Spaces:

Zalimannard
/

TimeSeriesHomework1_2

No application file

App Files Files Community

Kolesnikov Dmitry commited on Nov 21, 2025

Commit

b34a74f

1 Parent(s): eaf6e74

feat: Sarima и VAR для третей лабораторки

Browse files

Files changed (8) hide show

_lab3_input_tmp.csv +821 -0
requirements.txt +12 -6
src/lab3_functions.py +192 -0
src/lab3_pipeline.py +272 -0
src/lab4_functions.py +43 -0
src/lab5_functions.py +48 -0
src/streamlit_app.py +358 -2
src/streamlit_lab3.py +50 -0

_lab3_input_tmp.csv ADDED Viewed

	@@ -0,0 +1,821 @@

+index,timestamp,DailyNewCases,ActiveCases,DailyNewDeaths
+0,2020-02-15 00:00:00+03:00,0.0,0.0,0.0
+1,2020-02-16 00:00:00+03:00,0.0,0.0,0.0
+2,2020-02-17 00:00:00+03:00,0.0,0.0,0.0
+3,2020-02-18 00:00:00+03:00,0.0,0.0,0.0
+4,2020-02-19 00:00:00+03:00,0.0,0.0,0.0
+5,2020-02-20 00:00:00+03:00,0.0,0.0,0.0
+6,2020-02-21 00:00:00+03:00,0.0,0.0,0.0
+7,2020-02-22 00:00:00+03:00,0.0,0.0,0.0
+8,2020-02-23 00:00:00+03:00,0.0,0.0,0.0
+9,2020-02-24 00:00:00+03:00,0.0,0.0,0.0
+10,2020-02-25 00:00:00+03:00,0.0,0.0,0.0
+11,2020-02-26 00:00:00+03:00,0.0,0.0,0.0
+12,2020-02-27 00:00:00+03:00,0.0,0.0,0.0
+13,2020-02-28 00:00:00+03:00,0.0,0.0,0.0
+14,2020-02-29 00:00:00+03:00,0.0,0.0,0.0
+15,2020-03-01 00:00:00+03:00,0.0,0.0,0.0
+16,2020-03-02 00:00:00+03:00,1.0,1.0,0.0
+17,2020-03-03 00:00:00+03:00,0.0,1.0,0.0
+18,2020-03-04 00:00:00+03:00,0.0,1.0,0.0
+19,2020-03-05 00:00:00+03:00,4.0,5.0,0.0
+20,2020-03-06 00:00:00+03:00,6.0,11.0,0.0
+21,2020-03-07 00:00:00+03:00,1.0,11.0,0.0
+22,2020-03-08 00:00:00+03:00,3.0,14.0,0.0
+23,2020-03-09 00:00:00+03:00,3.0,17.0,0.0
+24,2020-03-10 00:00:00+03:00,0.0,17.0,0.0
+25,2020-03-11 00:00:00+03:00,8.0,25.0,0.0
+26,2020-03-12 00:00:00+03:00,6.0,31.0,0.0
+27,2020-03-13 00:00:00+03:00,11.0,37.0,0.0
+28,2020-03-14 00:00:00+03:00,14.0,51.0,0.0
+29,2020-03-15 00:00:00+03:00,4.0,55.0,0.0
+30,2020-03-16 00:00:00+03:00,30.0,85.0,0.0
+31,2020-03-17 00:00:00+03:00,21.0,106.0,0.0
+32,2020-03-18 00:00:00+03:00,33.0,139.0,0.0
+33,2020-03-19 00:00:00+03:00,52.0,190.0,1.0
+34,2020-03-20 00:00:00+03:00,54.0,240.0,0.0
+35,2020-03-21 00:00:00+03:00,53.0,289.0,0.0
+36,2020-03-22 00:00:00+03:00,61.0,350.0,0.0
+37,2020-03-23 00:00:00+03:00,71.0,420.0,0.0
+38,2020-03-24 00:00:00+03:00,57.0,472.0,0.0
+39,2020-03-25 00:00:00+03:00,163.0,626.0,2.0
+40,2020-03-26 00:00:00+03:00,182.0,799.0,0.0
+41,2020-03-27 00:00:00+03:00,196.0,987.0,1.0
+42,2020-03-28 00:00:00+03:00,228.0,1211.0,0.0
+43,2020-03-29 00:00:00+03:00,270.0,1462.0,4.0
+44,2020-03-30 00:00:00+03:00,302.0,1761.0,1.0
+45,2020-03-31 00:00:00+03:00,501.0,2199.0,8.0
+46,2020-04-01 00:00:00+03:00,440.0,2563.0,7.0
+47,2020-04-02 00:00:00+03:00,771.0,3283.0,6.0
+48,2020-04-03 00:00:00+03:00,601.0,3834.0,4.0
+49,2020-04-04 00:00:00+03:00,582.0,4355.0,9.0
+50,2020-04-05 00:00:00+03:00,658.0,4989.0,2.0
+51,2020-04-06 00:00:00+03:00,954.0,5890.0,2.0
+52,2020-04-07 00:00:00+03:00,1154.0,6945.0,11.0
+53,2020-04-08 00:00:00+03:00,1175.0,8029.0,5.0
+54,2020-04-09 00:00:00+03:00,1459.0,9357.0,13.0
+55,2020-04-10 00:00:00+03:00,1786.0,11028.0,18.0
+56,2020-04-11 00:00:00+03:00,1667.0,12433.0,12.0
+57,2020-04-12 00:00:00+03:00,2186.0,14349.0,24.0
+58,2020-04-13 00:00:00+03:00,2558.0,16710.0,18.0
+59,2020-04-14 00:00:00+03:00,2774.0,19238.0,22.0
+60,2020-04-15 00:00:00+03:00,3388.0,22306.0,28.0
+61,2020-04-16 00:00:00+03:00,3448.0,25402.0,34.0
+62,2020-04-17 00:00:00+03:00,4070.0,29145.0,41.0
+63,2020-04-18 00:00:00+03:00,4785.0,33423.0,40.0
+64,2020-04-19 00:00:00+03:00,6060.0,39201.0,48.0
+65,2020-04-20 00:00:00+03:00,4268.0,43270.0,44.0
+66,2020-04-21 00:00:00+03:00,5642.0,48434.0,51.0
+67,2020-04-22 00:00:00+03:00,5236.0,53066.0,57.0
+68,2020-04-23 00:00:00+03:00,4774.0,57327.0,42.0
+69,2020-04-24 00:00:00+03:00,5849.0,62439.0,60.0
+70,2020-04-25 00:00:00+03:00,5966.0,67657.0,66.0
+71,2020-04-26 00:00:00+03:00,6361.0,73435.0,66.0
+72,2020-04-27 00:00:00+03:00,6198.0,79007.0,47.0
+73,2020-04-28 00:00:00+03:00,6411.0,84235.0,73.0
+74,2020-04-29 00:00:00+03:00,5841.0,88141.0,105.0
+75,2020-04-30 00:00:00+03:00,7099.0,93806.0,101.0
+76,2020-05-01 00:00:00+03:00,7933.0,100042.0,96.0
+77,2020-05-02 00:00:00+03:00,9623.0,107819.0,53.0
+78,2020-05-03 00:00:00+03:00,10633.0,116768.0,58.0
+79,2020-05-04 00:00:00+03:00,10581.0,125817.0,76.0
+80,2020-05-05 00:00:00+03:00,10102.0,134054.0,95.0
+81,2020-05-06 00:00:00+03:00,10559.0,143065.0,86.0
+82,2020-05-07 00:00:00+03:00,11231.0,151732.0,88.0
+83,2020-05-08 00:00:00+03:00,10699.0,159528.0,98.0
+84,2020-05-09 00:00:00+03:00,10817.0,164933.0,104.0
+85,2020-05-10 00:00:00+03:00,11012.0,173467.0,88.0
+86,2020-05-11 00:00:00+03:00,11656.0,179534.0,94.0
+87,2020-05-12 00:00:00+03:00,10899.0,186615.0,107.0
+88,2020-05-13 00:00:00+03:00,10028.0,192056.0,96.0
+89,2020-05-14 00:00:00+03:00,9974.0,196410.0,93.0
+90,2020-05-15 00:00:00+03:00,10598.0,202199.0,113.0
+91,2020-05-16 00:00:00+03:00,9200.0,206340.0,119.0
+92,2020-05-17 00:00:00+03:00,9709.0,211748.0,94.0
+93,2020-05-18 00:00:00+03:00,8926.0,217747.0,91.0
+94,2020-05-19 00:00:00+03:00,9263.0,220974.0,115.0
+95,2020-05-20 00:00:00+03:00,8764.0,220341.0,135.0
+96,2020-05-21 00:00:00+03:00,8849.0,221774.0,127.0
+97,2020-05-22 00:00:00+03:00,8894.0,223374.0,150.0
+98,2020-05-23 00:00:00+03:00,9434.0,224558.0,139.0
+99,2020-05-24 00:00:00+03:00,8599.0,227641.0,153.0
+100,2020-05-25 00:00:00+03:00,8946.0,230996.0,92.0
+101,2020-05-26 00:00:00+03:00,8915.0,227406.0,174.0
+102,2020-05-27 00:00:00+03:00,8338.0,224504.0,161.0
+103,2020-05-28 00:00:00+03:00,8371.0,223916.0,174.0
+104,2020-05-29 00:00:00+03:00,8572.0,223992.0,232.0
+105,2020-05-30 00:00:00+03:00,8952.0,224551.0,181.0
+106,2020-05-31 00:00:00+03:00,9268.0,229267.0,138.0
+107,2020-06-01 00:00:00+03:00,9035.0,234146.0,162.0
+108,2020-06-02 00:00:00+03:00,8863.0,231719.0,182.0
+109,2020-06-03 00:00:00+03:00,8536.0,231105.0,178.0
+110,2020-06-04 00:00:00+03:00,8831.0,231101.0,169.0
+111,2020-06-05 00:00:00+03:00,8726.0,231626.0,144.0
+112,2020-06-06 00:00:00+03:00,8855.0,231576.0,197.0
+113,2020-06-07 00:00:00+03:00,8984.0,235083.0,134.0
+114,2020-06-08 00:00:00+03:00,8985.0,239999.0,112.0
+115,2020-06-09 00:00:00+03:00,8595.0,236714.0,171.0
+116,2020-06-10 00:00:00+03:00,8404.0,234516.0,216.0
+117,2020-06-11 00:00:00+03:00,8779.0,234754.0,174.0
+118,2020-06-12 00:00:00+03:00,8987.0,235338.0,183.0
+119,2020-06-13 00:00:00+03:00,8706.0,238659.0,114.0
+120,2020-06-14 00:00:00+03:00,8835.0,241966.0,119.0
+121,2020-06-15 00:00:00+03:00,8246.0,245580.0,143.0
+122,2020-06-16 00:00:00+03:00,8248.0,243868.0,193.0
+123,2020-06-17 00:00:00+03:00,7843.0,241481.0,194.0
+124,2020-06-18 00:00:00+03:00,7790.0,239468.0,182.0
+125,2020-06-19 00:00:00+03:00,7972.0,236816.0,181.0
+126,2020-06-20 00:00:00+03:00,7889.0,234358.0,161.0
+127,2020-06-21 00:00:00+03:00,7728.0,236858.0,109.0
+128,2020-06-22 00:00:00+03:00,7600.0,239658.0,95.0
+129,2020-06-23 00:00:00+03:00,7425.0,234917.0,153.0
+130,2020-06-24 00:00:00+03:00,7176.0,229546.0,154.0
+131,2020-06-25 00:00:00+03:00,7113.0,230225.0,92.0
+132,2020-06-26 00:00:00+03:00,6800.0,227861.0,176.0
+133,2020-06-27 00:00:00+03:00,6852.0,225325.0,188.0
+134,2020-06-28 00:00:00+03:00,6791.0,226277.0,104.0
+135,2020-06-29 00:00:00+03:00,6719.0,228560.0,93.0
+136,2020-06-30 00:00:00+03:00,6693.0,225879.0,154.0
+137,2020-07-01 00:00:00+03:00,6556.0,221938.0,216.0
+138,2020-07-02 00:00:00+03:00,6760.0,222504.0,147.0
+139,2020-07-03 00:00:00+03:00,6718.0,220131.0,176.0
+140,2020-07-04 00:00:00+03:00,6632.0,217609.0,168.0
+141,2020-07-05 00:00:00+03:00,6736.0,220340.0,134.0
+142,2020-07-06 00:00:00+03:00,6611.0,223237.0,135.0
+143,2020-07-07 00:00:00+03:00,6368.0,219856.0,198.0
+144,2020-07-08 00:00:00+03:00,6562.0,217614.0,173.0
+145,2020-07-09 00:00:00+03:00,6509.0,215142.0,176.0
+146,2020-07-10 00:00:00+03:00,6635.0,213851.0,174.0
+147,2020-07-11 00:00:00+03:00,6611.0,211896.0,188.0
+148,2020-07-12 00:00:00+03:00,6615.0,214766.0,130.0
+149,2020-07-13 00:00:00+03:00,6537.0,218239.0,104.0
+150,2020-07-14 00:00:00+03:00,6248.0,215508.0,175.0
+151,2020-07-15 00:00:00+03:00,6422.0,211350.0,156.0
+152,2020-07-16 00:00:00+03:00,6428.0,209168.0,167.0
+153,2020-07-17 00:00:00+03:00,6406.0,207707.0,186.0
+154,2020-07-18 00:00:00+03:00,6234.0,206327.0,124.0
+155,2020-07-19 00:00:00+03:00,6109.0,208860.0,95.0
+156,2020-07-20 00:00:00+03:00,5940.0,211457.0,85.0
+157,2020-07-21 00:00:00+03:00,5842.0,208364.0,153.0
+158,2020-07-22 00:00:00+03:00,5862.0,204392.0,165.0
+159,2020-07-23 00:00:00+03:00,5848.0,201816.0,147.0
+160,2020-07-24 00:00:00+03:00,5811.0,199029.0,154.0
+161,2020-07-25 00:00:00+03:00,5871.0,196388.0,146.0
+162,2020-07-26 00:00:00+03:00,5765.0,198966.0,77.0
+163,2020-07-27 00:00:00+03:00,5635.0,201437.0,85.0
+164,2020-07-28 00:00:00+03:00,5395.0,197794.0,150.0
+165,2020-07-29 00:00:00+03:00,5475.0,194984.0,169.0
