def ver03(test=False,
skip_copy=False,
publisher='J.X. Prochaska',
clobber=False,
version='v03.1',
out_path=None,
redo_dr14=False):
""" Build version 3.X
Reads several previous datasets from v1.X
Remakes the maindb using BOSS DR14 as the main driver
v3.0 will be BOSS DR14 only to speed up generation of the rest
Parameters
----------
test : bool, optional
Run test only
skip_copy : bool, optional
Skip copying the data from v01
Returns
-------
"""
import os
from specdb.specdb import IgmSpec
# Read v02
v02file = os.getenv('SPECDB') + '/IGMspec_DB_v02.1.hdf5'
#v01file_debug = igmspec.__path__[0]+'/tests/files/IGMspec_DB_v01_debug.hdf5'
print("Loading v02")
igmsp_v02 = IgmSpec(db_file=v02file)
v02hdf = igmsp_v02.hdf
#maindb = igmsp_v02.cat.copy()
# Start new file
if out_path is None:
out_path = '/scratch/IGMSpec/'
outfil = out_path + 'IGMspec_DB_{:s}.hdf5'.format(version)
# Clobber?
if not chk_clobber(outfil, clobber=clobber):
return
# Other bits
pair_groups = ['SDSS_DR7']
# Begin
hdf = h5py.File(outfil, 'w')
# Set/Check keys (and set idkey internally for other checks)
idkey = 'IGM_ID'
maindb, tkeys = sdbbu.start_maindb(idkey)
group_dict = {}
# BOSS DR14
new_groups = get_build_groups('v03')
gname = 'BOSS_DR14'
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict, skip_for_debug=True)
if not redo_dr14:
v030file = os.getenv('SPECDB') + '/IGMspec_DB_v03.0.hdf5'
igmsp_v030 = IgmSpec(db_file=v030file)
grp = hdf.create_group(gname)
# Copy spectra
#warnings.warn("GET THE DR14 spectra!")
igmsp_v030.hdf.copy(gname + '/spec', hdf[gname])
# Copy meta
igmsp_v030.hdf.copy(gname + '/meta', hdf[gname])
# Meta for maindb (a little risky as Meta needs to be aligned to the spectra but they should be)
meta = igmsp_v030['BOSS_DR14'].meta
meta.remove_column('IGM_ID')
maindb = sdbbu.add_ids(maindb,
meta,
flag_g,
tkeys,
idkey,
first=(flag_g == 1),
close_pairs=(gname in pair_groups),
debug=False)
#hdf[key+'/meta'] = meta
#for akey in v01hdf[key+'/meta'].attrs.keys():
# hdf[key+'/meta'].attrs[akey] = v01hdf[key+'/meta'].attrs[akey]
# SSA info
#new_groups[gname].add_ssa(hdf, gname)
else:
# BOSS DR14
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# IDs
maindb = sdbbu.add_ids(maindb,
meta,
flag_g,
tkeys,
idkey,
first=(flag_g == 1),
close_pairs=(gname in pair_groups),
debug=False)
# Spectra
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
# Pop me
new_groups.pop('BOSS_DR14')
# Loop on new v3 groups before copying in the others
for gname in new_groups.keys():
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname,
group_dict,
skip_for_debug=True)
# IDs
maindb = sdbbu.add_ids(maindb,
meta,
flag_g,
tkeys,
idkey,
first=(flag_g == 1),
close_pairs=(gname in pair_groups),
debug=False)
# Spectra
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
# Copy over all the old stuff
redo_groups = [] #'HD-LLS_DR1']
skip_groups = [
'BOSS_DR12'
] # 'SDSS_DR7'] #warnings.warn("NEED TO PUT BACK SDSS AND BOSS!")
skip_copy = False
if (not test) and (not skip_copy):
old1 = get_build_groups('v01')
old2 = get_build_groups('v02')
# Add v02 to v01 list
for key, item in old2.items():
old1[key] = item
# Loop on the combined
for key in old1.keys():
if key in ['catalog'] + redo_groups + skip_groups:
continue
print("Working on: {:s}".format(key))
grp = hdf.create_group(key)
# Meta
meta = Table(v02hdf[key + '/meta'].value)
meta.remove_column('IGM_ID')
# Survey flag
flag_g = sdbbu.add_to_group_dict(key,
group_dict,
skip_for_debug=True)
# IDs
maindb = sdbbu.add_ids(maindb,
meta,
flag_g,
tkeys,
idkey,
first=(flag_g == 1),
close_pairs=(key in pair_groups),
debug=False)
# Add meta to HDF5
#meta.rename_column('GRATING', 'DISPERSER')
hdf[key + '/meta'] = meta
for akey in v02hdf[key + '/meta'].attrs.keys():
hdf[key + '/meta'].attrs[akey] = v02hdf[key +
'/meta'].attrs[akey]
# SSA info
old1[key].add_ssa(hdf, key)
# Copy spectra
v02hdf.copy(key + '/spec', hdf[key])
skip_myers = False
if skip_myers:
warnings.warn("NEED TO INCLUDE MYERS!")
