y[j SrSSKrSSKrSSKrSSKJr SSKrSSKr SSK r Sr \"\ 5R5RRr\S- S- S- r\S- S- rS S S S S SS.r/SQr/SQrSS\S\S\S\S-4SjjrS\S\S\ R0S-4SjrS\ R0S\ R04SjrS\4Sjr\S:Xa\R<"\"55 gg)u Punto 1 della roadmap: scarica lo storico partite + quote da football-data.co.uk e produce un dataset unico e pulito pronto per il modello base e il backtesting. Uso: python scripts/download_data.py # default: 6 leghe, stagioni 2017/18 -> 2025/26 python scripts/download_data.py --leagues I1 E0 # solo Serie A e Premier League python scripts/download_data.py --seasons 2223 2324 2425 # stagioni specifiche python scripts/download_data.py --force # ri-scarica anche i file già in cache Output: data/raw/football-data/_
.csv (CSV originali, cache) data/processed/matches.csv (dataset unico normalizzato) N)Pathz:https://www.football-data.co.uk/mmz4281/{season}/{div}.csvdatarawz football-data processedzSerie AzSerie BzPremier LeaguezLa Liga BundesligazLigue 1)I1I2E0SP1D1F1) 171818191920202121222223232424252526)3DivDateTimeHomeTeamAwayTeamFTHGFTAGFTRHTHGHTAGHTRHSASHSTASTHCACHFAFHYAYHRARB365HB365DB365APSHPSDPSAAvgHAvgDAvgAMaxHMaxDMaxAB365CHB365CDB365CAPSCHPSCDPSCAAvgCHAvgCDAvgCAMaxCHMaxCDMaxCAzB365>2.5zB365<2.5zAvg>2.5zAvg<2.5seasondivforcereturnc [RSSS9 [USUS3- nUR5(a!U(d[SUR35 U$[ R XS9n[S5Hn[R"USS 9nURS :Xa[S US US 35 gUR5 URUR5 [SURS[UR5S-S35 Us $ [SU35 g![Ra>n[SUS-SUSU35 [ R""SUS--5 SnAMSnAff=f)zLScarica un CSV (con cache locale) e restituisce il path, o None se fallisce.Tparentsexist_ok_z.csvz [cache] )rFrG)timeoutiz [404] /z non esiste, saltoNz [ok]  (iz KB)z [retry z/3] : z [FAIL] )RAW_DIRmkdirexistsprintnameBASE_URLformatrangerequestsget status_coderaise_for_status write_bytescontentlenRequestExceptiontimesleep)rFrGrHdesturlattemptrespexcs 0C:\wamp64\www\Tigertips\scripts\download_data.py download_csvroAsZ MM$M. xqT* *D {{}}U 499+&' /// 1C8 *<<R0D3& 6(!C50BCD  ! ! #   T\\ * JtyykC ,=,E+FdK LK Jse   (( * Igk]$se2cU; < JJqGaK( ) ) *s<7D'5AD''E9;3E44E9pathcUR5SSS:XaSOSn[R"XSS9nURRR 5UlUR/S QS S 9nUR(ag[R"URS 9n[H(nXcR;aX6O[RXV'M* SUSSSUSS3US'[R "USSSSS9US'URS/S9nU$![ a%n[SURS U35 SnAgSnAff=f)z>Legge un CSV grezzo e lo normalizza alle colonne KEEP_COLUMNS.NrOsz utf-8-sigzlatin-1skip)encoding on_bad_linesz [ERRORE lettura] rU)rrrany)subsethowindex20rVrRSeasonrTmixedcoerce)dayfirstr]errors)rv) read_bytespdread_csvcolumnsstrstrip ExceptionrZr[dropnaempty DataFramery KEEP_COLUMNSnpnan to_datetime)rprFrsdfrmoutcols rnload_csvr[s*#'//"3BQ"7?"J;PY [[v FZZ^^))+ : FB xx ,,RXX &C!ZZ/27RVV AfQRj\2CM..VtGT\]CK **VH* %C