+166,2020-07-30 00:00:00+03:00,5509.0,191042.0,129.0
+167,2020-07-31 00:00:00+03:00,5482.0,187608.0,161.0
+168,2020-08-01 00:00:00+03:00,5462.0,184861.0,95.0
+169,2020-08-02 00:00:00+03:00,5427.0,186569.0,70.0
+170,2020-08-03 00:00:00+03:00,5394.0,188464.0,79.0
+171,2020-08-04 00:00:00+03:00,5159.0,185601.0,144.0
+172,2020-08-05 00:00:00+03:00,5204.0,183111.0,139.0
+173,2020-08-06 00:00:00+03:00,5267.0,180931.0,116.0
+174,2020-08-07 00:00:00+03:00,5241.0,178818.0,119.0
+175,2020-08-08 00:00:00+03:00,5212.0,177286.0,129.0
+176,2020-08-09 00:00:00+03:00,5189.0,179183.0,77.0
+177,2020-08-10 00:00:00+03:00,5118.0,180972.0,70.0
+178,2020-08-11 00:00:00+03:00,4945.0,179293.0,130.0
+179,2020-08-12 00:00:00+03:00,5102.0,177143.0,129.0
+180,2020-08-13 00:00:00+03:00,5057.0,175978.0,124.0
+181,2020-08-14 00:00:00+03:00,5065.0,174361.0,114.0
+182,2020-08-15 00:00:00+03:00,5061.0,172856.0,119.0
+183,2020-08-16 00:00:00+03:00,4969.0,174200.0,68.0
+184,2020-08-17 00:00:00+03:00,4892.0,175904.0,55.0
+185,2020-08-18 00:00:00+03:00,4748.0,173993.0,132.0
+186,2020-08-19 00:00:00+03:00,4828.0,171909.0,117.0
+187,2020-08-20 00:00:00+03:00,4785.0,170494.0,110.0
+188,2020-08-21 00:00:00+03:00,4870.0,169457.0,90.0
+189,2020-08-22 00:00:00+03:00,4921.0,168110.0,121.0
+190,2020-08-23 00:00:00+03:00,4852.0,169727.0,73.0
+191,2020-08-24 00:00:00+03:00,4744.0,171950.0,65.0
+192,2020-08-25 00:00:00+03:00,4696.0,169874.0,120.0
+193,2020-08-26 00:00:00+03:00,4676.0,168032.0,115.0
+194,2020-08-27 00:00:00+03:00,4711.0,166211.0,121.0
+195,2020-08-28 00:00:00+03:00,4829.0,165025.0,110.0
+196,2020-08-29 00:00:00+03:00,4941.0,163938.0,111.0
+197,2020-08-30 00:00:00+03:00,4980.0,166251.0,68.0
+198,2020-08-31 00:00:00+03:00,4993.0,168756.0,83.0
+199,2020-09-01 00:00:00+03:00,4729.0,167044.0,123.0
+200,2020-09-02 00:00:00+03:00,4952.0,166417.0,115.0
+201,2020-09-03 00:00:00+03:00,4995.0,165532.0,114.0
+202,2020-09-04 00:00:00+03:00,5110.0,164709.0,121.0
+203,2020-09-05 00:00:00+03:00,5205.0,164425.0,110.0
+204,2020-09-06 00:00:00+03:00,5195.0,166736.0,61.0
+205,2020-09-07 00:00:00+03:00,5185.0,169542.0,51.0
+206,2020-09-08 00:00:00+03:00,5099.0,167747.0,122.0
+207,2020-09-09 00:00:00+03:00,5218.0,166414.0,142.0
+208,2020-09-10 00:00:00+03:00,5363.0,165734.0,128.0
+209,2020-09-11 00:00:00+03:00,5504.0,165402.0,102.0
+210,2020-09-12 00:00:00+03:00,5488.0,165343.0,119.0
+211,2020-09-13 00:00:00+03:00,5449.0,168008.0,94.0
+212,2020-09-14 00:00:00+03:00,5509.0,170985.0,57.0
+213,2020-09-15 00:00:00+03:00,5529.0,170759.0,150.0
+214,2020-09-16 00:00:00+03:00,5670.0,170488.0,132.0
+215,2020-09-17 00:00:00+03:00,5762.0,170352.0,144.0
+216,2020-09-18 00:00:00+03:00,5905.0,170784.0,134.0
+217,2020-09-19 00:00:00+03:00,6065.0,171450.0,144.0
+218,2020-09-20 00:00:00+03:00,6148.0,174624.0,79.0
+219,2020-09-21 00:00:00+03:00,6196.0,178133.0,71.0
+220,2020-09-22 00:00:00+03:00,6215.0,178212.0,160.0
+221,2020-09-23 00:00:00+03:00,6431.0,178743.0,150.0
+222,2020-09-24 00:00:00+03:00,6595.0,179059.0,149.0
+223,2020-09-25 00:00:00+03:00,7212.0,181846.0,108.0
+224,2020-09-26 00:00:00+03:00,7523.0,183196.0,169.0
+225,2020-09-27 00:00:00+03:00,7867.0,187896.0,99.0
+226,2020-09-28 00:00:00+03:00,8135.0,193268.0,61.0
+227,2020-09-29 00:00:00+03:00,8232.0,194861.0,160.0
+228,2020-09-30 00:00:00+03:00,8481.0,197307.0,177.0
+229,2020-10-01 00:00:00+03:00,8945.0,200098.0,169.0
+230,2020-10-02 00:00:00+03:00,9412.0,203270.0,186.0
+231,2020-10-03 00:00:00+03:00,9859.0,207392.0,174.0
+232,2020-10-04 00:00:00+03:00,10499.0,214500.0,107.0
+233,2020-10-05 00:00:00+03:00,10888.0,222090.0,117.0
+234,2020-10-06 00:00:00+03:00,11615.0,227265.0,188.0
+235,2020-10-07 00:00:00+03:00,11115.0,231479.0,202.0
+236,2020-10-08 00:00:00+03:00,11493.0,235727.0,191.0
+237,2020-10-09 00:00:00+03:00,12126.0,240560.0,201.0
+238,2020-10-10 00:00:00+03:00,12846.0,246428.0,197.0
+239,2020-10-11 00:00:00+03:00,13634.0,255679.0,143.0
+240,2020-10-12 00:00:00+03:00,13592.0,265353.0,125.0
+241,2020-10-13 00:00:00+03:00,13868.0,271427.0,244.0
+242,2020-10-14 00:00:00+03:00,14231.0,277499.0,239.0
+243,2020-10-15 00:00:00+03:00,13754.0,282575.0,286.0
+244,2020-10-16 00:00:00+03:00,15150.0,289008.0,232.0
+245,2020-10-17 00:00:00+03:00,14922.0,295034.0,279.0
+246,2020-10-18 00:00:00+03:00,15099.0,304571.0,185.0
+247,2020-10-19 00:00:00+03:00,15982.0,315046.0,179.0
+248,2020-10-20 00:00:00+03:00,16319.0,321392.0,269.0
+249,2020-10-21 00:00:00+03:00,15700.0,325823.0,317.0
+250,2020-10-22 00:00:00+03:00,15971.0,330076.0,290.0
+251,2020-10-23 00:00:00+03:00,17340.0,335870.0,283.0
+252,2020-10-24 00:00:00+03:00,16521.0,340528.0,296.0
+253,2020-10-25 00:00:00+03:00,16710.0,349305.0,229.0
+254,2020-10-26 00:00:00+03:00,17347.0,358859.0,219.0
+255,2020-10-27 00:00:00+03:00,16550.0,362245.0,320.0
+256,2020-10-28 00:00:00+03:00,16202.0,365740.0,346.0
+257,2020-10-29 00:00:00+03:00,17717.0,368351.0,366.0
+258,2020-10-30 00:00:00+03:00,18283.0,371760.0,355.0
+259,2020-10-31 00:00:00+03:00,18140.0,374712.0,334.0
+260,2020-11-01 00:00:00+03:00,18665.0,382873.0,245.0
+261,2020-11-02 00:00:00+03:00,18257.0,390532.0,238.0
+262,2020-11-03 00:00:00+03:00,18648.0,393494.0,355.0
+263,2020-11-04 00:00:00+03:00,19768.0,397306.0,389.0
+264,2020-11-05 00:00:00+03:00,19404.0,404180.0,292.0
+265,2020-11-06 00:00:00+03:00,20582.0,407429.0,378.0
+266,2020-11-07 00:00:00+03:00,20396.0,410658.0,364.0
+267,2020-11-08 00:00:00+03:00,20498.0,419378.0,286.0
+268,2020-11-09 00:00:00+03:00,21798.0,430198.0,256.0
+269,2020-11-10 00:00:00+03:00,20977.0,435207.0,368.0
+270,2020-11-11 00:00:00+03:00,19851.0,436010.0,432.0
+271,2020-11-12 00:00:00+03:00,21608.0,438368.0,439.0
+272,2020-11-13 00:00:00+03:00,21983.0,441205.0,411.0
+273,2020-11-14 00:00:00+03:00,22702.0,444890.0,391.0
+274,2020-11-15 00:00:00+03:00,22572.0,452654.0,352.0
+275,2020-11-16 00:00:00+03:00,22778.0,461265.0,303.0
+276,2020-11-17 00:00:00+03:00,22410.0,461178.0,442.0
+277,2020-11-18 00:00:00+03:00,20985.0,456528.0,456.0
+278,2020-11-19 00:00:00+03:00,23610.0,454102.0,463.0
+279,2020-11-20 00:00:00+03:00,24318.0,453201.0,461.0
+280,2020-11-21 00:00:00+03:00,24822.0,451535.0,467.0
+281,2020-11-22 00:00:00+03:00,24581.0,457707.0,401.0
+282,2020-11-23 00:00:00+03:00,25173.0,466517.0,361.0
+283,2020-11-24 00:00:00+03:00,24326.0,467126.0,491.0
+284,2020-11-25 00:00:00+03:00,23675.0,464546.0,507.0
+285,2020-11-26 00:00:00+03:00,25487.0,464436.0,524.0
+286,2020-11-27 00:00:00+03:00,27543.0,464801.0,496.0
+287,2020-11-28 00:00:00+03:00,27100.0,464095.0,510.0
+288,2020-11-29 00:00:00+03:00,26683.0,468332.0,459.0
+289,2020-11-30 00:00:00+03:00,26338.0,477055.0,368.0
+290,2020-12-01 00:00:00+03:00,26402.0,478125.0,569.0
+291,2020-12-02 00:00:00+03:00,25345.0,475999.0,589.0
+292,2020-12-03 00:00:00+03:00,28145.0,474088.0,554.0
+293,2020-12-04 00:00:00+03:00,27403.0,472021.0,569.0
+294,2020-12-05 00:00:00+03:00,28782.0,472651.0,508.0
+295,2020-12-06 00:00:00+03:00,29039.0,479891.0,457.0
+296,2020-12-07 00:00:00+03:00,28142.0,488727.0,456.0
+297,2020-12-08 00:00:00+03:00,26097.0,489324.0,562.0
+298,2020-12-09 00:00:00+03:00,26190.0,488689.0,559.0
+299,2020-12-10 00:00:00+03:00,27927.0,490177.0,562.0
+300,2020-12-11 00:00:00+03:00,28585.0,491978.0,613.0
+301,2020-12-12 00:00:00+03:00,28137.0,493437.0,560.0
+302,2020-12-13 00:00:00+03:00,28080.0,500752.0,488.0
+303,2020-12-14 00:00:00+03:00,27328.0,509068.0,450.0
+304,2020-12-15 00:00:00+03:00,26689.0,510367.0,577.0
+305,2020-12-16 00:00:00+03:00,26509.0,509790.0,596.0
+306,2020-12-17 00:00:00+03:00,28214.0,510977.0,587.0
+307,2020-12-18 00:00:00+03:00,28552.0,512825.0,611.0
+308,2020-12-19 00:00:00+03:00,28209.0,514340.0,585.0
+309,2020-12-20 00:00:00+03:00,28948.0,521862.0,511.0
+310,2020-12-21 00:00:00+03:00,29350.0,531014.0,493.0
+311,2020-12-22 00:00:00+03:00,28776.0,535071.0,561.0
+312,2020-12-23 00:00:00+03:00,27250.0,537325.0,549.0
+313,2020-12-24 00:00:00+03:00,29935.0,539735.0,635.0
+314,2020-12-25 00:00:00+03:00,29018.0,540793.0,563.0
+315,2020-12-26 00:00:00+03:00,29258.0,541299.0,567.0
+316,2020-12-27 00:00:00+03:00,28284.0,544641.0,552.0
+317,2020-12-28 00:00:00+03:00,27787.0,551461.0,487.0
+318,2020-12-29 00:00:00+03:00,27002.0,553027.0,562.0
+319,2020-12-30 00:00:00+03:00,26513.0,549706.0,599.0
+320,2020-12-31 00:00:00+03:00,27747.0,547938.0,593.0
+321,2021-01-01 00:00:00+03:00,27039.0,548643.0,536.0
+322,2021-01-02 00:00:00+03:00,26301.0,555600.0,447.0
+323,2021-01-03 00:00:00+03:00,24150.0,559399.0,504.0
+324,2021-01-04 00:00:00+03:00,23351.0,561114.0,482.0
+325,2021-01-05 00:00:00+03:00,24246.0,562210.0,518.0
+326,2021-01-06 00:00:00+03:00,24217.0,562927.0,445.0
+327,2021-01-07 00:00:00+03:00,23541.0,562233.0,506.0
+328,2021-01-08 00:00:00+03:00,23652.0,563754.0,454.0
+329,2021-01-09 00:00:00+03:00,23309.0,562913.0,470.0
+330,2021-01-10 00:00:00+03:00,22851.0,561228.0,456.0
