def get_new_ids(maindb,
newdb,
idkey,
chk=True,
mtch_toler=None,
pair_sep=0.5 * u.arcsec,
close_pairs=False,
debug=False):
""" Generate new CAT_IDs for an input DB
Parameters
----------
maindb : Table
newdb : Table
RA, DEC assumed to be given by RA_GROUP and DEC_GROUP
chk : bool, optional
Perform some checks
idkey : str
Key for ID
mtch_toler : Quantity, optional
Matching tolerance; typically taken from the default
pair_sep : Angle, optional
Sepration at which a pair is considered 'real'
close_pairs : bool, optional
Input list includes close pairs (i.e. within mtch_toler)
Returns
-------
ids : ndarray (int)
Old IDs are filled with negative their value
New IDs are generated as needed
"""
if mtch_toler is None:
cdict = defs.get_cat_dict()
mtch_toler = cdict['match_toler']
IDs = np.zeros(len(newdb), dtype=int)
# Setup
c_main = SkyCoord(ra=maindb['RA'], dec=maindb['DEC'], unit='deg')
c_new = SkyCoord(ra=newdb['RA_GROUP'], dec=newdb['DEC_GROUP'], unit='deg')
# Check for pairs in the new list
pidx1, pidx2, pd2d, _ = c_new.search_around_sky(c_new, mtch_toler)
pairs = pd2d > pair_sep
if np.sum(pairs) and (not close_pairs):
print(
"Input catalog includes pairs closer than {:g} and wider than {:g}"
.format(mtch_toler, pair_sep))
raise IOError("Use close_pairs=True if appropriate")
# Find new sources (ignoring pairs at first)
idx, d2d, d3d = match_coordinates_sky(c_new, c_main, nthneighbor=1)
new = d2d > mtch_toler
# Old IDs
IDs[~new] = -1 * maindb[idkey][idx[~new]]
# Now deal with pairs
if np.sum(pairs) > 0:
# Check against catalog
pidx, pd2d, _ = match_coordinates_sky(c_new[pidx1][pairs],
c_main,
nthneighbor=1)
not_pair_match = pd2d > pair_sep
# Reset new -- It will get a new ID below -- np.where is needed to actually set new
new[pidx1[pairs][np.where(not_pair_match)[0]]] = True
# New IDs
nnew = np.sum(new)
new_idx = np.where(new)[0]
newID = np.max(maindb[idkey])
# Ingest
if nnew == 1:
IDs[new_idx] = newID + 1
elif nnew > 1: # Deal with duplicates
sub_c_new = c_new[new]
dup_idx, dup_d2d, _ = match_coordinates_sky(sub_c_new,
sub_c_new,
nthneighbor=2)
if close_pairs:
dups = dup_d2d < pair_sep
else:
dups = dup_d2d < mtch_toler
ndups = np.sum(dups)
# Not duplicates
IDs[new_idx[~dups]] = newID + 1 + np.arange(np.sum(~dups))
newID = max(np.max(IDs), newID)
# Duplicates
if ndups > 0:
warnings.warn(
"We found {:d} duplicates (e.g. multiple spectra). Hope this was expected"
.format(ndups // 2))
# Cut down to unique and restrict to new ones (there are at least 2 duplicates per match)
dup_idx = np.where(dups)[0]
dup_filled = np.array([False] * len(sub_c_new))
if debug:
pdb.set_trace()
for idup in dup_idx: # Ugly loop..