else:
# Copy from v02
_ = hdf.create_group('quasars')
v02hdf.copy('quasars', hdf['quasars'])
#myers.add_to_hdf(hdf)
# Setup groups
pair_groups = []
'''
# Loop over the old groups to update (as needed)
new_IDs = False
for gname in redo_groups:
print("Working to replace meta/spec for group: {:s}".format(gname))
# Meta
meta = old_groups[gname].grab_meta()
# Group flag
flag_g = group_dict[gname]
# IDs
if new_IDs:
pdb.set_trace() # NOT READY FOR THIS
#maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
# first=(flag_g==1), close_pairs=(gname in pair_groups))
else:
_, _, ids = sdbbu.set_new_ids(maindb, meta, idkey)
# Spectra
old_groups[gname].hdf5_adddata(hdf, gname, meta)
old_groups[gname].add_ssa(hdf, gname)
'''
'''
meta_only = False
# Loop over the new groups
for gname in new_groups:
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict, skip_for_debug=True)
# IDs
debug= False
#if gname == 'XQ-100':
# debug = True
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(gname in pair_groups),
debug=debug)
# Spectra
if not meta_only:
pdb.set_trace()
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
'''
# Check for duplicates -- There is 1 pair in SDSS (i.e. 2 duplicates)
if not sdbbu.chk_for_duplicates(maindb, dup_lim=2):
raise ValueError("Failed duplicates")
# Check stacking
if not sdbbu.chk_vstack(hdf):
print("Meta data will not stack using specdb.utils.clean_vstack")
print("Proceed to write at your own risk..")
pdb.set_trace()
# Finish
zpri = v02hdf['catalog'].attrs['Z_PRIORITY']
sdbbu.write_hdf(hdf,
str('igmspec'),
maindb,
zpri,
group_dict,
version,
Publisher=str(publisher))
print("Wrote {:s} DB file".format(outfil))
print("Update DB info in specdb.defs.dbase_info !!")
def ver01(test=False, clobber=False, publisher='J.X. Prochaska', **kwargs):
""" Build version 1.0
Parameters
----------
test : bool, optional
Run test only
Returns
-------
"""
pdb.set_trace() # THIS VERSION IS NOW FROZEN
raise IOError("THIS VERSION IS NOW FROZEN")
version = 'v01'
# HDF5 file
outfil = igmspec.__path__[0] + '/../DB/IGMspec_DB_{:s}.hdf5'.format(
version)
# Chk clobber
if os.path.isfile(outfil):
if clobber:
warnings.warn("Overwriting previous DB file {:s}".format(outfil))
else:
warnings.warn(
"Not overwiting previous DB file. Use clobber=True to do so")
return
# Begin
hdf = h5py.File(outfil, 'w')
''' Myers QSOs '''
myers.orig_add_to_hdf(hdf)
# Main DB Table
idkey = 'IGM_ID'
maindb, tkeys = sdbbu.start_maindb(idkey)
# Group dict
group_dict = {}
# Organize for main loop
groups = get_build_groups(version)
pair_groups = ['SDSS_DR7']
meta_only = False
# Loop over the groups
for gname in groups:
# Meta
if gname == 'SDSS_DR7':
meta = groups[gname].grab_meta(hdf)
else:
meta = groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict)
# IDs
maindb = sdbbu.add_ids(maindb,
meta,
flag_g,
tkeys,
idkey,
first=(flag_g == 1),
close_pairs=(gname in pair_groups))
# Spectra
if not meta_only:
groups[gname].hdf5_adddata(hdf, gname, meta, idkey)
# Check for duplicates -- There is 1 pair in SDSS (i.e. 2 duplicates)
if not sdbbu.chk_for_duplicates(maindb, dup_lim=2):
raise ValueError("Failed duplicates")
# Check for junk
zpri = defs.z_priority()
# Finish
sdbbu.write_hdf(hdf,
str('igmspec'),
maindb,
zpri,
group_dict,
version,
Publisher=str(publisher))
print("Wrote {:s} DB file".format(outfil))
print("Update DB info in specdb.defs.dbase_info !!")