+331,2021-01-11 00:00:00+03:00,23315.0,562321.0,436.0
+332,2021-01-12 00:00:00+03:00,22934.0,559969.0,531.0
+333,2021-01-13 00:00:00+03:00,22850.0,553595.0,566.0
+334,2021-01-14 00:00:00+03:00,24763.0,549832.0,570.0
+335,2021-01-15 00:00:00+03:00,24715.0,546356.0,555.0
+336,2021-01-16 00:00:00+03:00,24092.0,542547.0,590.0
+337,2021-01-17 00:00:00+03:00,23586.0,542212.0,481.0
+338,2021-01-18 00:00:00+03:00,22857.0,546265.0,471.0
+339,2021-01-19 00:00:00+03:00,21734.0,544151.0,586.0
+340,2021-01-20 00:00:00+03:00,21152.0,539416.0,597.0
+341,2021-01-21 00:00:00+03:00,21887.0,533789.0,612.0
+342,2021-01-22 00:00:00+03:00,21513.0,527404.0,580.0
+343,2021-01-23 00:00:00+03:00,20921.0,519987.0,559.0
+344,2021-01-24 00:00:00+03:00,21127.0,518178.0,491.0
+345,2021-01-25 00:00:00+03:00,19290.0,518009.0,456.0
+346,2021-01-26 00:00:00+03:00,18241.0,511888.0,564.0
+347,2021-01-27 00:00:00+03:00,17741.0,501113.0,594.0
+348,2021-01-28 00:00:00+03:00,19138.0,492901.0,575.0
+349,2021-01-29 00:00:00+03:00,19238.0,485401.0,534.0
+350,2021-01-30 00:00:00+03:00,19032.0,479419.0,512.0
+351,2021-01-31 00:00:00+03:00,18359.0,477253.0,485.0
+352,2021-02-01 00:00:00+03:00,17648.0,476295.0,437.0
+353,2021-02-02 00:00:00+03:00,16643.0,470027.0,539.0
+354,2021-02-03 00:00:00+03:00,16474.0,461153.0,526.0
+355,2021-02-04 00:00:00+03:00,16714.0,452800.0,521.0
+356,2021-02-05 00:00:00+03:00,16688.0,445379.0,527.0
+357,2021-02-06 00:00:00+03:00,16627.0,438678.0,497.0
+358,2021-02-07 00:00:00+03:00,16048.0,434410.0,432.0
+359,2021-02-08 00:00:00+03:00,15916.0,434038.0,407.0
+360,2021-02-09 00:00:00+03:00,15019.0,426732.0,530.0
+361,2021-02-10 00:00:00+03:00,14494.0,418115.0,536.0
+362,2021-02-11 00:00:00+03:00,15038.0,410639.0,553.0
+363,2021-02-12 00:00:00+03:00,15089.0,404501.0,507.0
+364,2021-02-13 00:00:00+03:00,14861.0,400095.0,502.0
+365,2021-02-14 00:00:00+03:00,14185.0,398656.0,430.0
+366,2021-02-15 00:00:00+03:00,14207.0,398534.0,394.0
+367,2021-02-16 00:00:00+03:00,13233.0,393681.0,459.0
+368,2021-02-17 00:00:00+03:00,12828.0,388123.0,467.0
+369,2021-02-18 00:00:00+03:00,13447.0,382360.0,480.0
+370,2021-02-19 00:00:00+03:00,13433.0,376686.0,470.0
+371,2021-02-20 00:00:00+03:00,12953.0,371675.0,480.0
+372,2021-02-21 00:00:00+03:00,12742.0,367988.0,417.0
+373,2021-02-22 00:00:00+03:00,12604.0,367312.0,337.0
+374,2021-02-23 00:00:00+03:00,11823.0,365762.0,417.0
+375,2021-02-24 00:00:00+03:00,11749.0,364910.0,383.0
+376,2021-02-25 00:00:00+03:00,11198.0,359560.0,446.0
+377,2021-02-26 00:00:00+03:00,11086.0,354496.0,428.0
+378,2021-02-27 00:00:00+03:00,11534.0,349571.0,439.0
+379,2021-02-28 00:00:00+03:00,11359.0,348160.0,379.0
+380,2021-03-01 00:00:00+03:00,11571.0,348121.0,333.0
+381,2021-03-02 00:00:00+03:00,10565.0,343279.0,441.0
+382,2021-03-03 00:00:00+03:00,10535.0,337668.0,452.0
+383,2021-03-04 00:00:00+03:00,11385.0,332455.0,475.0
+384,2021-03-05 00:00:00+03:00,11024.0,327553.0,462.0
+385,2021-03-06 00:00:00+03:00,11022.0,323107.0,441.0
+386,2021-03-07 00:00:00+03:00,10595.0,321758.0,368.0
+387,2021-03-08 00:00:00+03:00,10253.0,321310.0,379.0
+388,2021-03-09 00:00:00+03:00,9445.0,320488.0,336.0
+389,2021-03-10 00:00:00+03:00,9079.0,315751.0,466.0
+390,2021-03-11 00:00:00+03:00,9270.0,310556.0,459.0
+391,2021-03-12 00:00:00+03:00,9794.0,306368.0,486.0
+392,2021-03-13 00:00:00+03:00,9908.0,302933.0,475.0
+393,2021-03-14 00:00:00+03:00,10083.0,303209.0,395.0
+394,2021-03-15 00:00:00+03:00,9437.0,303975.0,404.0
+395,2021-03-16 00:00:00+03:00,9393.0,302281.0,443.0
+396,2021-03-17 00:00:00+03:00,8998.0,300097.0,427.0
+397,2021-03-18 00:00:00+03:00,9803.0,297379.0,460.0
+398,2021-03-19 00:00:00+03:00,9699.0,294298.0,443.0
+399,2021-03-20 00:00:00+03:00,9632.0,292259.0,392.0
+400,2021-03-21 00:00:00+03:00,9299.0,292444.0,371.0
+401,2021-03-22 00:00:00+03:00,9284.0,293577.0,361.0
+402,2021-03-23 00:00:00+03:00,8457.0,290747.0,427.0
+403,2021-03-24 00:00:00+03:00,8861.0,288852.0,401.0
+404,2021-03-25 00:00:00+03:00,9221.0,286799.0,393.0
+405,2021-03-26 00:00:00+03:00,9167.0,284681.0,405.0
+406,2021-03-27 00:00:00+03:00,8885.0,282842.0,387.0
+407,2021-03-28 00:00:00+03:00,9088.0,282964.0,336.0
+408,2021-03-29 00:00:00+03:00,8711.0,284102.0,293.0
+409,2021-03-30 00:00:00+03:00,8277.0,282382.0,409.0
+410,2021-03-31 00:00:00+03:00,8275.0,280073.0,408.0
+411,2021-04-01 00:00:00+03:00,9169.0,278612.0,383.0
+412,2021-04-02 00:00:00+03:00,8792.0,277172.0,400.0
+413,2021-04-03 00:00:00+03:00,9021.0,276191.0,384.0
+414,2021-04-04 00:00:00+03:00,8817.0,276439.0,357.0
+415,2021-04-05 00:00:00+03:00,8646.0,277690.0,343.0
+416,2021-04-06 00:00:00+03:00,8328.0,276727.0,389.0
+417,2021-04-07 00:00:00+03:00,8294.0,275202.0,374.0
+418,2021-04-08 00:00:00+03:00,8672.0,273951.0,365.0
+419,2021-04-09 00:00:00+03:00,9150.0,273037.0,402.0
+420,2021-04-10 00:00:00+03:00,8704.0,271760.0,402.0
+421,2021-04-11 00:00:00+03:00,8702.0,272895.0,337.0
+422,2021-04-12 00:00:00+03:00,8320.0,274282.0,277.0
+423,2021-04-13 00:00:00+03:00,8173.0,272506.0,338.0
+424,2021-04-14 00:00:00+03:00,8326.0,270986.0,399.0
+425,2021-04-15 00:00:00+03:00,8944.0,269307.0,398.0
+426,2021-04-16 00:00:00+03:00,8995.0,268796.0,397.0
+427,2021-04-17 00:00:00+03:00,9321.0,268887.0,398.0
+428,2021-04-18 00:00:00+03:00,8632.0,269739.0,389.0
+429,2021-04-19 00:00:00+03:00,8589.0,271164.0,346.0
+430,2021-04-20 00:00:00+03:00,8164.0,269318.0,379.0
+431,2021-04-21 00:00:00+03:00,8271.0,267546.0,399.0
+432,2021-04-22 00:00:00+03:00,8996.0,267211.0,397.0
+433,2021-04-23 00:00:00+03:00,8840.0,266246.0,398.0
+434,2021-04-24 00:00:00+03:00,8828.0,265421.0,399.0
+435,2021-04-25 00:00:00+03:00,8780.0,266329.0,332.0
+436,2021-04-26 00:00:00+03:00,8803.0,268145.0,356.0
+437,2021-04-27 00:00:00+03:00,8053.0,267767.0,392.0
+438,2021-04-28 00:00:00+03:00,7848.0,266808.0,387.0
+439,2021-04-29 00:00:00+03:00,9284.0,267286.0,364.0
+440,2021-04-30 00:00:00+03:00,8731.0,267214.0,397.0
+441,2021-05-01 00:00:00+03:00,9270.0,267455.0,392.0
+442,2021-05-02 00:00:00+03:00,8697.0,268471.0,342.0
+443,2021-05-03 00:00:00+03:00,8489.0,270257.0,336.0
+444,2021-05-04 00:00:00+03:00,7770.0,270935.0,337.0
+445,2021-05-05 00:00:00+03:00,7975.0,271044.0,360.0
+446,2021-05-06 00:00:00+03:00,7639.0,270544.0,351.0
+447,2021-05-07 00:00:00+03:00,8386.0,270532.0,376.0
+448,2021-05-08 00:00:00+03:00,8329.0,270236.0,370.0
+449,2021-05-09 00:00:00+03:00,8419.0,270804.0,334.0
+450,2021-05-10 00:00:00+03:00,8465.0,272174.0,321.0
+451,2021-05-11 00:00:00+03:00,8115.0,272951.0,329.0
+452,2021-05-12 00:00:00+03:00,8217.0,272199.0,355.0
+453,2021-05-13 00:00:00+03:00,8380.0,270838.0,392.0
+454,2021-05-14 00:00:00+03:00,9462.0,270151.0,393.0
+455,2021-05-15 00:00:00+03:00,8790.0,268711.0,364.0
+456,2021-05-16 00:00:00+03:00,8554.0,268301.0,391.0
+457,2021-05-17 00:00:00+03:00,9328.0,270108.0,340.0
+458,2021-05-18 00:00:00+03:00,8183.0,268955.0,364.0
+459,2021-05-19 00:00:00+03:00,7920.0,266924.0,390.0
+460,2021-05-20 00:00:00+03:00,9232.0,265777.0,396.0
+461,2021-05-21 00:00:00+03:00,8937.0,264986.0,378.0
+462,2021-05-22 00:00:00+03:00,8709.0,263964.0,386.0
+463,2021-05-23 00:00:00+03:00,8951.0,265261.0,357.0
+464,2021-05-24 00:00:00+03:00,8406.0,266898.0,319.0
+465,2021-05-25 00:00:00+03:00,7884.0,265646.0,393.0
+466,2021-05-26 00:00:00+03:00,8373.0,264478.0,406.0
+467,2021-05-27 00:00:00+03:00,9039.0,263356.0,402.0
+468,2021-05-28 00:00:00+03:00,9252.0,262819.0,404.0
+469,2021-05-29 00:00:00+03:00,9289.0,262457.0,401.0
+470,2021-05-30 00:00:00+03:00,9694.0,264410.0,355.0
+471,2021-05-31 00:00:00+03:00,8475.0,265831.0,339.0
+472,2021-06-01 00:00:00+03:00,9500.0,265965.0,372.0
+473,2021-06-02 00:00:00+03:00,8832.0,265383.0,394.0
+474,2021-06-03 00:00:00+03:00,8933.0,264540.0,393.0
+475,2021-06-04 00:00:00+03:00,8947.0,264580.0,377.0
+476,2021-06-05 00:00:00+03:00,9145.0,264761.0,399.0
+477,2021-06-06 00:00:00+03:00,9163.0,266204.0,351.0
+478,2021-06-07 00:00:00+03:00,9429.0,268547.0,330.0
+479,2021-06-08 00:00:00+03:00,9977.0,269262.0,379.0
+480,2021-06-09 00:00:00+03:00,10407.0,269456.0,399.0
+481,2021-06-10 00:00:00+03:00,11699.0,270676.0,383.0
+482,2021-06-11 00:00:00+03:00,12505.0,272597.0,396.0
+483,2021-06-12 00:00:00+03:00,13510.0,275722.0,399.0
+484,2021-06-13 00:00:00+03:00,14723.0,280922.0,357.0
+485,2021-06-14 00:00:00+03:00,13721.0,285960.0,371.0
+486,2021-06-15 00:00:00+03:00,14185.0,291169.0,379.0
+487,2021-06-16 00:00:00+03:00,13397.0,293914.0,396.0
+488,2021-06-17 00:00:00+03:00,14057.0,296350.0,416.0
+489,2021-06-18 00:00:00+03:00,17262.0,302205.0,453.0
+490,2021-06-19 00:00:00+03:00,17906.0,308961.0,466.0
+491,2021-06-20 00:00:00+03:00,17611.0,317493.0,450.0
+492,2021-06-21 00:00:00+03:00,17378.0,326070.0,440.0
+493,2021-06-22 00:00:00+03:00,16715.0,331122.0,546.0
+494,2021-06-23 00:00:00+03:00,17594.0,335508.0,548.0
+495,2021-06-24 00:00:00+03:00,20182.0,341617.0,568.0
