6OuLYsWMaC6cdO3YMpUqVkpTqtcwfdYVCgXHjxqmNZE9PT8epU6dQrVo1Sek0ubm5YfLkyTh+/HiW674MHTpUUrJ/ZbVOztOnT1mM5NHIkSMxc+ZMAMDly5cxfPhwBAUF4cCBAxg+fDh++uknyQmpIFSrVk3Vk/j28gkAT4vXZSxGtEz//v0xbNgwpKam4uOPPwYA7N+/H6NGjUJQUJDUbBEREQBe7/iXL1+GkZGR6jojIyNUrVoVI0aMkBVPw4oVK2BtbY1z585pnJWiUCi0ohiZOXMmnJ2d4evrCwDo0qULNm/eDAcHB+zevRtVq1aVnFA3REdHw9PTEwCwZcsW+Pj4YNq0aTh//jzatm0rOR0VlDd7EiMiIjBixAiMHDlSNe3BiRMnEBwcjFmzZsmKSPlV+JO+Uk4yMjLEqFGjhFKpFHp6ekJPT0+YmpqKiRMnyo6m0qdPH/Hs2TPZMfIkIyNDZGRkyI6hwcXFRRw7dkwIIUR4eLiwtrYWe/fuFf7+/qJFixaS0+mOYsWKiatXrwohhGjQoIFYtmyZEEKI6OhoYWJiIjMafSC1a9cWu3bt0mjftWuXqFGjhoRE9D44gFVLvXjxAteuXYOJiQnc3d05iU8+rVy5EnPnzlUtlOfu7o5hw4ahf//+kpO9ZmJighs3bsDR0RFff/01UlJSsGzZMty4cQN169bFkydPZEfUCe3bt8erV6/QoEEDTJ48GdHR0ShdujTCw8MxePBg3LhxQ3ZEKmAmJiY4f/48PDw81NqvXbuGGjVq4J9//pGUjPKDM7BqKXNzc9SuXRuVK1dmIZJPY8eOxddffw0fHx9s2rQJmzZtgo+PDwIDA/H999/LjgcAKFasGGJjYwG8HoTZvHlzAK8PhfGYd+4tXLgQBgYG2Lx5M5YsWYLSpUsDAH777Te0bt1acjr6EDw8PDBlyhS12apfvnyJKVOmaBQopP3YM0JFVokSJbBgwQJ069ZNrX3Dhg0YMmQIHj16JCnZvwYPHoydO3fC3d0dERERuHPnDszNzREWFoaZM2fi/PnzsiMSaaXTp0/Dx8cHGRkZqrFVFy9ehEKhwM6dO1GnTh3JCSkvOICViqz09HTUqlVLo71mzZpIS0uTkEjT3Llz4ezsjNjYWMyaNQvm5uYAXi/yFxAQIDmdbklPT8e2bdtw7do1KBQKVKxYEZ999hkMDPg1VxTVqVMH0dHRarNV+/r6onv37hpnzpH2Y88IFVlDhgyBoaEhQkJC1NpHjBiBf/75B4sWLZKUjAralStX0L59e/z999+oUKECAODGjRuwtbXFr7/+Ci8vL8kJiSgnLEaoyBoyZAjWrFkDR0dH1KtXDwBw8uRJxMbGws/PD4aGhqpt3y5YCltkZCRiYmLw6tUrtfb27dtLSqRb6tWrBzs7O6xevVo1wdmTJ0/Qp08fxMfH48SJE5ITUkH49ddf0aZNGxgaGuLXX3/NcVvuO7qFxQgVWblduVOhUODAgQMfOE3Wbt++jQ4dOuDy5ctq08JnTt7EQay5Y2JigrNnz6JSpUpq7VeuXEHt2rV5ZkURoaenhwcPHsDOzg56etmff6FQKLjv6BgeTKUi6+DBg7IjvNPXX38NFxcX/P7773B1dcXp06eRkJCAoKAgzJkzR3Y8nVGhQgX8/fffGsVIfHw83NzcJKWigpaRkZHlv0n38dReIolOnDiBSZMmwdbWFnp6etDT00PDhg0xffp0rZghVpslJiaqLtOmTcPQoUOxefNm3Lt3D/fu3cPmzZsxbNgw1TTxVPQ9ffpUdgTKJ/aMEEmUnp6uOoOmRIkSuH//PipUqAAnJydcv35dcjrtZm1trbE+SZcuXVRtmYe8fHx82GVfBL29lELnzp2xZcsWLqWgo1iMEElUuXJlXLp0Ca6urqhbty5mzZoFIyMj/Pjjj3B1dZUdT6vpwmE4+nCWLVuGtWvXAgD27duH33//HXv27MHGjRsxcuRIhIeHS05IecEBrEQS7d27F0lJSejYsSNu376NTz75BH/++SdsbGwQFhamWiyR3u3p06dYuXKlap4RDw8P+Pv7w8rKSnY0+gC4lELRwmKESMs8fvwYxYoVUzsEQTk7e/YsWrduDaVSiTp16kAIgbNnz+Kff/5BeHg4atSoITsiFbBSpUph8+bN8Pb2RoUKFTBlyhR07twZ169fR+3atZGYmCg7IuUBD9MQaYFbt24hKioKjRs3RvHixcH/I+RNYGAgfHx8sHz5ctWMq2lpaejfvz+GDRuGP/74Q3JCKmgdO3ZE9+7d4e7ujoSEBLRp0wYAcOHCBZ5BpYPYM0IkUUJCArp06YKDBw9CoVDg5s2bcHV1hb+/P6ytrREcHCw7ok4wMTFBREQEKlasqNYeGRmJWrVqITk5WVIy+lBSU1Mxf/58xMbGok+fPqhevToAYN68eTA3N9ealbkpd9gzQiRRYGAgDA0NERMTo7bSqK+vLwIDA1mM5JKlpSViYmI0ipHY2FhYWFhISkUfkqGhIUaMGKHRPmzYsMIPQ++NxQiRROHh4di7dy/KlCmj1u7u7o67d+9KSqV7fH194e/vjzlz5sDb2xsKhQJHjx7FyJEjNVZtpqLjxo0bOHToEOLj4zUmQRs3bpykVJQfLEaIJEpKSoKpqalG+6NHj2BsbCwhkW6aM2cOFAoF/Pz8VCsyGxoaYuDAgZgxY4bkdPQhLF++HAMHDkSJEiVQsmRJtQHfCoWCxYiO4ZgRIonatWuHGjVqYPLkybCwsMClS5fg5OSErl27IiMjA5s3b5YdUackJycjKioKQgi4ubllWehR0eDk5ISAgAB88803sqNQAWAxQiRRZGQkPvroI9SsWRMHDhxA+/btcfXqVTx+/BjHjh1DuXLlZEck0kqWlpa4cOECJwcsIrg2DZFEnp6euHjxIurUqYMWLVqoJkCLiIhgIUKUg86dO3OW1SKEPSNEkqWkpODSpUtZDsJr3769pFRE2m369OkICQlBu3bt4OXlBUNDQ7XrudCkbmExQiTRnj174Ofnh4SEBI2JzhQKBRd4I8qGi4tLttcpFArcvn27ENPQ+2IxQiSRm5sbWrVqhXHjxsHe3l52HCIiKViMEElkaWnJ8SFE7ynzZ4zrOekuDmAlkqhTp044dOiQ7BhEOmnNmjXw8vKCiYkJTExMUKVKFfz888+yY1E+sGeESKLk5GR07twZtra2HIRHlAchISEYO3YsBg8ejAYNGkAIgWPHjmHRokWYMmUKAgMDZUekPGAxQiTRihUrMGDAAJiYmMDGxkZjFkkOwiPKmouLCyZOnAg/Pz+19tWrV2PChAmIjo6WlIzyg8UIkUQlS5bE0KFDMXr0aOjp8agpUW4plUpcuXIFbm5uau03b96El5cXUlJSJCWj/OC3H5FEr169gq+vLwsRojxyc3PDxo0bNdrDwsLg7u4uIRG9D/aMEEkUGBgIW1tbfPvtt7KjEOmULVu2wNfXF82bN0eDBg1UKzXv378fGzduRIcOHWRHpDxgMUIk0dChQ7FmzRpUrVoVVapU0RjAGhISIikZkfY7d+4c5s6di2vXrkEIAU9PTwQFBaF69eqyo1EesRghkqhp06bZXqdQKHDgwIFCTENEJAeLESIi0jm7d++Gvr4+WrVqpda+d+9