def hdf5_writter(path, reso, lowl, highl, outputname):
'''
:param path: path to the spectra file
:param lowl float: lower limit for request redshift
:param highl float: higher limit for request redshift
:param reso: delta v in km/s to rebin the data
:param outputname: the name for the output file
:return: hdf5file
'''
outfil = str(outputname)
hdf = h5py.File(outfil, 'w')
#
gdict = {}
# creating group
for z in np.arange(lowl, highl, 0.05):
# hdf.create_group('z'+str(z)+'-'+str(z+0.05))
print(z)
spectest1 = []
spec2 = []
spectest1, spec2 = reading_data(path, reso, z, z + 0.05)
print(np.round(z, 2), len(spectest1.z))
# hdf_append = h5py.File('tmp'+str(z)+'.hdf', 'w')
# hdf.create_group('z' + str(z) + '-' + str(z + 0.05))
f = hdf.create_group('z' + str(np.round(z, 2)) + '-' +
str(np.round(z + 0.05, 2)))
npix = len(spec2.wavelength)
data = sdb_u.init_data(npix)
nspec = len(spectest1.z)
print(npix)
#creat dataset
spec_set = hdf['z' + str(np.round(z, 2)) + '-' +
str(np.round(z + 0.05, 2))].create_dataset(
'spec',
data=data,
chunks=True,
maxshape=(None, ),
compression='gzip')
spec_set.resize((nspec, ))
for ii in range(nspec):
data['flux'][0][:npix] = spec2[ii].flux # Should be flux values
data['sig'][0][:npix] = spec2[ii].sig # SHould be sigma values
# print (spec[ii].sig)
data['wave'][0][:npix] = spec2[
ii].wavelength # Should be wavelength values
# Fill
spec_set[ii] = data
# hdf.copy('z'+str(z)+'-'+str(z+0.05), hdf_append['z'+str(z)+'-'+str(z+0.05)])
# making meta data
# hdfnew = h5py.File('z2.8_specdb_test1.hdf', 'w')
# group = 'z2.8-2.85'
# _ = hdfnew.create_group(group)
group = 'z' + str(np.round(z, 2)) + '-' + str(np.round(z + 0.05, 2))
id_key = 'DESI_ID'
maindb, tkeys = spbu.start_maindb(id_key)
meta = Table()
meta['zem_GROUP'] = spectest1.z
meta['RA_GROUP'] = spectest1.ra
meta['DEC_GROUP'] = spectest1.dec
meta['EPOCH'] = 2000.
meta['sig_zem'] = 0.
meta['flag_zem'] = np.string_('DESI')
meta['STYPE'] = np.string_('QSO')
# Observation
meta['SPEC_FILE'] = np.array(spectest1.filename, dtype=float)
meta['DATE-OBS'] = spectest1.date
#
meta['GROUP_ID'] = np.arange(len(meta)).astype(int)
# Spectrograph
meta['R'] = 3000.
meta['TELESCOPE'] = np.string_('KPNO-4m')
meta['DISPERSER'] = np.string_('ALL')
meta['INSTR'] = np.string_('DESI')
meta['WV_MIN'] = 3800. # Should be the right value
meta['WV_MAX'] = 9900. # Should be the right value
meta['NPIX'] = 8000 # Should be the right value
meta['PLATE'] = np.array(np.tile(1, len(spectest1.id)), dtype=int)
meta['FIBERID'] = spectest1.id
meta['MOCK_ID'] = spectest1.id
flag_g = spbu.add_to_group_dict(group, gdict)
maindb = spbu.add_ids(maindb,
meta,
flag_g,
tkeys,
'DESI_ID',
first=(flag_g == flag_g))
hdf[group]['meta'] = meta
zpri = spb_defs.z_priority()
print(flag_g)
spbu.write_hdf(hdf,
str('DESI_v05'),
maindb,
zpri,
gdict,
str('v0.1'),
Publisher='jding')
hdf.close()
def ver02(test=False,
skip_copy=False,
publisher='J.X. Prochaska',
clobber=False,
version='v02',
out_path=None):
""" Build version 2.X
Reads previous datasets from v1.X
Parameters
----------
test : bool, optional
Run test only
skip_copy : bool, optional
Skip copying the data from v01
Returns
-------
"""
import os
from specdb.specdb import IgmSpec
# Read v01
v01file = os.getenv('SPECDB') + '/IGMspec_DB_v01.hdf5'
#v01file_debug = igmspec.__path__[0]+'/tests/files/IGMspec_DB_v01_debug.hdf5'
print("Loading v01")
igmsp_v01 = IgmSpec(db_file=v01file)
v01hdf = igmsp_v01.hdf
maindb = igmsp_v01.cat.copy()
# Start new file
if out_path is None:
out_path = igmspec.__path__[0] + '/../DB/'
outfil = out_path + 'IGMspec_DB_{:s}.hdf5'.format(version)
# Clobber?
if not chk_clobber(outfil, clobber=clobber):
return
# Begin
hdf = h5py.File(outfil, 'w')
# Copy over the old stuff
redo_groups = ['HD-LLS_DR1']
skip_groups = [
] #'BOSS_DR12', 'SDSS_DR7'] #warnings.warn("NEED TO PUT BACK SDSS AND BOSS!")