+496,2021-06-25 00:00:00+03:00,20393.0,347385.0,601.0
+497,2021-06-26 00:00:00+03:00,21665.0,354084.0,619.0
+498,2021-06-27 00:00:00+03:00,20538.0,361295.0,599.0
+499,2021-06-28 00:00:00+03:00,21650.0,369708.0,611.0
+500,2021-06-29 00:00:00+03:00,20616.0,374975.0,652.0
+501,2021-06-30 00:00:00+03:00,21042.0,378992.0,669.0
+502,2021-07-01 00:00:00+03:00,23543.0,384935.0,672.0
+503,2021-07-02 00:00:00+03:00,23218.0,389277.0,679.0
+504,2021-07-03 00:00:00+03:00,24439.0,395120.0,697.0
+505,2021-07-04 00:00:00+03:00,25142.0,404115.0,663.0
+506,2021-07-05 00:00:00+03:00,24353.0,413274.0,654.0
+507,2021-07-06 00:00:00+03:00,23378.0,417504.0,737.0
+508,2021-07-07 00:00:00+03:00,23962.0,420674.0,725.0
+509,2021-07-08 00:00:00+03:00,24818.0,423422.0,734.0
+510,2021-07-09 00:00:00+03:00,25766.0,426630.0,726.0
+511,2021-07-10 00:00:00+03:00,25082.0,433210.0,752.0
+512,2021-07-11 00:00:00+03:00,25033.0,440112.0,749.0
+513,2021-07-12 00:00:00+03:00,25140.0,448113.0,710.0
+514,2021-07-13 00:00:00+03:00,24702.0,452469.0,780.0
+515,2021-07-14 00:00:00+03:00,23827.0,454241.0,786.0
+516,2021-07-15 00:00:00+03:00,25293.0,457250.0,791.0
+517,2021-07-16 00:00:00+03:00,25704.0,460223.0,799.0
+518,2021-07-17 00:00:00+03:00,25116.0,463115.0,787.0
+519,2021-07-18 00:00:00+03:00,25018.0,468483.0,764.0
+520,2021-07-19 00:00:00+03:00,24633.0,473633.0,719.0
+521,2021-07-20 00:00:00+03:00,23770.0,474401.0,784.0
+522,2021-07-21 00:00:00+03:00,23704.0,474738.0,783.0
+523,2021-07-22 00:00:00+03:00,24471.0,475753.0,796.0
+524,2021-07-23 00:00:00+03:00,23811.0,476222.0,795.0
+525,2021-07-24 00:00:00+03:00,23947.0,477418.0,799.0
+526,2021-07-25 00:00:00+03:00,24072.0,482033.0,779.0
+527,2021-07-26 00:00:00+03:00,23239.0,488345.0,727.0
+528,2021-07-27 00:00:00+03:00,23032.0,490482.0,779.0
+529,2021-07-28 00:00:00+03:00,22420.0,491525.0,798.0
+530,2021-07-29 00:00:00+03:00,23270.0,493162.0,799.0
+531,2021-07-30 00:00:00+03:00,23564.0,495447.0,794.0
+532,2021-07-31 00:00:00+03:00,23807.0,498691.0,792.0
+533,2021-08-01 00:00:00+03:00,22804.0,503435.0,789.0
+534,2021-08-02 00:00:00+03:00,23508.0,511265.0,785.0
+535,2021-08-03 00:00:00+03:00,22010.0,513524.0,788.0
+536,2021-08-04 00:00:00+03:00,22589.0,515227.0,790.0
+537,2021-08-05 00:00:00+03:00,23120.0,517183.0,794.0
+538,2021-08-06 00:00:00+03:00,22660.0,518910.0,792.0
+539,2021-08-07 00:00:00+03:00,22320.0,520952.0,793.0
+540,2021-08-08 00:00:00+03:00,22866.0,527362.0,787.0
+541,2021-08-09 00:00:00+03:00,22160.0,534279.0,769.0
+542,2021-08-10 00:00:00+03:00,21378.0,536136.0,792.0
+543,2021-08-11 00:00:00+03:00,21571.0,536841.0,799.0
+544,2021-08-12 00:00:00+03:00,21932.0,537770.0,808.0
+545,2021-08-13 00:00:00+03:00,22277.0,539864.0,815.0
+546,2021-08-14 00:00:00+03:00,22144.0,541639.0,819.0
+547,2021-08-15 00:00:00+03:00,21624.0,546021.0,816.0
+548,2021-08-16 00:00:00+03:00,20765.0,550379.0,806.0
+549,2021-08-17 00:00:00+03:00,20958.0,552125.0,805.0
+550,2021-08-18 00:00:00+03:00,20914.0,551527.0,799.0
+551,2021-08-19 00:00:00+03:00,21058.0,547777.0,791.0
+552,2021-08-20 00:00:00+03:00,20992.0,547633.0,785.0
+553,2021-08-21 00:00:00+03:00,21000.0,547189.0,797.0
+554,2021-08-22 00:00:00+03:00,20564.0,551577.0,762.0
+555,2021-08-23 00:00:00+03:00,19454.0,554854.0,776.0
+556,2021-08-24 00:00:00+03:00,18833.0,554257.0,794.0
+557,2021-08-25 00:00:00+03:00,19536.0,553330.0,809.0
+558,2021-08-26 00:00:00+03:00,19630.0,552479.0,820.0
+559,2021-08-27 00:00:00+03:00,19509.0,551973.0,798.0
+560,2021-08-28 00:00:00+03:00,19492.0,551255.0,799.0
+561,2021-08-29 00:00:00+03:00,19286.0,552940.0,797.0
+562,2021-08-30 00:00:00+03:00,18325.0,556293.0,792.0
+563,2021-08-31 00:00:00+03:00,17813.0,554687.0,795.0
+564,2021-09-01 00:00:00+03:00,18368.0,553940.0,790.0
+565,2021-09-02 00:00:00+03:00,18985.0,553458.0,798.0
+566,2021-09-03 00:00:00+03:00,18856.0,552825.0,799.0
+567,2021-09-04 00:00:00+03:00,18780.0,552072.0,796.0
+568,2021-09-05 00:00:00+03:00,18645.0,554668.0,793.0
+569,2021-09-06 00:00:00+03:00,17856.0,557458.0,790.0
+570,2021-09-07 00:00:00+03:00,17425.0,556845.0,795.0
+571,2021-09-08 00:00:00+03:00,18024.0,555810.0,797.0
+572,2021-09-09 00:00:00+03:00,18380.0,553757.0,794.0
+573,2021-09-10 00:00:00+03:00,18341.0,554188.0,789.0
+574,2021-09-11 00:00:00+03:00,18891.0,554395.0,796.0
+575,2021-09-12 00:00:00+03:00,18554.0,557664.0,788.0
+576,2021-09-13 00:00:00+03:00,18178.0,562654.0,719.0
+577,2021-09-14 00:00:00+03:00,17837.0,563803.0,781.0
+578,2021-09-15 00:00:00+03:00,18841.0,564813.0,792.0
+579,2021-09-16 00:00:00+03:00,19594.0,566287.0,794.0
+580,2021-09-17 00:00:00+03:00,19905.0,568782.0,791.0
+581,2021-09-18 00:00:00+03:00,20329.0,572065.0,799.0
+582,2021-09-19 00:00:00+03:00,20174.0,578028.0,793.0
+583,2021-09-20 00:00:00+03:00,19744.0,585002.0,778.0
+584,2021-09-21 00:00:00+03:00,19179.0,587932.0,812.0
+585,2021-09-22 00:00:00+03:00,19706.0,590719.0,817.0
+586,2021-09-23 00:00:00+03:00,21438.0,594770.0,820.0
+587,2021-09-24 00:00:00+03:00,21379.0,599493.0,828.0
+588,2021-09-25 00:00:00+03:00,22041.0,604387.0,822.0
+589,2021-09-26 00:00:00+03:00,22498.0,612409.0,805.0
+590,2021-09-27 00:00:00+03:00,22236.0,620353.0,779.0
+591,2021-09-28 00:00:00+03:00,21559.0,623692.0,852.0
+592,2021-09-29 00:00:00+03:00,22430.0,626809.0,857.0
+593,2021-09-30 00:00:00+03:00,23888.0,631004.0,867.0
+594,2021-10-01 00:00:00+03:00,24522.0,634684.0,887.0
+595,2021-10-02 00:00:00+03:00,25219.0,641165.0,886.0
+596,2021-10-03 00:00:00+03:00,25769.0,650653.0,890.0
+597,2021-10-04 00:00:00+03:00,25781.0,661025.0,883.0
+598,2021-10-05 00:00:00+03:00,25110.0,666672.0,895.0
+599,2021-10-06 00:00:00+03:00,25133.0,671035.0,929.0
+600,2021-10-07 00:00:00+03:00,27550.0,677331.0,924.0
+601,2021-10-08 00:00:00+03:00,27246.0,683075.0,936.0
+602,2021-10-09 00:00:00+03:00,29362.0,690420.0,968.0
+603,2021-10-10 00:00:00+03:00,28647.0,700831.0,962.0
+604,2021-10-11 00:00:00+03:00,29409.0,713823.0,957.0
+605,2021-10-12 00:00:00+03:00,28190.0,720334.0,973.0
+606,2021-10-13 00:00:00+03:00,28717.0,726266.0,984.0
+607,2021-10-14 00:00:00+03:00,31299.0,734909.0,986.0
+608,2021-10-15 00:00:00+03:00,32196.0,743839.0,998.0
+609,2021-10-16 00:00:00+03:00,33208.0,754162.0,1002.0
+610,2021-10-17 00:00:00+03:00,34303.0,768751.0,997.0
+611,2021-10-18 00:00:00+03:00,34325.0,785647.0,998.0
+612,2021-10-19 00:00:00+03:00,33740.0,794946.0,1015.0
+613,2021-10-20 00:00:00+03:00,34073.0,802760.0,1028.0
+614,2021-10-21 00:00:00+03:00,36339.0,812168.0,1036.0
+615,2021-10-22 00:00:00+03:00,37141.0,822792.0,1064.0
+616,2021-10-23 00:00:00+03:00,37678.0,833318.0,1075.0
+617,2021-10-24 00:00:00+03:00,35660.0,845122.0,1072.0
+618,2021-10-25 00:00:00+03:00,37930.0,861293.0,1069.0
+619,2021-10-26 00:00:00+03:00,36446.0,869660.0,1106.0
+620,2021-10-27 00:00:00+03:00,36582.0,875968.0,1123.0
+621,2021-10-28 00:00:00+03:00,40096.0,885587.0,1159.0
+622,2021-10-29 00:00:00+03:00,39849.0,893811.0,1163.0
+623,2021-10-30 00:00:00+03:00,40251.0,903993.0,1160.0
+624,2021-10-31 00:00:00+03:00,40993.0,916713.0,1158.0
+625,2021-11-01 00:00:00+03:00,40402.0,932773.0,1155.0
+626,2021-11-02 00:00:00+03:00,39008.0,939698.0,1178.0
+627,2021-11-03 00:00:00+03:00,40443.0,946145.0,1189.0
+628,2021-11-04 00:00:00+03:00,40217.0,953239.0,1195.0
+629,2021-11-05 00:00:00+03:00,40735.0,964177.0,1192.0
+630,2021-11-06 00:00:00+03:00,41335.0,975123.0,1188.0
+631,2021-11-07 00:00:00+03:00,39165.0,986303.0,1179.0
+632,2021-11-08 00:00:00+03:00,39400.0,998931.0,1190.0
+633,2021-11-09 00:00:00+03:00,39160.0,1004844.0,1211.0
+634,2021-11-10 00:00:00+03:00,38058.0,1007098.0,1239.0
+635,2021-11-11 00:00:00+03:00,40759.0,1013464.0,1237.0
+636,2021-11-12 00:00:00+03:00,40123.0,1018707.0,1235.0
+637,2021-11-13 00:00:00+03:00,39256.0,1022920.0,1241.0
+638,2021-11-14 00:00:00+03:00,38823.0,1030703.0,1219.0
+639,2021-11-15 00:00:00+03:00,38420.0,1040210.0,1211.0
+640,2021-11-16 00:00:00+03:00,36818.0,1041627.0,1240.0
+641,2021-11-17 00:00:00+03:00,36626.0,1040618.0,1247.0
+642,2021-11-18 00:00:00+03:00,37374.0,1040327.0,1251.0
+643,2021-11-19 00:00:00+03:00,37156.0,1039225.0,1254.0
+644,2021-11-20 00:00:00+03:00,37120.0,1038919.0,1254.0
+645,2021-11-21 00:00:00+03:00,36970.0,1042133.0,1252.0
+646,2021-11-22 00:00:00+03:00,35681.0,1047860.0,1241.0
+647,2021-11-23 00:00:00+03:00,33996.0,1044562.0,1243.0
+648,2021-11-24 00:00:00+03:00,33558.0,1040198.0,1240.0
+649,2021-11-25 00:00:00+03:00,33796.0,1034306.0,1238.0
+650,2021-11-26 00:00:00+03:00,34690.0,1031616.0,1235.0
+651,2021-11-27 00:00:00+03:00,33946.0,1027829.0,1239.0
+652,2021-11-28 00:00:00+03:00,33548.0,1029507.0,1224.0
+653,2021-11-29 00:00:00+03:00,33860.0,1034458.0,1209.0
+654,2021-11-30 00:00:00+03:00,32648.0,1032435.0,1229.0
+655,2021-12-01 00:00:00+03:00,32837.0,1028367.0,1226.0
+656,2021-12-02 00:00:00+03:00,33389.0,1025350.0,1221.0
+657,2021-12-03 00:00:00+03:00,32930.0,1020549.0,1217.0
