"""Kepler Bonus background-M-star pilot. Exploratory, not a planet classifier. Dates are explicitly converted BKJD -> BTJD by subtracting 2167 days. Last ceil(25%) usable quarters reserved before period search. Flux selection uses documented PSF FLUX, quality==0, positive errors and PSFFRAC>=0.5. The search uses an assumed rho=5 solar (not a measured stellar parameter). No radii/planet validation are inferred from that approximation. """ import os for k in ('OMP_NUM_THREADS','OPENBLAS_NUM_THREADS','MKL_NUM_THREADS'):os.environ[k]='1' import argparse,hashlib,io,json,sys,time,traceback from pathlib import Path from concurrent.futures import ProcessPoolExecutor,as_completed import numpy as np,pandas as pd,requests from astropy.io import fits ROOT=Path(__file__).resolve().parent PROJECT=ROOT.parents[1] sys.path.insert(0,str(PROJECT)) import run_search as R from improvements.independent_vet import vet import fastbls _seasons=fastbls.seasons def short_blocks(t,gap=20.,max_len=None):return _seasons(t,gap=gap,max_len=90. if max_len is None else max_len) fastbls.seasons=short_blocks def dump(p,d): p.write_text(json.dumps(d,indent=2,default=lambda x:x.item() if hasattr(x,'item') else str(x))) def select(limit): d=pd.read_csv(ROOT/'eligible_before_catalogcheck.csv',dtype={'psf_avail':str}) primary=pd.read_csv(ROOT.parent/'kepler_primary_targets.csv') d=d[~d.kic.isin(primary.kepid)] # Frozen catalog-based priority, before looking at any photometry. d=d.sort_values(['phot_g_mean_mag','gaia_designation']).head(limit).copy() d['source_id']=d.gaia_designation.str.split().str[-1] d.to_csv(ROOT/'pilot_targets.csv',index=False) req={'service':'Mast.Caom.Filtered','format':'json','pagesize':50000,'page':1,'params':{'columns':'*','filters':[{'paramName':'provenance_name','values':['KBONUS-BKG']},{'paramName':'target_name','values':d.gaia_designation.tolist()}]}} r=requests.post('https://mast.stsci.edu/api/v0/invoke',data={'request':json.dumps(req)},timeout=90);r.raise_for_status();j=r.json() if j.get('status')!='COMPLETE':raise ValueError(j.get('status')) dump(ROOT/'pilot_observations.json',j) obs=pd.DataFrame(j['data']);obs=obs.drop_duplicates('target_name');obs.to_csv(ROOT/'pilot_observations.csv',index=False) return d.merge(obs[['target_name','dataURL']],left_on='gaia_designation',right_on='target_name') def acquire(row,min_quarters=6): sid=str(row['source_id']);npz=ROOT/'lightcurves'/f'{sid}.npz';fp=ROOT/'fits'/f'{sid}.fits' if npz.exists():return dict(np.load(npz)) url='https://mast.stsci.edu/api/v0.1/Download/file?uri='+row['dataURL'] if not fp.exists(): r=requests.get(url,timeout=90);r.raise_for_status() if r.content[:6]!=b'SIMPLE':raise ValueError('Not FITS') fp.write_bytes(r.content) parts=[];quarters=[] with fits.open(fp,memmap=False) as h: exp=float(h[0].header.get('XPOSURE',1765.464)) for hd in h: if not hd.name.startswith('LIGHTCURVE_Q'):continue q=int(hd.header['QUARTER']);frac=float(hd.header.get('PSFFRAC',0));d=hd.data if frac<.5:continue t=np.asarray(d['TIME'],float)-2167;f=np.asarray(d['FLUX'],float);e=np.asarray(d['FLUX_ERR'],float) good=(d['SAP_QUALITY']==0)&np.isfinite(t)&np.isfinite(f)&np.isfinite(e)&(e>0) if good.sum()<300:continue med=np.median(f[good]) if med<=0:continue n=good.sum();sap=np.asarray(d['SAP_FLUX'],float);sm=np.nanmedian(sap[good]) parts.append({'time':t[good],'flux':f[good]/med,'flux_err':e[good]/med,'sector':np.full(n,q),'exptime':np.full(n,exp),'sap_flux':sap[good]/sm if sm>0 else np.full(n,np.nan),'cx':np.asarray(d['CENTROID_COLUMN'])[good],'cy':np.asarray(d['CENTROID_ROW'])[good]}) quarters.append({'quarter':q,'psffrac':frac,'crowdsap':float(hd.header.get('CROWDSAP',0)),'points':int(n)}) if len(parts)=9 and st['n_good_events']>=3 and ht['n_good_events']>=3 and ht['snr']>=4 and st['odd_even_sigma']<3 and abs(st['secondary']/st['secondary_err'])<3 and st['max_ses_frac']<.5) s={'rank':rank+1,'period':p,'t0_btjd':t0,'duration_hr':dur*24,'depth_ppm':depth*1e6,'bls_snr':snr,'discovery':st,'holdout':ht,'screen_pass':passed} if passed: s['independent_vet']=vet(raw,p,t0,dur,holdout=tuple(hold),controls=24,holdout_was_unseen=True) # Independent extraction, not an independent telescope. No BLS fitting on SAP. sap=dict(raw);sap['flux']=d['sap_flux'];valid=np.isfinite(sap['flux']);sap={k:v[valid] for k,v in sap.items()} if len(sap['time'])>100:s['sap_fixed_ephemeris_vet']=vet(sap,p,t0,dur,holdout=tuple(hold),controls=0,holdout_was_unseen=True) out['signals'].append(s) if snr<9:break keep&=np.abs((t-t0+p/2)%p-p/2)>dur out['status']='ok' except Exception as ex:out.update(status='error',error=repr(ex),trace=traceback.format_exc()) out['seconds']=time.time()-start;dump(dest,out) return out def main(): ap=argparse.ArgumentParser();ap.add_argument('--limit',type=int,default=64);ap.add_argument('--workers',type=int,default=5);ap.add_argument('--inverted',action='store_true');a=ap.parse_args() for folder in ('fits','lightcurves','results','manifests'):(ROOT/folder).mkdir(exist_ok=True) if (ROOT/'pilot_observations.csv').exists() and len(pd.read_csv(ROOT/'pilot_targets.csv'))>=a.limit: d=pd.read_csv(ROOT/'pilot_targets.csv',dtype={'source_id':str}).merge(pd.read_csv(ROOT/'pilot_observations.csv')[['target_name','dataURL']],left_on='gaia_designation',right_on='target_name').head(a.limit) else:d=select(a.limit) with ProcessPoolExecutor(a.workers) as ex: for f in as_completed([ex.submit(one,row,a.inverted) for row in d.to_dict('records')]): r=f.result();print(json.dumps({k:r.get(k) for k in ('source_id','kic','status','error','seconds')}|{'signals':[{'period':s['period'],'snr':s['bls_snr'],'holdout_snr':s['holdout']['snr'],'pass':s['screen_pass']} for s in r['signals']]}),flush=True) if __name__=='__main__':main()