if dup_filled[idup]: # Already filled as a duplicate
continue
dcoord = sub_c_new[idup]
sep = dcoord.separation(sub_c_new)
if close_pairs:
isep = np.where(sep < pair_sep)[0]
else:
isep = np.where(sep < mtch_toler)[0]
# ID
newID += 1
IDs[new_idx[isep]] = newID
dup_filled[isep] = True # Avoids the other dup(s)
if chk:
print("The following sources were previously in the DB")
print(newdb[~new])
'''
if close_pairs: # TEST SDSS
tc = SkyCoord(ra=210.053222, dec=31.581701, unit='deg')#, (210.053552, 31.58131)]>
isep = np.argmin(tc.separation(c_new))
pdb.set_trace()
IDs[isep]
'''
# Return
return IDs
def mk_meta(files, ztbl, fname=False, stype='QSO', skip_badz=False,
mdict=None, parse_head=None, debug=False, chkz=False,
mtbl_file=None,
verbose=False, specdb=None, sdb_key=None, **kwargs):
""" Generate a meta Table from an input list of files
Parameters
----------
files : list
List of FITS files
ztbl : Table
Table of redshifts. Must include RA, DEC, ZEM, ZEM_SOURCE
Used for RA/DEC if fname=False; then requires SPEC_FILE too
fname : bool, optional
Attempt to parse RA/DEC from the file name
Format must be
SDSSJ######(.##)+/-######(.#)[x]
where x cannot be a #. or +/-
stype : str, optional
Description of object type (e.g. 'QSO', 'Galaxy', 'SN')
specdb : SpecDB, optional
Database object to grab ID values from
Requires sdb_key
sdb_key : str, optional
ID key in SpecDB object
skip_badz : bool, optional
Skip spectra without a parseable redshift (using the Myers catalog)
parse_head : dict, optional
Parse header for meta info with this dict
mdict : dict, optional
Input meta data in dict form e.g. mdict=dict(INSTR='ESI')
chkz : bool, optional
If any sources have no parseable redshift, hit a set_trace
mtbl_file : str
Filename of input meta table. Current allowed extensions are _meta.ascii or _meta.fits
and they must be readable by Table.read(). The values in this table will overwrite
any others generated. Table must include a SPEC_FILE column to link meta data
Returns
-------
meta : Table
Meta table
"""
if specdb is not None:
if sdb_key is None:
raise IOError("Must specify sdb_key if you are passing in specdb")
#
coordlist = []
snames = []
for ifile in files:
sname = ifile.split('/')[-1]
snames.append(sname)
if fname:
# Starting index
if 'SDSSJ' in ifile:
i0 = ifile.find('SDSSJ')+4
else:
i0 = ifile.rfind('J')+1
# Find end (ugly)
for ii in range(i0+1,99999):
if ifile[ii] in ('0','1','2','3','4','5','6','7','8','9',
'.','+','-'):
continue
else:
i1 = ii
break
# Deal with .fits
if ifile[i1-1] == '.':
i1 -= 1
# Get coord
try:
coord = ltu.radec_to_coord(ifile[i0:i1])
except (UnboundLocalError, ValueError):
pdb.set_trace()
else:
mt = np.where(ztbl['SPEC_FILE'] == sname)[0]
if len(mt) != 1:
raise IndexError("NO MATCH FOR {:s}".format(sname))
coord = ltu.radec_to_coord((ztbl['RA'][mt],
ztbl['DEC'][mt]))[0]
coordlist.append(coord)
ras = np.array([coord.ra.degree for coord in coordlist])
decs= np.array([coord.dec.degree for coord in coordlist])
coords = SkyCoord(ra=ras, dec=decs, unit='deg')
# Generate maindb Table
#maindb, tkeys = spbu.start_maindb(private=True)
# Fill
meta = Table()
meta['RA_GROUP'] = coords.ra.deg
meta['DEC_GROUP'] = coords.dec.deg
meta['STYPE'] = [str(stype)]*len(meta)
zem, zsource, ZQ = spzu.zem_from_radec(meta['RA_GROUP'], meta['DEC_GROUP'],
ztbl, **kwargs)
badz = zem <= 0.
if np.sum(badz) > 0:
if skip_badz:
warnings.warn("Skipping {:d} entries without a parseable redshift".format(
np.sum(badz)))
badz[:] = False # Risky!
else:
if chkz: # Turn this on to hit a stop instead of an Exception
pdb.set_trace()
else:
raise ValueError("{:d} entries without a parseable redshift".format(
np.sum(badz)))
meta['zem_GROUP'] = zem
meta['sig_zem'] = 0. # Need to add
meta['flag_zem'] = zsource
if ZQ is not None:
meta['ZQ'] = ZQ
# Cut?
meta = meta[~badz]
# specdb IDs
if sdb_key is not None:
meta[sdb_key] = [-9999]*len(meta)
if sdb_key not in meta.keys():
meta[sdb_key] = [-9999]*len(meta)
c_igmsp = SkyCoord(ra=specdb.qcat.cat['RA'], dec=specdb.qcat.cat['DEC'], unit='deg')
c_new = SkyCoord(ra=meta['RA_GROUP'], dec=meta['DEC_GROUP'], unit='deg')
# Find new sources
idx, d2d, d3d = match_coordinates_sky(c_new, c_igmsp, nthneighbor=1)
cdict = defs.get_cat_dict()
mtch = d2d < cdict['match_toler']
meta[sdb_key][mtch] = specdb.qcat.cat[sdb_key][idx[mtch]]
# Stack (primarily as a test)
'''
try:
maindb = vstack([maindb,meta], join_type='exact')
except:
pdb.set_trace()
'''
# SPEC_FILE
meta['SPEC_FILE'] = np.array(files)[~badz]
root_names = np.array(snames)[~badz]
# Try Header?
if parse_head is not None:
# Setup to store
plist = {}
for key in parse_head.keys():
plist[key] = []
# Loop on files
for count, sfile in enumerate(meta['SPEC_FILE']):
if verbose:
print('Parsing {:s}'.format(sfile))
try:
head = fits.open(sfile)[0].header
except FileNotFoundError: # Try for compressed
head = fits.open(sfile+'.gz')[0].header
# Call it
_ = spbu.parse_header(parse_head, head, mdict=mdict,
plist=plist, count=count)
'''
for key,item in parse_head.items():
# R
if key == 'R':
if parse_head[key] is True:
try:
plist[key].append(spbu.set_resolution(head))
except ValueError:
if mdict is not None:
try:
plist[key].append(mdict['R'])
except KeyError:
print("Key error!")