+658,2021-12-04 00:00:00+03:00,32974.0,1017126.0,1215.0
+659,2021-12-05 00:00:00+03:00,32602.0,1017929.0,1206.0
+660,2021-12-06 00:00:00+03:00,32136.0,1020811.0,1184.0
+661,2021-12-07 00:00:00+03:00,31096.0,1016110.0,1182.0
+662,2021-12-08 00:00:00+03:00,30752.0,1008707.0,1179.0
+663,2021-12-09 00:00:00+03:00,30209.0,1001941.0,1181.0
+664,2021-12-10 00:00:00+03:00,30873.0,995981.0,1176.0
+665,2021-12-11 00:00:00+03:00,30288.0,988652.0,1171.0
+666,2021-12-12 00:00:00+03:00,29929.0,986058.0,1132.0
+667,2021-12-13 00:00:00+03:00,29558.0,985934.0,1121.0
+668,2021-12-14 00:00:00+03:00,28343.0,979048.0,1145.0
+669,2021-12-15 00:00:00+03:00,28363.0,970636.0,1142.0
+670,2021-12-16 00:00:00+03:00,28486.0,960834.0,1133.0
+671,2021-12-17 00:00:00+03:00,27743.0,950060.0,1080.0
+672,2021-12-18 00:00:00+03:00,27434.0,938377.0,1076.0
+673,2021-12-19 00:00:00+03:00,27967.0,932666.0,1023.0
+674,2021-12-20 00:00:00+03:00,27022.0,928610.0,1019.0
+675,2021-12-21 00:00:00+03:00,25907.0,913271.0,1027.0
+676,2021-12-22 00:00:00+03:00,25264.0,895193.0,1020.0
+677,2021-12-23 00:00:00+03:00,25667.0,878213.0,1002.0
+678,2021-12-24 00:00:00+03:00,24703.0,860705.0,998.0
+679,2021-12-25 00:00:00+03:00,24946.0,842563.0,981.0
+680,2021-12-26 00:00:00+03:00,23721.0,828031.0,968.0
+681,2021-12-27 00:00:00+03:00,23210.0,816589.0,937.0
+682,2021-12-28 00:00:00+03:00,21922.0,793615.0,935.0
+683,2021-12-29 00:00:00+03:00,21119.0,771026.0,932.0
+684,2021-12-30 00:00:00+03:00,21073.0,748169.0,926.0
+685,2021-12-31 00:00:00+03:00,20638.0,727203.0,912.0
+686,2022-01-01 00:00:00+03:00,19751.0,712963.0,847.0
+687,2022-01-02 00:00:00+03:00,18233.0,703409.0,811.0
+688,2022-01-03 00:00:00+03:00,16343.0,694880.0,835.0
+689,2022-01-04 00:00:00+03:00,15903.0,682878.0,834.0
+690,2022-01-05 00:00:00+03:00,15772.0,672241.0,828.0
+691,2022-01-06 00:00:00+03:00,15316.0,663806.0,802.0
+692,2022-01-07 00:00:00+03:00,16735.0,657719.0,787.0
+693,2022-01-08 00:00:00+03:00,16568.0,653042.0,796.0
+694,2022-01-09 00:00:00+03:00,16246.0,647774.0,763.0
+695,2022-01-10 00:00:00+03:00,15830.0,642973.0,741.0
+696,2022-01-11 00:00:00+03:00,17525.0,634499.0,783.0
+697,2022-01-12 00:00:00+03:00,17946.0,625354.0,745.0
+698,2022-01-13 00:00:00+03:00,21155.0,619785.0,740.0
+699,2022-01-14 00:00:00+03:00,23820.0,617914.0,739.0
+700,2022-01-15 00:00:00+03:00,27179.0,617786.0,723.0
+701,2022-01-16 00:00:00+03:00,29230.0,623599.0,686.0
+702,2022-01-17 00:00:00+03:00,30726.0,633899.0,670.0
+703,2022-01-18 00:00:00+03:00,31252.0,639899.0,688.0
+704,2022-01-19 00:00:00+03:00,33899.0,650180.0,698.0
+705,2022-01-20 00:00:00+03:00,38850.0,663868.0,684.0
+706,2022-01-21 00:00:00+03:00,49513.0,687970.0,692.0
+707,2022-01-22 00:00:00+03:00,49485.82,718976.0,681.0
+708,2022-01-23 00:00:00+03:00,49458.64,758457.0,679.0
+709,2022-01-24 00:00:00+03:00,49431.46,801197.0,655.0
+710,2022-01-25 00:00:00+03:00,49404.28,841921.0,681.0
+711,2022-01-26 00:00:00+03:00,49377.1,887759.0,657.0
+712,2022-01-27 00:00:00+03:00,49349.92,946156.0,665.0
+713,2022-01-28 00:00:00+03:00,49322.74,1014017.0,673.0
+714,2022-01-29 00:00:00+03:00,49295.56,1014783.3,668.0
+715,2022-01-30 00:00:00+03:00,49268.38,1015549.6,617.0
+716,2022-01-31 00:00:00+03:00,49241.2,1016315.9,621.0
+717,2022-02-01 00:00:00+03:00,49214.02,1017082.2,663.0
+718,2022-02-02 00:00:00+03:00,49186.84,1017848.5,678.0
+719,2022-02-03 00:00:00+03:00,49159.66,1018614.8,667.0
+720,2022-02-04 00:00:00+03:00,49132.48,1019381.1,682.0
+721,2022-02-05 00:00:00+03:00,49105.3,1020147.4,714.0
+722,2022-02-06 00:00:00+03:00,49078.12,1020913.7,661.0
+723,2022-02-07 00:00:00+03:00,49050.94,1021680.0,609.0
+724,2022-02-08 00:00:00+03:00,49023.76,1022446.3,698.0
+725,2022-02-09 00:00:00+03:00,48996.58,1023212.6,669.0
+726,2022-02-10 00:00:00+03:00,48969.4,1023978.9,701.0
+727,2022-02-11 00:00:00+03:00,48942.22,1024745.2,722.0
+728,2022-02-12 00:00:00+03:00,48915.04,1025511.5,729.0
+729,2022-02-13 00:00:00+03:00,48887.86,1026277.8,706.0
+730,2022-02-14 00:00:00+03:00,48860.68,1027044.1,683.0
+731,2022-02-15 00:00:00+03:00,48833.5,1027810.4,704.0
+732,2022-02-16 00:00:00+03:00,48806.32,1028576.7,748.0
+733,2022-02-17 00:00:00+03:00,48779.14,1029343.0,790.0
+734,2022-02-18 00:00:00+03:00,48751.96,1030109.3,784.0
+735,2022-02-19 00:00:00+03:00,48724.78,1030875.6,798.0
+736,2022-02-20 00:00:00+03:00,48697.6,1031641.9,745.0
+737,2022-02-21 00:00:00+03:00,48670.42,1032408.2,735.0
+738,2022-02-22 00:00:00+03:00,48643.24,1033174.5,796.0
+739,2022-02-23 00:00:00+03:00,48616.06,1033940.8,785.0
+740,2022-02-24 00:00:00+03:00,48588.88,1034707.1,762.0
+741,2022-02-25 00:00:00+03:00,48561.7,1035473.4,787.0
+742,2022-02-26 00:00:00+03:00,48534.52,1036239.7,793.0
+743,2022-02-27 00:00:00+03:00,48507.34,1037006.0,769.0
+744,2022-02-28 00:00:00+03:00,48480.16,1037772.3,733.0
+745,2022-03-01 00:00:00+03:00,48452.98,1038538.6,786.0
+746,2022-03-02 00:00:00+03:00,48425.8,1039304.9,784.0
+747,2022-03-03 00:00:00+03:00,48398.62,1040071.2,781.0
+748,2022-03-04 00:00:00+03:00,48371.44,1040837.5,776.0
+749,2022-03-05 00:00:00+03:00,48344.26,1041603.8,750.0
+750,2022-03-06 00:00:00+03:00,48317.08,1042370.1,744.0
+751,2022-03-07 00:00:00+03:00,48289.9,1043136.4,668.0
+752,2022-03-08 00:00:00+03:00,48262.72,1043902.7,652.0
+753,2022-03-09 00:00:00+03:00,48235.54,1044669.0,645.0
+754,2022-03-10 00:00:00+03:00,48208.36,1045435.3,665.0
+755,2022-03-11 00:00:00+03:00,48181.18,1046201.6,674.0
+756,2022-03-12 00:00:00+03:00,48154.0,1046967.9,630.0
+757,2022-03-13 00:00:00+03:00,44989.0,1047734.2,596.0
+758,2022-03-14 00:00:00+03:00,41055.0,1048500.5,533.0
+759,2022-03-15 00:00:00+03:00,36678.0,1049266.8,558.0
+760,2022-03-16 00:00:00+03:00,36519.0,1050033.1,576.0
+761,2022-03-17 00:00:00+03:00,34819.0,1050799.4,561.0
+762,2022-03-18 00:00:00+03:00,34442.0,1051565.7,524.0
+763,2022-03-19 00:00:00+03:00,32958.0,1052332.0,495.0
+764,2022-03-20 00:00:00+03:00,31035.0,1008258.0,434.0
+765,2022-03-21 00:00:00+03:00,28709.0,974620.0,409.0
+766,2022-03-22 00:00:00+03:00,26394.0,925622.0,472.0
+767,2022-03-23 00:00:00+03:00,26826.0,881397.0,429.0
+768,2022-03-24 00:00:00+03:00,25387.0,839890.0,418.0
+769,2022-03-25 00:00:00+03:00,25382.0,803014.0,398.0
+770,2022-03-26 00:00:00+03:00,24072.0,765758.0,395.0
+771,2022-03-27 00:00:00+03:00,23280.0,741160.0,338.0
+772,2022-03-28 00:00:00+03:00,21101.0,726180.0,335.0
+773,2022-03-29 00:00:00+03:00,19660.0,698272.0,339.0
+774,2022-03-30 00:00:00+03:00,20145.0,666980.0,352.0
+775,2022-03-31 00:00:00+03:00,19277.0,635416.0,345.0
+776,2022-04-01 00:00:00+03:00,19164.0,608240.0,342.0
+777,2022-04-02 00:00:00+03:00,17949.0,581302.0,340.0
+778,2022-04-03 00:00:00+03:00,16828.0,562131.0,304.0
+779,2022-04-04 00:00:00+03:00,15291.0,544285.0,287.0
+780,2022-04-05 00:00:00+03:00,13947.0,520457.0,316.0
+781,2022-04-06 00:00:00+03:00,14661.0,499521.0,291.0
+782,2022-04-07 00:00:00+03:00,14355.0,478285.0,287.0
+783,2022-04-08 00:00:00+03:00,14311.0,454632.0,280.0
+784,2022-04-09 00:00:00+03:00,13573.0,427817.0,288.0
+785,2022-04-10 00:00:00+03:00,13056.0,411242.0,259.0
+786,2022-04-11 00:00:00+03:00,11855.0,396094.0,248.0
+787,2022-04-12 00:00:00+03:00,10910.0,375452.0,281.0
+788,2022-04-13 00:00:00+03:00,11754.0,358804.0,267.0
+789,2022-04-14 00:00:00+03:00,11348.0,348045.0,254.0
+790,2022-04-15 00:00:00+03:00,11432.0,339693.0,261.0
+791,2022-04-16 00:00:00+03:00,11095.0,331542.0,240.0
+792,2022-04-17 00:00:00+03:00,10263.0,325076.0,233.0
+793,2022-04-18 00:00:00+03:00,9434.0,319973.0,213.0
+794,2022-04-19 00:00:00+03:00,8640.0,311615.0,235.0
+795,2022-04-20 00:00:00+03:00,9195.0,304387.0,223.0
+796,2022-04-21 00:00:00+03:00,8875.0,298883.0,197.0
+797,2022-04-22 00:00:00+03:00,9001.0,294209.0,195.0
+798,2022-04-23 00:00:00+03:00,8829.0,289837.0,171.0
+799,2022-04-24 00:00:00+03:00,8446.0,287607.0,168.0
+800,2022-04-25 00:00:00+03:00,7651.0,286244.0,159.0
+801,2022-04-26 00:00:00+03:00,7107.0,281276.0,176.0
+802,2022-04-27 00:00:00+03:00,7705.0,277341.0,163.0
+803,2022-04-28 00:00:00+03:00,7681.0,273793.0,166.0
+804,2022-04-29 00:00:00+03:00,7710.0,270301.0,161.0
+805,2022-04-30 00:00:00+03:00,7363.0,266485.0,157.0
+806,2022-05-01 00:00:00+03:00,7047.0,264652.0,147.0
+807,2022-05-02 00:00:00+03:00,6207.0,263221.0,136.0
+808,2022-05-03 00:00:00+03:00,5466.0,261510.0,125.0
+809,2022-05-04 00:00:00+03:00,5093.0,259810.0,129.0
+810,2022-05-05 00:00:00+03:00,5011.0,255959.0,139.0
+811,2022-05-06 00:00:00+03:00,5541.0,251956.0,136.0
+812,2022-05-07 00:00:00+03:00,5500.0,248675.0,132.0
+813,2022-05-08 00:00:00+03:00,5447.0,247322.0,118.0
+814,2022-05-09 00:00:00+03:00,5030.0,246290.0,103.0
+815,2022-05-10 00:00:00+03:00,4531.0,245706.0,101.0
+816,2022-05-11 00:00:00+03:00,4102.0,244667.0,98.0
+817,2022-05-12 00:00:00+03:00,4065.0,241691.0,111.0
+818,2022-05-13 00:00:00+03:00,4896.0,239225.0,105.0
+819,2022-05-14 00:00:00+03:00,5047.0,236787.0,107.0