skip_copy = False
if (not test) and (not skip_copy):
old_groups = get_build_groups('v01')
for key in v01hdf.keys():
if key in ['catalog', 'quasars'] + redo_groups + skip_groups:
continue
else:
#v01hdf.copy(key, hdf) # ONE STOP SHOPPING
grp = hdf.create_group(key)
# Copy spectra
v01hdf.copy(key + '/spec', hdf[key])
# Modify v01 meta and add
if key == 'BOSS_DR12':
meta = boss.add_coflag(v01hdf)
else:
meta = Table(v01hdf[key + '/meta'].value)
meta.rename_column('GRATING', 'DISPERSER')
hdf[key + '/meta'] = meta
for akey in v01hdf[key + '/meta'].attrs.keys():
hdf[key +
'/meta'].attrs[akey] = v01hdf[key +
'/meta'].attrs[akey]
# SSA info
old_groups[key].add_ssa(hdf, key)
skip_myers = False
if skip_myers:
warnings.warn("NEED TO INCLUDE MYERS!")
else:
myers.add_to_hdf(hdf)
# Setup groups
old_groups = get_build_groups('v01')
pair_groups = []
group_dict = igmsp_v01.qcat.group_dict
# Set/Check keys (and set idkey internally for other checks)
idkey = 'IGM_ID'
_, tkeys = sdbbu.start_maindb(idkey)
mkeys = list(maindb.keys())
for key in tkeys:
assert key in mkeys
# Loop over the old groups to update (as needed)
new_IDs = False
for gname in redo_groups:
print("Working to replace meta/spec for group: {:s}".format(gname))
# Meta
meta = old_groups[gname].grab_meta()
# Group flag
flag_g = group_dict[gname]
# IDs
if new_IDs:
pdb.set_trace() # NOT READY FOR THIS
#maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
# first=(flag_g==1), close_pairs=(gname in pair_groups))
else:
_, _, ids = sdbbu.set_new_ids(maindb, meta, idkey)
# Spectra
old_groups[gname].hdf5_adddata(hdf, gname, meta)
old_groups[gname].add_ssa(hdf, gname)
meta_only = False
new_groups = get_build_groups(version)
# Loop over the new groups
for gname in new_groups:
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname,
group_dict,
skip_for_debug=True)
# IDs
debug = False
#if gname == 'XQ-100':
# debug = True
maindb = sdbbu.add_ids(maindb,
meta,
flag_g,
tkeys,
idkey,
first=(flag_g == 1),
close_pairs=(gname in pair_groups),
debug=debug)
# Spectra
if not meta_only:
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
# Check for duplicates -- There is 1 pair in SDSS (i.e. 2 duplicates)
if not sdbbu.chk_for_duplicates(maindb, dup_lim=2):
raise ValueError("Failed duplicates")
# Check stacking
if not sdbbu.chk_vstack(hdf):
print("Meta data will not stack using specdb.utils.clean_vstack")
print("Proceed to write at your own risk..")
pdb.set_trace()
# Finish
zpri = v01hdf['catalog'].attrs['Z_PRIORITY']
sdbbuwrite_hdf(hdf,
str('igmspec'),
maindb,
zpri,
group_dict,
version,
Publisher=str(publisher))
print("Wrote {:s} DB file".format(outfil))
print("Update DB info in specdb.defs.dbase_info !!")