eZGRkoE2bNpKSUX5w1BwREemc0aNHZ7l2kxACo0ePlpCI3geLESIi0jk3b96Ep6enRnvFihVx69YtCYnofbAYISIinWNlZZXlpIC3bt2CmZmZhET0PliMEBGRzmnfvj2GDRuGqKgoVdutW7cQFBSE9u3bS0xG+cEBrEREpHOePXuG1q1b4+zZsyhTpgwA4N69e2jUqBG2bt0Ka2truQEpT1iMEBGRThJCYN++fbh48aJq1d7GjRvLjkX5wGKEiIiKLC8vL+zevRuOjo6yo1AOOGaEiIiKrDt37iA1NVV2DHoHFiNEREQkFYsRIiIikorFCBEREUnFYoSIiIikYjFCREREUrEYISIinRMdHZ2r7ZYtWwZ7e/sPnIbeF+cZISIinaOvr4/GjRvD398fnTp1glKplB2J3gN7RoiISOdcvHgR1atXR1BQEEqWLImvvvoKp0+flh2L8ok9I0REpLPS0tKwY8cOhIaG4rfffoO7uzv8/f3Rq1cv2Nrayo5HucRihIiIdN7Lly+xePFijBkzBq9evYKhoSF8fX0xc+ZMODg4yI5H78DDNEREpLPOnj2LgIAAODg4ICQkBCNGjEBUVBQOHDiAv/76C59++qnsiJQL7BkhIiKdExISgp9++gnXr19H27Zt0b9/f7Rt2xZ6ev/+H/vWrVuoWLEi0tLSJCal3GAxQkREOsfd3R39+vVD3759UbJkySy3efXqFTZs2IDevXsXcjrKKxYjREREJBXHjBARkc756aefsGnTJo32TZs2YfXq1RIS0ftgMUJERDpnxowZKFGihEa7nZ0dpk2bJiERvQ8WI0REpHPu3r0LFxcXjXYnJyfExMRISETvg8UIERHpHDs7O1y6dEmj/eLFi7CxsZGQiN4HixEiItI5Xbt2xdChQ3Hw4EGkp6cjPT0dBw4cwNdff42uXbvKjkd5xLNpiIhI57x69Qq9evXCpk2bYGBgAADIyMiAn58fli5dCiMjI8kJKS9YjBARkc66ceMGLl68CBMTE3h5ecHJyUl2JMoHFiNEREQklYHsAERERHmVnp6O0NBQ7N+/H/Hx8cjIyFC7/sCBA5KSUX6wGCEiIp3z9ddfIzQ0FO3atUPlypWhUChkR6L3wMM0RESkc0qUKIE1a9agbdu2sqNQAeCpvUREpHOMjIzg5uYmOwYVEBYjRESkc4KCgjB//nywc79o4GEaIiLSOR06dMDBgwdRvHhxVKpUCYaGhmrXb926VVIyyg8OYCUiIp1jbW2NDh06yI5BBYQ9I0RERCQVx4wQERGRVDxMQ0REOmnz5s3YuHEjYmJi8OrVK7Xrzp8/LykV5Qd7RoiISOf88MMP6Nu3L+zs7BAREYE6derAxsYGt2/fRps2bWTHozzimBEiItI5FStWxPjx49GtWzdYWFjg4sWLcHV1xbhx4/D48WMsXLhQdkTKA/aMEBGRzomJiYG3tzcAwMTEBM+fPwcA9OrVCxs2bJAZjfKBxQgREemckiVLIiEhAQDg5OSEkydPAgCio6M5EZoOYjFCREQ65+OPP8aOHTsAAP7+/ggMDESLFi3g6+vL+Ud0EMeMEBGRzsnIyEBGRgYMDF6fFLpx40YcPXoUbm5uGDBgAIyMjCQnpLxgMUJERERScZ4RIiLSSU+fPsXp06cRHx+PjIwMtev8/PwkpaL8YM8IERHpnB07dqBHjx5ISkqChYUFFAqF6jqFQoHHjx9LTEd5xWKEiIh0Tvny5dG2bVtMmzYNpqamsuPQe2IxQkREOsfMzAyXL1+Gq6ur7ChUAHhqLxER6ZxWrVrh7NmzsmNQAeEAViIi0jnt2rXDyJEjERkZCS8vLxgaGqpd3759e0nJKD94mIaIiHSOnl72HfsKhQLp6emFmIbeF4sRIiIikopjRoiIiEgqjhkhIiKd88MPP2TZrlAooFQq4ebmhsaNG0NfX7+Qk1F+8DANERHpHBcXFzx8+BDJyckoVqwYhBB4+vQpTE1NYW5ujvj4eLi6uuLgwYNwdHSUHZfegYdpiIhI50ybNg21a9fGzZs3kZCQgMePH+PGjRuoW7cu5s+fj5iYGJQsWRKBgYGyo1IusGeEiIh0Trly5bBlyxZUq1ZNrT0iIgKff/45bt++jePHj+Pzzz9HXFycnJCUa+wZISIinRMXF4e0tDSN9rS0NDx48AAAUKpUKTx//rywo1E+sBghIiKd07RpU3z11VeIiIhQtUVERGDgwIH4+OOPAQCXL1+Gi4uLrIiUByxGiIhI56xcuRLFixdHzZo1YWxsDGNjY9SqVQvFixfHypUrAQDm5uYIDg6WnJRyg2NGiIhIZ/3555+4ceMGhBCoWLEiKlSoIDsS5QPnGSEiIp3l6uoKhUKBcuXKwcCAP2m6iodpiIhI5yQnJ8Pf3x+mpqaoVKkSYmJiAABDhw7FjBkzJKejvGIxQkREOmfMmDG4ePEiDh06BKVSqWpv3rw5wsLCJCaj/GCfFhER6Zzt27cjLCwM9erVg0KhULV7enoiKipKYjLKD/aMEBGRznn48CHs7Ow02pOSktSKE9INLEaIiEjn1K5dG7t27VL9nVmALF++HPXr15cVi/KJh2mIiEjnTJ8+Ha1bt0ZkZCTS0tIwf/58XL16FSdOnMDhw4dlx6M8Ys8IERHpHG9vbxw/fhzJyckoV64cwsPDYW9vjxMnTqBmzZqy41EecdIzIiLSKampqfjyyy8xduxYuLq6yo5DBYA9I0REpFMMDQ2xbds22TGoALEYISIindOhQwds375ddgwqIBzASkREOsfNzQ2TJ0/G8ePHUbNmTZiZmaldP3ToUEnJKD84ZoSIiHSOi4tLttcpFArcvn27ENPQ+2IxQkREOi3zZ4yTnekujhkhIiKdtHLlSlSuXBlKpRJKpRKVK1fGihUrZMeifOCYESIi0jljx47F3LlzMWTIENWMqydOnEBgYCDu3LmDKVOmSE5IecHDNEREpHNKlCiBBQsWoFu3bmrtGzZswJAhQ/Do0SNJySg/eJiGiIh0Tnp6OmrVqqXRXrNmTaSlpUlIRO+DxQgREemcnj17YsmSJRrtP/74I3r06CEhEb0PHqYhIiKdM2TIEKxZswaOjo6oV68eAODkyZOIjY2Fn58fDA0NVduGhITIikm5xGKEiIh0TtOmTXO1nUKhwIEDBz5wGnpfLEaIiIhIKo4ZISIiIqlYjBAREZFULEaIiIhIKhYjREREJBWLESIiIpKKxQgRERFJxWKEiIiIpPo/txt9uOfOcr0AAAAASUVORK5CYII=",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "eval_df_roberta.plot(x='type', y=['f1_strict', 'f1_relaxed'], title= f'Strict and Relax F1_Score for Roberta base model on Chia', kind=\"bar\", rot=90)\n",
+    "plt.xlabel(\"\")\n",
+    "plt.show()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "honlp",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}