requirements.txt CHANGED Viewed

@@ -1,11 +1,17 @@
 pandas
 numpy
-streamlit
-pytz
 plotly
 statsmodels
-scikit-learn
-matplotlib
-pdfkit
 scipy
-seaborn

+streamlit
 pandas
 numpy
 plotly
 statsmodels
 scipy
+scikit-learn
+pytz
+pmdarima
+arch
+prophet           # если хотите Prophet (fbprophet -> prophet)
+lightgbm
+xgboost
+torch
+tensorflow        # альтернативно для Keras
+tbats             # если нужен TBATS
+autogluon.timeseries  # для AutoGluon (очень тяжёлый)

src/lab3_functions.py ADDED Viewed

	@@ -0,0 +1,192 @@

+# TimeSeriesHomework/src/lab3_functions.py
+"""
+Набор вспомогательных функций для ЛР3:
+- обёртки для SARIMAX, auto_arima (pmdarima), VAR, GARCH (arch) и т.п.
+- forecast helpers
+- простые метрики
+Файл не использует абсолютных путей и предназначен для импорта в проекте.
+"""
+import warnings
+warnings.filterwarnings("ignore")
+from typing import Tuple, Dict, Any, Optional, List
+import numpy as np
+import pandas as pd
+# optional heavy deps
+try:
+    import pmdarima as pm
+    PM_AVAILABLE = True
+except Exception:
+    PM_AVAILABLE = False
+try:
+    from statsmodels.tsa.statespace.sarimax import SARIMAX
+    from statsmodels.tsa.api import VAR
+    STATSMODELS_AVAILABLE = True
+except Exception:
+    STATSMODELS_AVAILABLE = False
+try:
+    from arch import arch_model
+    ARCH_AVAILABLE = True
+except Exception:
+    ARCH_AVAILABLE = False
+# sklearn metrics used for convenience (optional)
+try:
+    from sklearn.metrics import mean_absolute_error, mean_squared_error
+    SKLEARN_AVAILABLE = True
+except Exception:
+    SKLEARN_AVAILABLE = False
+def is_pandas_series(x: Any) -> bool:
+    return isinstance(x, (pd.Series,))
+def mae_rmse(y_true, y_pred) -> Dict[str, float]:
+    y_true = np.array(y_true)
+    y_pred = np.array(y_pred)
+    if SKLEARN_AVAILABLE:
+        mae = float(mean_absolute_error(y_true, y_pred))
+        rmse = float(np.sqrt(mean_squared_error(y_true, y_pred)))
+    else:
+        mae = float(np.mean(np.abs(y_true - y_pred)))
+        rmse = float(np.sqrt(np.mean((y_true - y_pred) ** 2)))
+    return {"MAE": mae, "RMSE": rmse}
+def fit_auto_arima(series: pd.Series, seasonal: bool = False, m: int = 1, **kwargs):
+    """
+    Подбор ARIMA через pmdarima.auto_arima. Возвращает обученную модель pmdarima.
+    """
+    if not PM_AVAILABLE:
+        raise ImportError("pmdarima не установлен. Установите pmdarima (pip install pmdarima).")
+    if not is_pandas_series(series):
+        series = pd.Series(series)
+    series_clean = series.dropna()
+    if series_clean.empty:
+        raise ValueError("Пустая серия передана в fit_auto_arima.")
+    model = pm.auto_arima(series_clean, seasonal=seasonal, m=m, error_action="ignore", suppress_warnings=True, **kwargs)
+    return model
+def fit_sarimax(series: pd.Series, order: Tuple[int, int, int] = (1, 0, 0),
+                seasonal_order: Tuple[int, int, int, int] = (0, 0, 0, 0),
+                enforce_stationarity: bool = False, enforce_invertibility: bool = True, **fit_kwargs):
+    """
+    Обучает SARIMAX (statsmodels). Возвращает результат fit() (SARIMAXResults).
+    """
+    if not STATSMODELS_AVAILABLE:
+        raise ImportError("statsmodels не установлен. Установите statsmodels.")
+    if not is_pandas_series(series):
+        series = pd.Series(series)
+    series_clean = series.dropna()
+    if series_clean.empty:
+        raise ValueError("Пустая серия передана в fit_sarimax.")
+    model = SARIMAX(series_clean, order=order, seasonal_order=seasonal_order,
+                    enforce_stationarity=enforce_stationarity, enforce_invertibility=enforce_invertibility)
+    res = model.fit(disp=False, **fit_kwargs)
+    return res
+def forecast_sarimax(fit_res, steps: int, alpha: float = 0.05) -> Tuple[np.ndarray, Tuple[np.ndarray, np.ndarray]]:
+    """
+    Делает прогноз из обученного SARIMAX-результата (res.get_forecast).
+    Возвращает (mean, (lower, upper)) — numpy arrays длины steps.
+    """
+    if hasattr(fit_res, "get_forecast"):
+        fc = fit_res.get_forecast(steps=steps)
+        mean = np.asarray(fc.predicted_mean)
+        try:
+            conf = fc.conf_int(alpha=alpha)
+            lower = np.asarray(conf.iloc[:, 0])
+            upper = np.asarray(conf.iloc[:, 1])
+        except Exception:
+            lower = np.full(len(mean), np.nan)
+            upper = np.full(len(mean), np.nan)
+        return mean, (lower, upper)
+    else:
+        # fallback на forecast
+        try:
+            f = fit_res.forecast(steps=steps)
+            mean = np.asarray(f)
+            lower = np.full(len(mean), np.nan)
+            upper = np.full(len(mean), np.nan)
+            return mean, (lower, upper)
+        except Exception as e:
+            raise ValueError(f"Не удалось получить прогноз из объекта результата: {e}")
+def fit_var(df: pd.DataFrame, maxlags: int = 15):
+    """
+    Обучает VAR на multivariate dataframe (pandas DataFrame). Возвращает fitted VARResults.
+    """
+    if not STATSMODELS_AVAILABLE:
+        raise ImportError("statsmodels не установлен. Установите statsmodels.")
+    if not isinstance(df, pd.DataFrame):
+        raise ValueError("fit_var ожидает pd.DataFrame с несколькими числовыми колонками.")
+    df_clean = df.dropna()
+    if df_clean.shape[0] < 3:
+        raise ValueError("Недостаточно наблюдений для VAR.")
+    model = VAR(df_clean)
+    sel = model.select_order(maxlags=maxlags)
+    best_lag = None
+    try:
+        if hasattr(sel, "selected_orders"):
+            so = sel.selected_orders
+            for k in ("aic", "bic", "fpe", "hqic"):
+                val = so.get(k, None)
+                if val is not None:
+                    best_lag = int(val)
+                    break
+    except Exception:
+        best_lag = None
+    if best_lag is None or best_lag < 1:
+        best_lag = 1
+    fitted = model.fit(maxlags=best_lag)
+    return fitted
+def forecast_var(fitted_var, steps: int) -> pd.DataFrame:
+    """
+    Multi-step forecasting for VARResults. Возвращает DataFrame прогнозов (columns = variables).
+    """
+    try:
+        forecast = fitted_var.forecast(fitted_var.endog[-fitted_var.k_ar:], steps=steps)
+        cols = fitted_var.names
+        idx = range(1, steps + 1)
+        return pd.DataFrame(forecast, columns=cols, index=idx)
+    except Exception as e:
+        raise ValueError(f"Ошибка при прогнозе VAR: {e}")
+def fit_garch(series: pd.Series, p: int = 1, q: int = 1):
+    """
+    Обучает GARCH(p,q) (arch package). Возвращает объект результата fit() из arch.
+    """
+    if not ARCH_AVAILABLE:
+        raise ImportError("arch не установлен. Установите arch (pip install arch).")
+    if not is_pandas_series(series):
+        series = pd.Series(series)
+    series_clean = series.dropna()
+    if series_clean.empty:
+        raise ValueError("Пустая серия передана в fit_garch.")
+    am = arch_model(series_clean, vol="Garch", p=p, q=q, dist="normal")
+    res = am.fit(disp="off")
+    return res
+def safe_summary(obj) -> str:
+    try:
+        return str(obj.summary())
+    except Exception:
+        return repr(obj)
+# краткий тест при запуске модуля напрямую
+if __name__ == "__main__":
+    print("lab3_functions: доступные функции:",
+          [n for n in dir() if n.startswith("fit_") or n.startswith("forecast_")])

src/lab3_pipeline.py ADDED Viewed

	@@ -0,0 +1,272 @@

+# TimeSeriesHomework/src/lab3_pipeline.py
+"""
+lab3_pipeline.py
+Простой воспроизводимый pipeline для ЛР №3:
+- загрузка CSV/Parquet
+- предобработка / ресемплинг
+- Box-Cox / log (если возможно)
+- тесты стационарности (ADF, KPSS)
+- создание лагов / роллов
+- простые модели (Naive, SeasonalNaive, SARIMAX, auto_arima если доступен)
+- VAR если мультивариатные данные
+- генерация простого HTML-отчёта
+"""
+import argparse
+import os
+import math
+import time
+from typing import Tuple, Dict, Any, Optional, List
+import numpy as np
+import pandas as pd
+import matplotlib.pyplot as plt
+# statsmodels (нужен для большинства тестов/моделей)
+try:
+    from statsmodels.tsa.stattools import adfuller, kpss
+    from statsmodels.tsa.seasonal import seasonal_decompose
+    from statsmodels.stats.diagnostic import acorr_ljungbox
+    STATSMODELS_AVAILABLE = True
+except Exception as e:
+    STATSMODELS_AVAILABLE = False
+    print("warning: statsmodels not available:", e)
+# optional heavy deps
+try:
+    import pmdarima as pm
+    PM_AVAILABLE = True
+except Exception:
+    PM_AVAILABLE = False
+try:
+    from scipy.stats import boxcox, boxcox_normmax
+    SCIPY_AVAILABLE = True
+except Exception:
+    SCIPY_AVAILABLE = False
+# ---- helpers / metrics ----
+def mae(y_true, y_pred): return np.mean(np.abs(y_true - y_pred))
+def rmse(y_true, y_pred): return math.sqrt(np.mean((y_true - y_pred)**2))
+def mape(y_true, y_pred): return np.mean(np.abs((y_true - y_pred) / (y_true + 1e-9))) * 100.0
+def smape(y_true, y_pred): return 100.0 * np.mean(2.0 * np.abs(y_pred - y_true) / (np.abs(y_true) + np.abs(y_pred) + 1e-9))
+# ---- IO / preprocessing ----
+def load_data(path: str, timestamp_col: str = 'timestamp', tz: Optional[str] = None) -> pd.DataFrame:
+    if path.endswith('.parquet'):
+        df = pd.read_parquet(path)
+    else:
+        df = pd.read_csv(path)
+    if timestamp_col not in df.columns:
+        raise ValueError(f"timestamp column '{timestamp_col}' not found in data")
+    df[timestamp_col] = pd.to_datetime(df[timestamp_col], errors='coerce')
+    if tz is not None:
+        try:
+            df[timestamp_col] = df[timestamp_col].dt.tz_localize(tz)
+        except Exception:
+            try:
+                df[timestamp_col] = df[timestamp_col].dt.tz_convert(tz)
+            except Exception:
+                pass
+    df = df.sort_values(timestamp_col).drop_duplicates(subset=[timestamp_col])
+    df = df.set_index(timestamp_col)
+    return df
+def resample_and_interpolate(df: pd.DataFrame, freq: str = 'D', method: str = 'linear') -> pd.DataFrame:
+    dfr = df.resample(freq).asfreq()
+    if method == 'linear':
+        return dfr.interpolate(method='linear')
+    elif method == 'ffill':
+        return dfr.fillna(method='ffill')
+    else:
+        return dfr.fillna(method='ffill').interpolate()
+# ---- transforms ----
+def apply_log(series: pd.Series):
+    if (series <= 0).any():
+        raise ValueError("log transform requires positive values")
+    return np.log(series)
+def apply_boxcox(series: pd.Series, lmbda: Optional[float] = None):
+    if not SCIPY_AVAILABLE:
+        raise ImportError("scipy required for boxcox transformations")
+    s = series.dropna()
+    if (s <= 0).any():
+        raise ValueError("Box-Cox requires positive values")
+    if lmbda is None:
+        lmbda = boxcox_normmax(s, brack=(-2,2))
+    transformed = boxcox(s, lmbda)
+    out = pd.Series(index=s.index, data=transformed)
+    return out, lmbda
+def difference(series: pd.Series, d=1):
+    return series.diff(d)
+# ---- feature engineering ----
+def make_lags(df: pd.DataFrame, col: str, lags: List[int]):
+    for l in lags:
+        df[f'{col}_lag_{l}'] = df[col].shift(l)
+    return df
+def make_rolls(df: pd.DataFrame, col: str, windows: List[int]):
+    for w in windows:
+        df[f'{col}_roll_mean_{w}'] = df[col].rolling(window=w, min_periods=1).mean()
+        df[f'{col}_roll_std_{w}']  = df[col].rolling(window=w, min_periods=1).std()
+        df[f'{col}_roll_min_{w}']  = df[col].rolling(window=w, min_periods=1).min()
+        df[f'{col}_roll_max_{w}']  = df[col].rolling(window=w, min_periods=1).max()
+    return df
+def make_time_features(df: pd.DataFrame, index_name: Optional[str] = None):
+    idx = df.index if index_name is None else pd.to_datetime(df[index_name])
+    df['dayofweek'] = idx.dayofweek
+    df['month'] = idx.month
+    df['sin_week'] = np.sin(2 * np.pi * df['dayofweek'] / 7)
+    df['cos_month'] = np.cos(2 * np.pi * (df['month'] - 1) / 12)
+    return df
+def chronological_split(df: pd.DataFrame, frac_train=0.7, frac_val=0.15):
+    n = len(df)
+    i_train = int(n * frac_train)
+    i_val = i_train + int(n * frac_val)
+    train = df.iloc[:i_train].copy()
+    val = df.iloc[i_train:i_val].copy()
+    test = df.iloc[i_val:].copy()
+    return train, val, test
+# ---- simple models / wrappers ----
+def naive_forecast(train: pd.Series, steps: int):
+    last = train.iloc[-1]
+    return np.full(steps, last)
+def seasonal_naive_forecast(series: pd.Series, season: int, steps: int):
+    last = series.iloc[-season:]
+    reps = int(np.ceil(steps / season))
+    arr = np.tile(last.values, reps)[:steps]
+    return arr
+def fit_sarimax_wrapper(series: pd.Series, order=(1,0,0), seasonal_order=(0,0,0,0), **kwargs):
+    if not STATSMODELS_AVAILABLE:
+        raise ImportError("statsmodels required for SARIMAX")
+    from statsmodels.tsa.statespace.sarimax import SARIMAX
+    m = SARIMAX(series.dropna(), order=order, seasonal_order=seasonal_order,
+                enforce_stationarity=False, enforce_invertibility=False)
+    res = m.fit(disp=False, **kwargs)
+    return res
+def forecast_sarimax(fit_res, steps: int, alpha: float = 0.05):
+    if hasattr(fit_res, "get_forecast"):
+        fc = fit_res.get_forecast(steps=steps)
+        mean = np.asarray(fc.predicted_mean)
+        try:
+            conf = fc.conf_int(alpha=alpha)
+            lower = np.asarray(conf.iloc[:,0])
+            upper = np.asarray(conf.iloc[:,1])
+        except Exception:
+            lower = np.full(len(mean), np.nan)
+            upper = np.full(len(mean), np.nan)
+        return mean, (lower, upper)
+    else:
+        f = fit_res.forecast(steps=steps)
+        mean = np.asarray(f)
+        return mean, (np.full(len(mean), np.nan), np.full(len(mean), np.nan))
+# ---- report generator ----
+def generate_report_html(out_path: str, plots: List[plt.Figure], tables: Dict[str, pd.DataFrame], title="LR3 report"):
+    html_parts = [f"<html><head><meta charset='utf-8'><title>{title}</title></head><body><h1>{title}</h1>"]
+    for name, df in tables.items():
+        html_parts.append(f"<h2>{name}</h2>")
+        html_parts.append(df.to_html(classes='table table-striped', border=0))
+    for i, fig in enumerate(plots):
+        img_path = f"plot_{i}.png"
+        fig.savefig(img_path, bbox_inches='tight')
+        html_parts.append(f"<h3>Figure {i+1}</h3><img src='{img_path}' style='max-width:100%'>")
+    html_parts.append("</body></html>")
+    with open(out_path, 'w', encoding='utf-8') as f:
+        f.write("\n".join(html_parts))
+    print("Report saved to", out_path)
+# ---- main pipeline ----
+def run_pipeline(data_path: str, timestamp_col: str, target_col: str, out_report: str = 'lab3_report.html', freq='D'):
+    df = load_data(data_path, timestamp_col)
+    if target_col not in df.columns:
+        numeric_cols = df.select_dtypes(include=[np.number]).columns.tolist()
+        if not numeric_cols:
+            raise ValueError("No numeric columns found for target")
+        target_col = numeric_cols[0]
+    series = df[target_col].astype('float').copy()
+    series = resample_and_interpolate(series.to_frame(), freq=freq).iloc[:,0]
+    df_all = series.to_frame(name=target_col)
+    df_all = make_time_features(df_all)
+    df_all = make_lags(df_all, target_col, [1,2,7,30])
+    df_all = make_rolls(df_all, target_col, [7,30])
+    train, val, test = chronological_split(df_all, 0.7, 0.15)
+    y_train = train[target_col]
+    y_val = val[target_col]
+    y_test = test[target_col]
+    results = []
+    horizons = [1,7,30]
+    for h in horizons:
+        results.append({'model':'naive','h':h,'pred':naive_forecast(y_train,h)})
+        results.append({'model':'seasonal_naive','h':h,'pred':seasonal_naive_forecast(y_train, season=7, steps=h)})
+    # SARIMAX baseline
+    sarimax_res = None
+    try:
+        sarimax_res = fit_sarimax_wrapper(y_train, order=(1,1,1))
+        for h in horizons:
+            p, ci = forecast_sarimax(sarimax_res, steps=h)
+            results.append({'model':'sarimax(1,1,1)','h':h,'pred':p})
+    except Exception as e:
+        print("SARIMAX fit failed:", e)
+    # pmdarima auto_arima
+    if PM_AVAILABLE:
+        try:
+            auto = pm.auto_arima(y_train.dropna(), seasonal=False, error_action='ignore', suppress_warnings=True)
+            for h in horizons:
+                p = auto.predict(n_periods=h)
+                results.append({'model':'auto_arima','h':h,'pred':p})
+        except Exception as e:
+            print("auto_arima failed:", e)
+    # Evaluate
+    eval_tables = []
+    plots = []
+    for rec in results:
+        pred = np.asarray(rec['pred']).ravel()
+        y_true = y_test.values[:len(pred)]
+        if len(y_true)==0:
+            continue
+        eval_tables.append({
+            'model':rec['model'],
+            'h':rec['h'],
+            'MAE':mae(y_true,pred),
+            'RMSE':rmse(y_true,pred),
+            'MAPE':mape(y_true,pred),
+            'SMAPE':smape(y_true,pred)
+        })
+        fig, ax = plt.subplots(figsize=(8,3))
+        ax.plot(y_train.index[-100:], y_train.values[-100:], label='train')
+        ax.plot(y_test.index[:len(pred)], y_true, label='test')
+        ax.plot(y_test.index[:len(pred)], pred, label=f'pred_{rec["model"]}_h{rec["h"]}')
+        ax.legend()
+        plots.append(fig)
+    eval_df = pd.DataFrame(eval_tables)
+    tables = {'evaluation': eval_df}
+    generate_report_html(out_report, plots, tables, title="Lab3 Report")
+    print("Done. Report:", out_report)
+# CLI runner
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--data', required=True)
+    parser.add_argument('--timestamp', default='timestamp')
+    parser.add_argument('--target', default='target')
+    parser.add_argument('--out', default='lab3_report.html')
+    parser.add_argument('--freq', default='D')
+    args = parser.parse_args()
+    run_pipeline(args.data, args.timestamp, args.target, out_report=args.out, freq=args.freq)