def generate_by_refs(input_refs, outfile, version):
"""
Build a specDB file according to the input references
Args:
input_refs (list):
List of references from which to build the specDB
outfile (str):
Output filename
version (str):
Version number
"""
# Not elegant but it works
all_folders = glob.glob(db_path+'/*/*')
all_refs = [os.path.basename(ifolder) for ifolder in all_folders]
# z_tbl
allz_tbl = Table()
# Loop in input refs
all_spec_files = []
refs_list = []
for ref in input_refs:
idx = all_refs.index(ref)
# Redshift tables
z_tbl = load_z_tables(all_folders[idx])
allz_tbl = vstack([allz_tbl, z_tbl])
# Grab the list of spectra
specs = glob.glob(os.path.join(all_folders[idx], 'J*_spec.fits'))
if len(specs) == 0:
continue
# Save
all_spec_files += specs
refs_list += [ref]*len(specs)
# Get it started
# HDF5 file
hdf = h5py.File(outfile, 'w')
# Defs
zpri = defs.z_priority()
# Main DB Table
id_key = 'FRB_ID'
maindb, tkeys = spbu.start_maindb(id_key)
tkeys += ['ZQ']
gdict = {}
# Loop on Instruments
#pair_groups = ['MUSE']
pair_groups = []
badf = None
for instr in all_instruments:
print("Working on {}".format(instr))
fits_files, irefs = grab_files(all_spec_files, refs_list, instr)
if len(fits_files) == 0:
continue
# Option dicts
mwargs = {}
mwargs['toler'] = 1.0 * units.arcsec # Require an
skipz = False
swargs = {}
# Meta
parse_head, mdict, fname = None, None, True
if instr == 'SDSS':
mdict = dict(DISPERSER='BOTH', R=2000., TELESCOPE='SDSS 2.5-M', INSTR='SDSS')
parse_head = {'DATE-OBS': 'MJD'}
maxpix = 4000
scale = 1e-17
elif instr == 'FORS2':
mdict = dict(TELESCOPE='VLT', INSTR='FORS2')
parse_head = {'DATE-OBS': 'MJD', 'DISPERSER': 'DISPNAME', 'R': True}
maxpix = 2050
scale = 1e-17
elif instr == 'MUSE':
mdict = dict(TELESCOPE='VLT', R=2000.)
parse_head = {'DATE-OBS': 'MJD-OBS', 'DISPERSER': 'DISPNAME', 'INSTR': 'INSTRUME'}
maxpix = 4000
scale = 1e-20
elif instr == 'KCWI':
mdict = dict(TELESCOPE='Keck-2')
parse_head = {'DATE-OBS': 'MJD', 'DISPERSER': 'DISPNAME', 'INSTR': 'INSTRUME', 'R': True}
maxpix = 4000
scale = 1e-17
elif instr == 'MagE':
parse_head = {'R': True, 'DATE-OBS': 'MJD-OBS', 'TELESCOPE': 'TELESCOP',
'INSTR': 'INSTRUME', 'DISPERSER': 'DISPNAME'}
maxpix = 18000
scale = 1e-17
elif instr == 'GMOS-S':
mdict = dict(TELESCOPE='Gemini-S', INSTR='GMOS-S')
parse_head = {'R': True, 'DATE-OBS': 'MJD-OBS', 'DISPERSER': 'DISPNAME'}
maxpix = 3500
scale = 1e-17
elif instr == 'LRISb':
mdict = dict(TELESCOPE='Keck-1')
parse_head = {'DATE-OBS': 'MJD', 'DISPERSER': 'DISPNAME', 'INSTR': 'INSTRUME'}
maxpix = 2050 # 2x binning
elif instr == 'GMOS-N':
mdict = dict(TELESCOPE='Gemini-N', INSTR='GMOS-N')
parse_head = {'R': True, 'DATE-OBS': 'MJD-OBS',
'DISPERSER': 'DISPNAME'}
maxpix = 3500
scale = 1e-17
elif instr == 'LRISr':
mdict = dict(TELESCOPE='Keck-1')
parse_head = {'DATE-OBS': 'MJD', 'DISPERSER': 'DISPNAME', 'INSTR': 'INSTRUME'}
maxpix = 2050
scale = 1e-17
elif instr == 'DEIMOS':
mdict = dict(TELESCOPE='Keck-2')
parse_head = {'DATE-OBS': 'MJD', 'DISPERSER': 'DISPNAME',
'INSTR': 'INSTRUME'}
maxpix = 9000
scale = 1e-17
elif instr == 'Goodman':
mdict = dict(TELESCOPE='SOAR', INSTR='Goodman')
parse_head = {'DATE-OBS': 'MJD', 'DISPERSER': 'DISPNAME',
'R': True}
maxpix = 2048
scale = 1e-17
elif instr == 'XSHOOTER':
mdict = dict(TELESCOPE='VLT')
parse_head = {'DATE-OBS': 'MJD', 'DISPERSER': 'DISPNAME', 'INSTR': 'INSTRUME'}