pdb.set_trace()
else:
print('Bad inputs to set_resolution()')
pdb.set_trace()
plist[key].append(0.)
else:
raise ValueError("Set something else for R")
elif key == 'DATE-OBS':
if 'MJD' in item:
tval = Time(head[item], format='mjd')
tval.format = 'iso'
else:
tval = Time(head[item].replace('/','-'), format='iso')
tval.out_submfit = 'date'
plist[key].append(tval.iso)
else:
plist[key].append(head[item])
# INSTRUMENT SPECIFIC
try:
instr = head['INSTRUME']
except KeyError:
instr = 'none'
# LRIS
if 'LRIS' in instr:
# Init
for skey in ['DISPERSER', 'INSTR', 'R']:
if skey not in plist.keys():
plist[skey] = []
# Figure out detector
try:
det = head['DETECTOR']
except KeyError:
if 'BLUE' in instr:
det = 'LRIS-B'
else:
if head['OUTFILE'] == 'lred':
det = 'LRIS-R'
else:
det = 'LRIS-B'
# Add
if 'LRIS-R' in det:
if len(plist['DISPERSER']) == count:
plist['DISPERSER'].append(head['GRANAME'])
if len(plist['INSTR']) == count:
plist['INSTR'].append('LRISr')
else:
plist['INSTR'][-1] = 'LRISr'
else:
if len(plist['DISPERSER']) == count:
plist['DISPERSER'].append(head['GRISNAME'])
if len(plist['INSTR']) == count:
plist['INSTR'].append('LRISb')
else:
plist['INSTR'][-1] = 'LRISb'
# Resolution
res = Rdicts[plist['INSTR'][-1]][plist['DISPERSER'][-1]]
try:
sname = head['SLITNAME']
except KeyError:
swidth = 1.
else:
swidth = defs.slit_width(sname, LRIS=True)
plist['R'].append(res/swidth)
'''
# Finish
for key in plist.keys():
try:
meta[key] = plist[key]
except:
embed(header='338 of private')
# mdict
if mdict is not None:
for key,item in mdict.items():
meta[key] = [item]*len(meta)
# EPOCH
if 'EPOCH' not in meta.keys():
warnings.warn("EPOCH not defined. Filling with 2000.")
meta['EPOCH'] = 2000.
# GROUP ID
meta['GROUP_ID'] = np.arange(len(meta)).astype(int)
# Fill in empty columns with warning
mkeys = meta.keys()
req_clms = defs.get_req_clms(sdb_key=sdb_key)
for clm in req_clms:
if clm not in mkeys:
if clm not in ['NPIX','WV_MIN','WV_MAX']: # File in ingest_spec
warnings.warn("Meta Column {:s} not defined. Filling with DUMMY".format(clm))
if clm == 'DATE-OBS':
meta[clm] = ['9999-1-1']*len(meta)
else:
meta[clm] = ['DUMMY']*len(meta)
# Input meta table
if mtbl_file is not None:
# Read
if '_meta.ascii' in mtbl_file:
imtbl = Table.read(mtbl_file, format='ascii')
elif '_meta.fits' in mtbl_file:
imtbl = Table.read(mtbl_file)
else:
raise IOError("Input meta table must have either an ascii or fits extension")
# Check length
if len(imtbl) != len(meta):
raise IOError("Input meta table must have same length as self-generated one")
# Check for SPEC_FILE
if 'SPEC_FILE' not in imtbl.keys():
raise ValueError("Input meta table must include SPEC_FILE column")
# Loop to get indices
idx = []
for row in imtbl:
imt = np.where(root_names == row['SPEC_FILE'])[0]
if len(imt) == 0:
print("No match to spec file {:s}. Will ignore".format(row['SPEC_FILE']))
elif len(imt) == 1:
idx.append(imt[0])
else:
raise ValueError("Two entries with the same SPEC_FILE. Something went wrong..")
idx = np.array(idx)
# Loop on keys
for key in imtbl.keys():
# Skip?
if key in ['SPEC_FILE']:
continue
if key in meta.keys():
pdb.set_trace()
else:
# Add Column
meta.add_column(imtbl[key][idx])
# Return
if debug:
meta[['RA_GROUP', 'DEC_GROUP', 'SPEC_FILE']].pprint(max_width=120)
pdb.set_trace()
return meta