src/lab4_functions.py ADDED Viewed

	@@ -0,0 +1,43 @@

+# lab4_functions.py
+"""
+Простейший набор для ЛР №4:
+- Feature engineering (lags, rolling)
+- Обучение Ridge, Lasso, RandomForest, LightGBM (опционально)
+- TimeSeriesSplit wrapper
+"""
+import numpy as np
+import pandas as pd
+from typing import List, Dict
+from sklearn.linear_model import Ridge, Lasso
+from sklearn.ensemble import RandomForestRegressor
+from sklearn.model_selection import TimeSeriesSplit, cross_val_score
+from sklearn.metrics import mean_absolute_error
+import warnings
+warnings.filterwarnings("ignore")
+try:
+    import lightgbm as lgb
+    LGB_AVAILABLE = True
+except Exception:
+    LGB_AVAILABLE = False
+def make_lag_features(df: pd.DataFrame, target:str, lags:List[int]=[1,7,30]):
+    dfc = df.copy().set_index('timestamp').sort_index()
+    for l in lags:
+        dfc[f'{target}_lag_{l}'] = dfc[target].shift(l)
+    dfc = dfc.dropna().reset_index()
+    return dfc
+def train_baselines(X_train, y_train):
+    models = {}
+    models['Ridge'] = Ridge().fit(X_train, y_train)
+    models['Lasso'] = Lasso().fit(X_train, y_train)
+    models['RF'] = RandomForestRegressor(n_estimators=100, random_state=42).fit(X_train, y_train)
+    if LGB_AVAILABLE:
+        models['LightGBM'] = lgb.LGBMRegressor(n_estimators=100).fit(X_train, y_train)
+    return models
+def cv_score_ts(model, X, y, n_splits=5, scoring='neg_mean_absolute_error'):
+    tscv = TimeSeriesSplit(n_splits=n_splits)
+    scores = cross_val_score(model, X, y, cv=tscv, scoring=scoring)
+    return scores.mean()

src/lab5_functions.py ADDED Viewed

	@@ -0,0 +1,48 @@

+# lab5_functions.py
+"""
+Минимальный LSTM example для ЛР №5 (PyTorch).
+Если у вас нет torch — функции выбросят понятную ошибку.
+"""
+import numpy as np
+import pandas as pd
+try:
+    import torch
+    import torch.nn as nn
+    TORCH_AVAILABLE = True
+except Exception:
+    TORCH_AVAILABLE = False
+if TORCH_AVAILABLE:
+    class SimpleLSTM(nn.Module):
+        def __init__(self, input_size, hidden_size=64, num_layers=1, out_size=1):
+            super().__init__()
+            self.lstm = nn.LSTM(input_size, hidden_size, num_layers, batch_first=True)
+            self.fc = nn.Linear(hidden_size, out_size)
+        def forward(self, x):
+            out, _ = self.lstm(x)
+            return self.fc(out[:, -1, :])
+    def create_sequences(series, lookback=30, horizon=1):
+        X, y = [], []
+        for i in range(len(series) - lookback - horizon + 1):
+            X.append(series[i:i+lookback])
+            y.append(series[i+lookback:i+lookback+horizon])
+        return np.array(X), np.array(y)
+    def train_lstm(series, lookback=30, epochs=20, lr=1e-3):
+        X, y = create_sequences(series, lookback=lookback, horizon=1)
+        X = torch.from_numpy(X).float().unsqueeze(-1)  # (N, L, 1)
+        y = torch.from_numpy(y).float().squeeze(-1)
+        model = SimpleLSTM(input_size=1)
+        opt = torch.optim.Adam(model.parameters(), lr=lr)
+        loss_fn = nn.MSELoss()
+        model.train()
+        for ep in range(epochs):
+            pred = model(X)
+            loss = loss_fn(pred.squeeze(), y)
+            opt.zero_grad(); loss.backward(); opt.step()
+        return model
+else:
+    def train_lstm(*args, **kwargs):
+        raise ImportError("PyTorch не установлен. Установите torch, чтобы использовать LSTM.")

src/streamlit_app.py CHANGED Viewed

@@ -15,6 +15,29 @@ from statsmodels.tsa.seasonal import seasonal_decompose
 from statsmodels.stats.outliers_influence import variance_inflation_factor
 from statsmodels.tools import add_constant
 st.set_page_config(page_title="Анализ временных рядов", layout="wide", initial_sidebar_state="expanded")
 MOSCOW = pytz.timezone("Europe/Moscow")
@@ -33,15 +56,39 @@ except ImportError as e:
     LAB2_AVAILABLE = False
     st.warning(f"Функции ЛР №2 недоступны: {e}")
 # Навигация между лабораторными работами
 st.sidebar.title("🧪 Лабораторные работы")
 lab_choice = st.sidebar.radio(
     "Выберите лабораторную работу:",
-    ["ЛР №1: Введение в анализ временных рядов", "ЛР №2: Прогнозирование временных рядов"],
     index=0
 )
 # ---------------- Utilities ----------------
 def detect_date_column(df: pd.DataFrame) -> Optional[str]:
     candidates = [c for c in df.columns if any(k in c.lower() for k in ("date", "time", "timestamp", "dt", "day"))]
@@ -1876,11 +1923,320 @@ def render_lab2():
     else:
         st.warning("⚠️ Сначала постройте модели в разделе 'Этап 6-7', чтобы увидеть результаты сравнения.")
 # ---------------- Главный код: выбор лабораторной работы ----------------
 if lab_choice == "ЛР №1: Введение в анализ временных рядов":
     render_lab1()
 elif lab_choice == "ЛР №2: Прогнозирование временных рядов":
     render_lab2()
 else:
     st.info("Выберите лабораторную работу в боковой панели")