maxpix = 33000
scale = 1e-17
else:
embed(header='172')
# Meta
full_meta = pbuild.mk_meta(fits_files, allz_tbl, mdict=mdict, fname=fname,
verbose=True, parse_head=parse_head, skip_badz=skipz,
stype='GAL',
chkz=True, **mwargs)
full_meta['Ref'] = irefs
# Survey flag
flag_g = spbu.add_to_group_dict(instr, gdict, skip_for_debug=True)
# IDs
#if 'MUSE' in instr:
# embed(header='278 of build specdb')
maindb = spbu.add_ids(maindb, full_meta, flag_g, tkeys, id_key,
first=(flag_g==1),
mtch_toler=1.*units.arcsec,
close_pairs=(instr in pair_groups))
# Ingest --
pbuild.ingest_spectra(hdf, instr, full_meta, max_npix=maxpix, verbose=False,
badf=badf, grab_conti=False, scale=scale, **swargs)
# Write
spbu.write_hdf(hdf, str('FRB'), maindb, zpri, gdict, version, Publisher=str('JXP'))
print("Wrote {:s} DB file".format(outfile))
print("You probably need to move it into SPECDB")
def mk_db(dbname, tree, outfil, iztbl, version='v00', id_key='PRIV_ID',
publisher='Unknown', **kwargs):
""" Generate the DB
Parameters
----------
dbname : str
Name for the database
tree : str
Path to top level of the tree of FITS files
Typically, each branch in the tree corresponds to a single instrument
outfil : str
Output file name for the hdf5 file
iztbl : Table or str
If Table, see meta() docs for details on its format
If str, it must be 'igmspec' and the user must have that DB downloaded
version : str, optional
Version code
Returns
-------
"""
from specdb import defs
# ztbl
if isinstance(iztbl, str):
if iztbl == 'igmspec':
from specdb.specdb import IgmSpec
igmsp = IgmSpec()
ztbl = Table(igmsp.idb.hdf['quasars'][...])
elif isinstance(iztbl, Table):
ztbl = iztbl
else:
raise IOError("Bad type for ztbl")
# Find the branches
branches = glob.glob(tree+'/*')
branches.sort()
# HDF5 file
hdf = h5py.File(outfil,'w')
# Defs
zpri = defs.z_priority()
gdict = {}
# Main DB Table
maindb, tkeys = spbu.start_maindb(id_key)
# MAIN LOOP
for ss,branch in enumerate(branches):
# Skip files
if not os.path.isdir(branch):
continue
print('Working on branch: {:s}'.format(branch))
# Files
fits_files, out_tup = grab_files(branch)
meta_file, mtbl_file, ssa_file = out_tup
# Meta
maxpix, phead, mdict, stype = 10000, None, None, 'QSO'
if meta_file is not None:
# Load
meta_dict = ltu.loadjson(meta_file)
# Maxpix
if 'maxpix' in meta_dict.keys():
maxpix = meta_dict['maxpix']
# STYPE
if 'stype' in meta_dict.keys():
stype = meta_dict['stype']
# Parse header
if 'parse_head' in meta_dict.keys():
phead = meta_dict['parse_head']
if 'meta_dict' in meta_dict.keys():
mdict = meta_dict['meta_dict']
full_meta = mk_meta(fits_files, ztbl, mtbl_file=mtbl_file,
parse_head=phead, mdict=mdict, **kwargs)
# Update group dict
group_name = branch.split('/')[-1]
flag_g = spbu.add_to_group_dict(group_name, gdict)
# IDs
maindb = add_ids(maindb, full_meta, flag_g, tkeys, 'PRIV_ID', first=(flag_g==1))
# Ingest
ingest_spectra(hdf, group_name, full_meta, max_npix=maxpix, **kwargs)
# SSA
if ssa_file is not None:
user_ssa = ltu.loadjson(ssa_file)
ssa_dict = default_fields(user_ssa['Title'], flux=user_ssa['flux'], fxcalib=user_ssa['fxcalib'])
hdf[group_name]['meta'].attrs['SSA'] = json.dumps(ltu.jsonify(ssa_dict))
# Check stacking
if not spbu.chk_vstack(hdf):
print("Meta data will not stack using specdb.utils.clean_vstack")
print("Proceed to write at your own risk..")