J#  #DII;b67sAD D>D99D>rc/SQn[R"[RURS9n[R"[RURS9n[R"[RURS9n[R"[R URSS9nUHupgp[R "XSS9n [R "XSS9n [R "XSS9n U RS5U RS5-U RS5-UR5-n U R5(dMSU - SU - SU - npX-U-nUU- U X-'UU- U X='UU- U XM'XU 'M URS5US 'URS5US 'URS5US 'XPS 'U$) u Converte le quote in probabilità implicite normalizzate (senza margine del bookmaker). Priorità: Pinnacle closing > media closing > Bet365 closing > Pinnacle pre-match > media pre-match > Bet365 pre-match. ))r=r>r?pinnacle_closing)r@rArB avg_closing)r:r;r< b365_closing)r1r2r3pinnacle)r4r5r6avg)r.r/r0b365rxstring)rydtyper})rrTImpProbHImpProbDImpProbA ImpProbSource) rSeriesrrryNA to_numericgtisnaruround)rpriorityprob_hprob_dprob_asourcecol_hcol_dcol_ar[odds_hodds_dodds_avalidraw_hraw_draw_a overrounds rnadd_implied_probabilitiesrws HYYrvvRXX .F YYrvvRXX .F YYrvvRXX .F YYruuBHHH =F%-!ery:ry:ry: ! vyy|+fiil:V[[]Jyy{{ &j!f*a&jeuME) *E2 *E2 *E2 u &.\\!_BzN\\!_BzN\\!_BzN  Ic 8[R"SS9nURSS[[5[[5SSSR [5S3S 9 URS S[ S S S 9 URSSSS9 UR5n[S[UR5S[UR5S35 /nURHPnURH=n[X4URS9nUcM[XS5nUcM,URU5 M? MR U(d [S5 g[ R""USS9nUR%SS/5R'SS9n[)U5n[*R-SSS9 [*S- nUR/USS 9 [S!5 [S"[U535 [S#USR15R35S$USR55R3535 [S%US&R75R95S'US&R75R;5S(S35 US)R=/S*Q5n [S+U R95S'U R;5S(S35 [S,5 USR?5RA5H*upJ[S-[RCXD5S.SU 35 M, [S/U35 g0)1Nz6Scarica storico partite + quote da football-data.co.uk) descriptionz --leagues+DIVzCodici lega (default:  ))nargsdefaultchoicesmetavarhelpz --seasonsSSEEzDStagioni in formato '2324' per 2023/24 (default: 2017/18 -> 2025/26))rrrrz--force store_trueuRi-scarica i file già in cache)actionrz Download: z leghe x z stagioni)rHzNessun dato scaricato.rTT) ignore_indexrr)droprKz matches.csvFrxz === Riepilogo dataset ===zPartite totali: zPeriodo: z -> zCon quote (1X2): rrSz.1%r)rrrzCon quote closing: z Partite per lega:z z<16z Salvato in: r)"argparseArgumentParser add_argumentlistLEAGUESjoinDEFAULT_SEASONS parse_argsrZreleaguesseasonsrorHrappendrconcat sort_values reset_indexr PROCESSED_DIRrXto_csvmindatemaxnotnasummeanisin value_countsitemsr`) parserargsframesrFrGrprmatchesout_pathclosingcounts rnmainrs  $ $1i jF  3W tT[} %.DSXXgEVDWWX,Y[  3QWce  ,=^_    D Js4<<()3t||3D2EY OP F,,<D|$'B~ b!    &'iiT2G!!65/2>>D>IG'0Gt4},H NN85N) '( !#g, 01 !'&/"5"5"7"<"<">!?tGFODWDWDYD^D^D`Ca bc !'*"5";";"="A"A"C!DEj!'')..0 5Q 89o&++,_`G !'++-7<<>#2Fa HI  en11399;  7;;s(-Qug67< N8* %& r__main__)F)__doc__rsysrgpathlibrnumpyrpandasrr_r\__file__resolveparent PROJECT_ROOTrWrrrrrboolrorrrintr__name__exitrrnrs   GH~%%'..55  % '/ 9v% 3         [ $3tt 44)<8&",,&2<<&R, c, ^ zHHTVr