 from statsmodels.stats.outliers_influence import variance_inflation_factor
 from statsmodels.tools import add_constant
+import sys
+import os
+# Добавляем папку с модулями (путь к файлу streamlit_app.py)
+# Это гарантирует, что Python найдёт lab3_pipeline и lab3_functions в той же папке src
+sys.path.append(os.path.dirname(os.path.abspath(__file__)))
+try:
+    import lab3_functions as lab3
+    from lab3_pipeline import run_pipeline
+    LAB3_AVAILABLE = True
+except Exception as e:
+    LAB3_AVAILABLE = False
+    lab3 = None
+    _LAB3_IMPORT_ERROR = str(e)
+    # Для дебага выведем в консоль (и в Streamlit, если он загружен)
+    print("lab3 import failed:", _LAB3_IMPORT_ERROR)
+    try:
+        import streamlit as _st
+        _st.warning("lab3 import failed: " + _LAB3_IMPORT_ERROR)
+    except Exception:
+        pass
 st.set_page_config(page_title="Анализ временных рядов", layout="wide", initial_sidebar_state="expanded")
 MOSCOW = pytz.timezone("Europe/Moscow")
     LAB2_AVAILABLE = False
     st.warning(f"Функции ЛР №2 недоступны: {e}")
+# после импорта lab2_functions
+try:
+    import lab3_functions as lab3
+    LAB3_AVAILABLE = True
+except Exception:
+    LAB3_AVAILABLE = False
+try:
+    import lab4_functions as lab4
+    LAB4_AVAILABLE = True
+except Exception:
+    LAB4_AVAILABLE = False
+try:
+    import lab5_functions as lab5
+    LAB5_AVAILABLE = True
+except Exception:
+    LAB5_AVAILABLE = False
 # Навигация между лабораторными работами
 st.sidebar.title("🧪 Лабораторные работы")
 lab_choice = st.sidebar.radio(
     "Выберите лабораторную работу:",
+    ["ЛР №1: Введение в анализ временных рядов",
+     "ЛР №2: Прогнозирование временных рядов",
+     "ЛР №3: Классические модели",
+     "ЛР №4: ML для TS",
+     "ЛР №5: Deep Learning для TS"
+    ],
     index=0
 )
 # ---------------- Utilities ----------------
 def detect_date_column(df: pd.DataFrame) -> Optional[str]:
     candidates = [c for c in df.columns if any(k in c.lower() for k in ("date", "time", "timestamp", "dt", "day"))]
     else:
         st.warning("⚠️ Сначала постройте модели в разделе 'Этап 6-7', чтобы увидеть результаты сравнения.")
+import streamlit.components.v1 as components
+import pandas as pd
+from pathlib import Path
+def render_lab3():
+    st.title("🧪 ЛР №3: Сравнительный анализ классических моделей")
+    if not LAB3_AVAILABLE:
+        st.error("Модуль lab3_functions / lab3_pipeline недоступен. Добавьте файлы в папку src.")
+        if '_LAB3_IMPORT_ERROR' in globals():
+            st.write("Ошибка импорта:", _LAB3_IMPORT_ERROR)
+        return
+    st.info("Загрузи/выбери датасет в ЛР1 и запусти preprocess — затем сюда.")
+    if 'df_clean' not in st.session_state:
+        st.info("Сначала выполните ЛР1: загрузите и предобработайте датасет.")
+        return
+    df = st.session_state['df_clean'].copy()
+    numeric_cols = df.select_dtypes(include=[np.number]).columns.tolist() if 'pd' in globals() else df.select_dtypes(include=['number']).columns.tolist()
+    if not numeric_cols:
+        st.error("Нет числовых колонок.")
+        return
+    target = st.selectbox("Выберите target для ЛР3", numeric_cols)
+    # Убедимся, что timestamp есть в датафрейме (если он в индексе — reset_index сделает колонку)
+    if 'timestamp' not in df.columns:
+        try:
+            df_reset = df.reset_index()
+            if 'timestamp' in df_reset.columns:
+                df = df_reset
+            else:
+                st.error("Колонка 'timestamp' не найдена. Убедитесь, что в df_clean есть временная метка.")
+                return
+        except Exception:
+            st.error("Не удалось найти колонку 'timestamp'.")
+            return
+    series = df.set_index('timestamp')[target].dropna()
+    st.write("Длина ряда:", len(series))
+    # ---- существующие экспандеры ARIMA/VAR оставляем без изменений ----
+    # === ARIMA / auto_arima expander ===
+    with st.expander("ARIMA / auto_arima (SARIMAX)"):
+        import numpy as _np
+        import pandas as _pd
+        import matplotlib.pyplot as _plt
+        use_auto = st.checkbox("Использовать auto_arima (pmdarima)", value=False)
+        col_freq = st.selectbox("Частота ряда (inference если None)", ["auto", "D", "W", "M"], index=0)
+        if use_auto:
+            max_p = st.number_input("max_p (для auto_arima)", min_value=0, max_value=10, value=3, step=1)
+            max_d = st.number_input("max_d", min_value=0, max_value=3, value=2, step=1)
+            max_q = st.number_input("max_q", min_value=0, max_value=10, value=3, step=1)
+            seasonal = st.checkbox("Ищем сезонность (seasonal)", value=False)
+            seasonal_period = st.number_input("m (период сезонности, если seasonal=True)", min_value=0, max_value=366, value=7, step=1)
+            if st.button("Запустить auto_arima"):
+                try:
+                    with st.spinner("Подбираю модель auto_arima..."):
+                        model = lab3.fit_auto_arima(series, seasonal=seasonal, m=int(seasonal_period),
+                                                    start_p=0, start_q=0, max_p=int(max_p), max_q=int(max_q),
+                                                    max_d=int(max_d))
+                    st.success("auto_arima выполнен")
+                    # pmdarima object имеет summary или order/seasonal_order
+                    try:
+                        st.text(str(model.summary()))
+                    except Exception:
+                        st.write("Выбран порядок:", getattr(model, "order", None))
+                        st.write("seasonal_order:", getattr(model, "seasonal_order", None))
+                    st.session_state['_lab3_last_model'] = ("auto_arima", model)
+                except Exception as e:
+                    st.error("Ошибка auto_arima: " + str(e))
+        else:
+            # ручная SARIMAX
+            st.write("Параметры SARIMAX (ручной режим)")
+            p = st.number_input("p (AR)", min_value=0, max_value=10, value=1, step=1)
+            d = st.number_input("d (diff)", min_value=0, max_value=3, value=0, step=1)
+            q = st.number_input("q (MA)", min_value=0, max_value=10, value=0, step=1)
+            sp = st.number_input("P (seasonal AR)", min_value=0, max_value=5, value=0, step=1)
+            sd = st.number_input("D (seasonal diff)", min_value=0, max_value=2, value=0, step=1)
+            sq = st.number_input("Q (seasonal MA)", min_value=0, max_value=5, value=0, step=1)
+            s = st.number_input("s (seasonal period)", min_value=0, max_value=366, value=7, step=1)
+            if st.button("Построить SARIMAX (ручной)"):
+                try:
+                    with st.spinner("Обучаю SARIMAX..."):
+                        res = lab3.fit_sarimax(series, order=(int(p), int(d), int(q)),
+                                               seasonal_order=(int(sp), int(sd), int(sq), int(s)))
+                    st.success("SARIMAX обучен")
+                    try:
+                        st.text(res.summary().as_text())
+                    except Exception:
+                        st.write("Модель успешно обучена. Объект результата:", type(res))
+                    st.session_state['_lab3_last_model'] = ("sarimax", res)
+                except Exception as e:
+                    st.error("Ошибка SARIMAX: " + str(e))
+        # Forecast controls (общие)
+        st.markdown("**Прогноз (forecast)**")
+        fh = st.slider("Горизонт прогнозирования (h)", min_value=1, max_value=90, value=7)
+        if st.button("Сделать прогноз (последние точки + future)"):
+            if '_lab3_last_model' not in st.session_state:
+                st.error("Сначала обучите модель (auto_arima или SARIMAX).")
+            else:
+                mtype, mobj = st.session_state['_lab3_last_model']
+                try:
+                    if mtype == "auto_arima":
+                        # pmdarima.predict возвращает массив
+                        preds = mobj.predict(n_periods=fh)
+                        # создаём индекс для будущих дат
+                        if col_freq == "auto":
+                            freq = _pd.infer_freq(series.index) or "D"
+                        else:
+                            freq = col_freq
+                        last = series.index[-1]
+                        future_index = _pd.date_range(start=last + _pd.tseries.frequencies.to_offset(freq), periods=fh, freq=freq)
+                        pred_series = _pd.Series(preds, index=future_index, name=f"pred_auto_arima_h{fh}")
+                        st.line_chart(_pd.concat([series.tail(200), pred_series]))
+                        st.write(pred_series.to_frame("forecast"))
+                    else:
+                        # SARIMAX result
+                        mean, (low, high) = lab3.forecast_sarimax(mobj, steps=fh)
+                        if col_freq == "auto":
+                            freq = _pd.infer_freq(series.index) or "D"
+                        else:
+                            freq = col_freq
+                        last = series.index[-1]
+                        future_index = _pd.date_range(start=last + _pd.tseries.frequencies.to_offset(freq), periods=fh, freq=freq)
+                        pred_series = _pd.Series(mean, index=future_index)
+                        low_s = _pd.Series(low, index=future_index)
+                        high_s = _pd.Series(high, index=future_index)
+                        fig, ax = _plt.subplots(figsize=(10, 4))
+                        ax.plot(series.tail(200).index, series.tail(200).values, label="history")
+                        ax.plot(pred_series.index, pred_series.values, label="forecast")
+                        ax.fill_between(pred_series.index, low_s.values, high_s.values, alpha=0.25, label="CI")
+                        ax.legend()
+                        st.pyplot(fig)
+                        st.write(pred_series.to_frame("forecast"))
+                except Exception as e:
+                    st.error("Ошибка при прогнозе: " + str(e))
+        # Residuals diagnostics
+        if st.checkbox("Показать диагностику остатков (последняя обученная модель)"):
+            if '_lab3_last_model' not in st.session_state:
+                st.info("Нет обученной модели в сессии.")
+            else:
+                try:
+                    mtype, mobj = st.session_state['_lab3_last_model']
+                    if mtype == "sarimax":
+                        resid = mobj.resid.dropna()
+                    elif mtype == "auto_arima":
+                        # у pmdarima есть attribute resid_ или можно считать inplace prediction
+                        try:
+                            resid = _pd.Series(mobj.resid())
+                        except Exception:
+                            resid = None
+                    else:
+                        resid = None
+                    if resid is None or len(resid) == 0:
+                        st.warning("Не удалось получить остатки для этой модели.")
+                    else:
+                        fig1, ax1 = _plt.subplots(1, 1, figsize=(8, 2.5))
+                        ax1.plot(resid)
+                        ax1.set_title("Остатки (time series)")
+                        st.pyplot(fig1)
+                        fig2, ax2 = _plt.subplots(1, 1, figsize=(6, 3))
+                        ax2.hist(resid, bins=30)
+                        ax2.set_title("Гистограмма остатков")
+                        st.pyplot(fig2)
+                        # Ljung-Box, если доступен
+                        try:
+                            from statsmodels.stats.diagnostic import acorr_ljungbox
+                            lb = acorr_ljungbox(resid, lags=[10], return_df=True)
+                            st.write("Ljung-Box (lag=10):")
+                            st.write(lb)
+                        except Exception:
+                            st.info("acorr_ljungbox недоступен в окружении.")
+                except Exception as e:
+                    st.error("Ошибка диагностики: " + str(e))
+    # === VAR (multivariate) expander ===
+    with st.expander("VAR (векторная авторегрессия) — multivariate"):
+        import pandas as _pd
+        import matplotlib.pyplot as _plt
+        st.write("Выберите колонки для VAR (минимум 2):")
+        var_cols = st.multiselect("Колонки для VAR", options=numeric_cols, default=numeric_cols[:3])
+        maxlags = st.number_input("maxlags (select_order)", min_value=1, max_value=20, value=5, step=1)
+        if st.button("Построить VAR"):
+            if len(var_cols) < 2:
+                st.error("Нужно хотя бы 2 колонки для VAR.")
+            else:
+                try:
+                    df_var = df.set_index('timestamp')[var_cols].dropna()
+                    with st.spinner("Обучаю VAR..."):
+                        var_res = lab3.fit_var(df_var, maxlags=int(maxlags))
+                    st.success("VAR обучен, lag order = %s" % getattr(var_res, "k_ar", "unknown"))
+                    try:
+                        st.text(var_res.summary().as_text())
+                    except Exception:
+                        st.write("VAR results object:", type(var_res))
+                    st.session_state['_lab3_last_var'] = var_res
+                except Exception as e:
+                    st.error("Ошибка VAR: " + str(e))
+        # Forecast for VAR
+        st.markdown("**Прогноз VAR**")
+        if st.button("Сделать VAR-прогноз (30 шагов)"):
+            if '_lab3_last_var' not in st.session_state:
+                st.error("Сначала обучите VAR.")
+            else:
+                try:
+                    var_res = st.session_state['_lab3_last_var']
+                    steps = st.number_input("h (шагов вперёд)", min_value=1, max_value=365, value=30, step=1)
+                    fut = lab3.forecast_var(var_res, steps=int(steps))
+                    st.write("Прогноз (DataFrame):")
+                    st.dataframe(fut)
+                    # Нарисуем прогноз по каждой переменной вместе с последними историческими точками
+                    df_var = df.set_index('timestamp')[var_cols].dropna()
+                    last_hist = df_var.tail(50)
+                    fig, ax = _plt.subplots(len(var_cols), 1, figsize=(10, 3*len(var_cols)), sharex=True)
+                    if len(var_cols) == 1:
+                        ax = [ax]
+                    for i, col in enumerate(var_cols):
+                        ax[i].plot(last_hist.index, last_hist[col].values, label=f"history_{col}")
+                        # построим future index
+                        freq = _pd.infer_freq(df_var.index) or "D"
+                        future_index = _pd.date_range(start=df_var.index[-1] + _pd.tseries.frequencies.to_offset(freq),
+                                                      periods=len(fut), freq=freq)
+                        ax[i].plot(future_index, fut[col].values, label=f"forecast_{col}")
+                        ax[i].legend()
+                    st.pyplot(fig)
+                except Exception as e:
+                    st.error("Ошибка при VAR-прогнозе: " + str(e))
+    # --- Новый блок: полный pipeline (run_pipeline) ---
+    st.markdown("---")
+    st.subheader("Запустить полный pipeline ЛР3 (генерация HTML-отчёта)")
+    out_report = st.text_input("Путь для отчёта (HTML)", "./lab3_report.html")
+    freq = st.selectbox("Частота ресемплинга", ['D','W','M'])
+    if st.button("Запустить полный pipeline и сформировать отчёт"):
+        tmp = "./_lab3_input_tmp.csv"
+        # сохраняем df с колонкой timestamp
+        df.to_csv(tmp, index=False)
+        st.info("Запускаю pipeline — это может занять время.")
+        try:
+            # run_pipeline ожидает путь к файлу
+            run_pipeline(tmp, timestamp_col='timestamp', target_col=target, out_report=out_report, freq=freq)
+            st.success("Готово — отчёт сохранён: " + out_report)
+            if Path(out_report).exists():
+                with open(out_report, 'r', encoding='utf-8') as f:
+                    html = f.read()
+                components.html(html, height=800, scrolling=True)
+            else:
+                st.info("Отчёт создан, но файл не найден по указанному пути: " + out_report)
+        except Exception as e:
+            st.error("Ошибка при выполнении pipeline: " + str(e))
+def render_lab4():
+    st.title("🧪 ЛР №4: ML-модели")
+    if not LAB4_AVAILABLE:
+        st.error("Модуль lab4_functions недоступен.")
+        return
+    if 'df_clean' not in st.session_state:
+        st.info("Сначала выполните ЛР1: загрузите и предобработайте датасет.")
+        return
+    df = st.session_state['df_clean']
+    numeric_cols = df.select_dtypes(include=[np.number]).columns.tolist()
+    target = st.selectbox("Target", numeric_cols)
+    if st.button("Сгенерировать простые лаги и обучить модели"):
+        dfr = lab4.make_lag_features(df, target, lags=[1,7,30])
+        X = dfr.drop(columns=['timestamp', target])
+        y = dfr[target]
+        models = lab4.train_baselines(X, y)
+        st.write("Обученные модели:", list(models.keys()))
+        for name, m in models.items():
+            st.write(name, getattr(m, 'score', lambda X,y: None)(X, y) )
+def render_lab5():
+    st.title("🧪 ЛР №5: Deep Learning")
+    if not LAB5_AVAILABLE:
+        st.error("Модуль lab5_functions недоступен.")
+        return
+    if 'df_clean' not in st.session_state:
+        st.info("Сначала выполните ЛР1.")
+        return
+    df = st.session_state['df_clean']
+    numeric_cols = df.select_dtypes(include=[np.number]).columns.tolist()
+    target = st.selectbox("Target для DL", numeric_cols)
+    if st.button("Обучить простой LSTM (demo)"):
+        s = df.set_index('timestamp')[target].dropna().values
+        try:
+            model = lab5.train_lstm(s, lookback=30, epochs=5)
+            st.success("Модель обучена (demo).")
+        except Exception as e:
+            st.error(str(e))
 # ---------------- Главный код: выбор лабораторной работы ----------------
 if lab_choice == "ЛР №1: Введение в анализ временных рядов":
     render_lab1()
 elif lab_choice == "ЛР №2: Прогнозирование временных рядов":
     render_lab2()
+elif lab_choice == "ЛР №3: Классические модели":
+    render_lab3()
+elif lab_choice == "ЛР №4: ML для TS":
+    render_lab4()
+elif lab_choice == "ЛР №5: Deep Learning для TS":
+    render_lab5()
 else:
     st.info("Выберите лабораторную работу в боковой панели")

src/streamlit_lab3.py ADDED Viewed

	@@ -0,0 +1,50 @@

+"""
+streamlit_lab3.py
+Упрощённый Streamlit-интерфейс для ЛР №3.
+Позволяет загрузить CSV/Parquet, настроить параметры и запустить pipeline (lab3_pipeline.run_pipeline).
+"""
+import streamlit as st
+import os
+import pandas as pd
+import tempfile
+from pathlib import Path
+# импортируем наш pipeline (если он в PATH)
+try:
+    from lab3_pipeline import run_pipeline
+    PIPELINE_AVAILABLE = True
+except Exception as e:
+    PIPELINE_AVAILABLE = False
+    st = None
+def main():
+    st.title('ЛР №3 — Временные ряды (pipeline)')
+    uploaded = st.file_uploader("Загрузите CSV/Parquet с колонкой timestamp и target", type=['csv','parquet'])
+    if uploaded is None:
+        st.info("Загрузите файл, или введите путь к локальному файлу на сервере.")
+    else:
+        with tempfile.NamedTemporaryFile(delete=False, suffix=Path(uploaded.name).suffix) as tmp:
+            tmp.write(uploaded.getbuffer())
+            tmp_path = tmp.name
+        st.write("Файл загружен:", tmp_path)
+        timestamp = st.text_input("Имя колонки с временной меткой", "timestamp")
+        target = st.text_input("Имя колонки с целевой переменной", "target")
+        freq = st.selectbox("Частота ресемплинга", ['D','W','M'])
+        outpath = './lab3_report.html'
+        if st.button("Запустить ЛР3"):
+            st.info("Запуск pipeline — может занять время. Выходной файл: " + outpath)
+            try:
+                run_pipeline(tmp_path, timestamp, target, out_report=outpath, freq=freq)
+                with open(outpath, 'r', encoding='utf-8') as f:
+                    html = f.read()
+                st.markdown("### Отчёт")
+                st.components.v1.html(html, height=800, scrolling=True)
+            except Exception as e:
+                st.error("Ошибка при выполнении pipeline: " + str(e))
+if __name__ == '__main__':
+    if PIPELINE_AVAILABLE:
+        main()
+    else:
+        print("lab3_pipeline недоступен. Положите lab3_pipeline.py в PYTHONPATH или /mnt/data")