pdb.set_trace()
# Write
write_hdf(hdf, str(dbname), maindb, zpri, gdict, version,
Publisher=publisher)
print("Wrote {:s} DB file".format(outfil))
def ver01(test=False, clobber=False, publisher='J.X. Prochaska', **kwargs):
""" Build version 1.0
Parameters
----------
test : bool, optional
Run test only
Returns
-------
"""
pdb.set_trace() # THIS VERSION IS NOW FROZEN
raise IOError("THIS VERSION IS NOW FROZEN")
version = 'v01'
# HDF5 file
outfil = igmspec.__path__[0]+'/../DB/IGMspec_DB_{:s}.hdf5'.format(version)
# Chk clobber
if os.path.isfile(outfil):
if clobber:
warnings.warn("Overwriting previous DB file {:s}".format(outfil))
else:
warnings.warn("Not overwiting previous DB file. Use clobber=True to do so")
return
# Begin
hdf = h5py.File(outfil,'w')
''' Myers QSOs '''
myers.orig_add_to_hdf(hdf)
# Main DB Table
idkey = 'IGM_ID'
maindb, tkeys = sdbbu.start_maindb(idkey)
# Group dict
group_dict = {}
# Organize for main loop
groups = get_build_groups(version)
pair_groups = ['SDSS_DR7']
meta_only = False
# Loop over the groups
for gname in groups:
# Meta
if gname == 'SDSS_DR7':
meta = groups[gname].grab_meta(hdf)
else:
meta = groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict)
# IDs
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(gname in pair_groups))
# Spectra
if not meta_only:
groups[gname].hdf5_adddata(hdf, gname, meta, idkey)
# Check for duplicates -- There is 1 pair in SDSS (i.e. 2 duplicates)
if not sdbbu.chk_for_duplicates(maindb, dup_lim=2):
raise ValueError("Failed duplicates")
# Check for junk
zpri = defs.z_priority()
# Finish
sdbbu.write_hdf(hdf, str('igmspec'), maindb, zpri,
group_dict, version, Publisher=str(publisher))
print("Wrote {:s} DB file".format(outfil))
print("Update DB info in specdb.defs.dbase_info !!")
def ver03(test=False, skip_copy=False, publisher='J.X. Prochaska', clobber=False,
version='v03.1', out_path=None, redo_dr14=False):
""" Build version 3.X
Reads several previous datasets from v1.X
Remakes the maindb using BOSS DR14 as the main driver
v3.0 will be BOSS DR14 only to speed up generation of the rest
Parameters
----------
test : bool, optional
Run test only
skip_copy : bool, optional
Skip copying the data from v01
Returns
-------
"""
import os
from specdb.specdb import IgmSpec
# Read v02
v02file = os.getenv('SPECDB')+'/IGMspec_DB_v02.1.hdf5'
#v01file_debug = igmspec.__path__[0]+'/tests/files/IGMspec_DB_v01_debug.hdf5'
print("Loading v02")
igmsp_v02 = IgmSpec(db_file=v02file)
v02hdf = igmsp_v02.hdf
#maindb = igmsp_v02.cat.copy()
# Start new file
if out_path is None:
out_path = '/scratch/IGMSpec/'
outfil = out_path + 'IGMspec_DB_{:s}.hdf5'.format(version)
# Clobber?
if not chk_clobber(outfil, clobber=clobber):
return
# Other bits
pair_groups = ['SDSS_DR7']
# Begin
hdf = h5py.File(outfil,'w')
# Set/Check keys (and set idkey internally for other checks)
idkey = 'IGM_ID'
maindb, tkeys = sdbbu.start_maindb(idkey)
group_dict = {}
# BOSS DR14
new_groups = get_build_groups('v03')
gname = 'BOSS_DR14'
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict, skip_for_debug=True)
if not redo_dr14:
v030file = os.getenv('SPECDB')+'/IGMspec_DB_v03.0.hdf5'
igmsp_v030 = IgmSpec(db_file=v030file)
grp = hdf.create_group(gname)
# Copy spectra
#warnings.warn("GET THE DR14 spectra!")
igmsp_v030.hdf.copy(gname+'/spec', hdf[gname])
# Copy meta
igmsp_v030.hdf.copy(gname+'/meta', hdf[gname])
# Meta for maindb (a little risky as Meta needs to be aligned to the spectra but they should be)
meta = igmsp_v030['BOSS_DR14'].meta
meta.remove_column('IGM_ID')
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(gname in pair_groups),
debug=False)
#hdf[key+'/meta'] = meta
#for akey in v01hdf[key+'/meta'].attrs.keys():
# hdf[key+'/meta'].attrs[akey] = v01hdf[key+'/meta'].attrs[akey]
# SSA info
#new_groups[gname].add_ssa(hdf, gname)
else:
# BOSS DR14
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# IDs
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(gname in pair_groups),
debug=False)
# Spectra
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
# Pop me
new_groups.pop('BOSS_DR14')
# Loop on new v3 groups before copying in the others
for gname in new_groups.keys():
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict, skip_for_debug=True)
# IDs
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(gname in pair_groups),
debug=False)
# Spectra
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
# Copy over all the old stuff
redo_groups = []#'HD-LLS_DR1']
skip_groups = ['BOSS_DR12']# 'SDSS_DR7'] #warnings.warn("NEED TO PUT BACK SDSS AND BOSS!")
