import pandas as pd,numpy as np,glob,os,re
pd.set_option('display.width',250);pd.set_option('display.max_rows',200)
out=open('results/options_report.txt','w')
def pr(*a):
    s=' '.join(str(x) for x in a); print(s); out.write(s+'\n')
def load(pattern):
    fr=[]
    for f in glob.glob(pattern):
        d=pd.read_parquet(f); m=re.search(r'_(\d{4}-\d\d-\d\d)_(\d+)(CE|PE)\.parquet',f)
        d['expiry']=m.group(1); d['strike']=int(m.group(2)); d['type']=m.group(3); fr.append(d[['ts','close','volume','oi','expiry','strike','type']])
    return pd.concat(fr)
LOT=65
cfg=[('2026-10-23','../lows/data/opt/2026-10-23_2026-10-27_*.parquet',[('13:20','13:37'),('13:37','13:41'),('13:41','13:47'),('13:47','13:54'),('13:54','14:10')]),
     ('2026-10-13','data/opt_2026-10-13_*.parquet',[('09:15','09:30'),('09:30','10:00'),('14:15','14:30'),('14:30','14:45'),('14:45','15:00'),('15:00','15:14')])]
for day,pat,phases in cfg:
    o=load(pat); o['ts']=pd.to_datetime(o.ts)
    pr('\n=====',day,'contracts',o.groupby(['expiry','type']).size().to_dict())
    rows=[]
    for e,g in o.groupby('expiry'):
        for a,b in phases:
            ta,tb=pd.Timestamp(f'{day} {a}'),pd.Timestamp(f'{day} {b}')
            for t in ['CE','PE']:
                gg=g[g.type==t]
                piv=gg.pivot_table(index='ts',columns='strike',values='oi').ffill()  # OI carried forward only between prints of the same contract (no fabrication of price)
                vol=gg[(gg.ts>ta)&(gg.ts<=tb)].groupby('strike').volume.sum()
                if ta not in piv.index or tb not in piv.index: continue
                doi=(piv.loc[tb]-piv.loc[ta])/LOT
                px=gg.pivot_table(index='ts',columns='strike',values='close')
                for k in doi.index:
                    rows.append(dict(expiry=e,phase=f'{a}->{b}',type=t,strike=k,dOI_lots=doi[k],vol_lots=vol.get(k,0)/LOT,px_a=px[k].get(ta,np.nan),px_b=px[k].get(tb,np.nan)))
    R=pd.DataFrame(rows); R.to_csv(f'results/options_strike_phases_{day}.csv',index=False)
    # bands relative to spot at phase start
    spot=pd.read_csv('../data/1m/NIFTY_1m.csv',parse_dates=['ts']); spot['ts']=spot.ts.dt.tz_localize(None); spot=spot.set_index('ts').close
    S=[]
    for (e,ph,t),g in R.groupby(['expiry','phase','type']):
        s0=spot[pd.Timestamp(f"{day} {ph[:5]}")]; atm=round(s0/50)*50
        itm=g[(g.strike<atm) if t=='CE' else (g.strike>atm)]; otm=g[(g.strike>atm) if t=='CE' else (g.strike<atm)]; at=g[g.strike==atm]
        S.append(dict(expiry=e,phase=ph,type=t,atm=atm,dOI_ATM=at.dOI_lots.sum(),dOI_OTM=otm.dOI_lots.sum(),dOI_ITM=itm.dOI_lots.sum(),dOI_all=g.dOI_lots.sum(),vol_all=g.vol_lots.sum()))
    S=pd.DataFrame(S); S.to_csv(f'results/options_bands_{day}.csv',index=False); pr(S.round(0).to_string(index=False))
    # top strikes in the key phase
    key='13:37->13:41' if day=='2026-10-23' else '09:15->09:30'
    for kp in ([key,'13:41->13:47'] if day=='2026-10-23' else [key,'14:30->14:45']):
        x=R[R.phase==kp].copy(); x=x.reindex(x.dOI_lots.abs().sort_values(ascending=False).index).head(10); pr('\nTop |dOI| strikes',kp); pr(x.round(1).to_string(index=False))
out.close()