skip_copy = False
if (not test) and (not skip_copy):
old1 = get_build_groups('v01')
old2 = get_build_groups('v02')
# Add v02 to v01 list
for key,item in old2.items():
old1[key] = item
# Loop on the combined
for key in old1.keys():
if key in ['catalog']+redo_groups+skip_groups:
continue
print("Working on: {:s}".format(key))
grp = hdf.create_group(key)
# Meta
meta = Table(v02hdf[key+'/meta'].value)
meta.remove_column('IGM_ID')
# Survey flag
flag_g = sdbbu.add_to_group_dict(key, group_dict, skip_for_debug=True)
# IDs
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(key in pair_groups),
debug=False)
# Add meta to HDF5
#meta.rename_column('GRATING', 'DISPERSER')
hdf[key+'/meta'] = meta
for akey in v02hdf[key+'/meta'].attrs.keys():
hdf[key+'/meta'].attrs[akey] = v02hdf[key+'/meta'].attrs[akey]
# SSA info
old1[key].add_ssa(hdf, key)
# Copy spectra
v02hdf.copy(key+'/spec', hdf[key])
skip_myers = False
if skip_myers:
warnings.warn("NEED TO INCLUDE MYERS!")
else:
# Copy from v02
_ = hdf.create_group('quasars')
v02hdf.copy('quasars', hdf['quasars'])
#myers.add_to_hdf(hdf)
# Setup groups
pair_groups = []
'''
# Loop over the old groups to update (as needed)
new_IDs = False
for gname in redo_groups:
print("Working to replace meta/spec for group: {:s}".format(gname))
# Meta
meta = old_groups[gname].grab_meta()
# Group flag
flag_g = group_dict[gname]
# IDs
if new_IDs:
pdb.set_trace() # NOT READY FOR THIS
#maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
# first=(flag_g==1), close_pairs=(gname in pair_groups))
else:
_, _, ids = sdbbu.set_new_ids(maindb, meta, idkey)
# Spectra
old_groups[gname].hdf5_adddata(hdf, gname, meta)
old_groups[gname].add_ssa(hdf, gname)
'''
'''
meta_only = False
# Loop over the new groups
for gname in new_groups:
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict, skip_for_debug=True)
# IDs
debug= False
#if gname == 'XQ-100':
# debug = True
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(gname in pair_groups),
debug=debug)
# Spectra
if not meta_only:
pdb.set_trace()
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
'''
# Check for duplicates -- There is 1 pair in SDSS (i.e. 2 duplicates)
if not sdbbu.chk_for_duplicates(maindb, dup_lim=2):
raise ValueError("Failed duplicates")
# Check stacking
if not sdbbu.chk_vstack(hdf):
print("Meta data will not stack using specdb.utils.clean_vstack")
print("Proceed to write at your own risk..")
pdb.set_trace()
# Finish
zpri = v02hdf['catalog'].attrs['Z_PRIORITY']
sdbbu.write_hdf(hdf, str('igmspec'), maindb, zpri,
group_dict, version, Publisher=str(publisher))
print("Wrote {:s} DB file".format(outfil))
print("Update DB info in specdb.defs.dbase_info !!")
def ver02(test=False, skip_copy=False, publisher='J.X. Prochaska', clobber=False,
version='v02', out_path=None):
""" Build version 2.X
Reads previous datasets from v1.X
Parameters
----------
test : bool, optional
Run test only
skip_copy : bool, optional
Skip copying the data from v01
Returns
-------
"""
import os
from specdb.specdb import IgmSpec
# Read v01
v01file = os.getenv('SPECDB')+'/IGMspec_DB_v01.hdf5'
#v01file_debug = igmspec.__path__[0]+'/tests/files/IGMspec_DB_v01_debug.hdf5'
print("Loading v01")
igmsp_v01 = IgmSpec(db_file=v01file)
v01hdf = igmsp_v01.hdf
maindb = igmsp_v01.cat.copy()
# Start new file
if out_path is None:
out_path = igmspec.__path__[0]+'/../DB/'
outfil = out_path + 'IGMspec_DB_{:s}.hdf5'.format(version)
# Clobber?
if not chk_clobber(outfil, clobber=clobber):
return
# Begin
hdf = h5py.File(outfil,'w')
# Copy over the old stuff
redo_groups = ['HD-LLS_DR1']
skip_groups = []#'BOSS_DR12', 'SDSS_DR7'] #warnings.warn("NEED TO PUT BACK SDSS AND BOSS!")
skip_copy = False
if (not test) and (not skip_copy):
old_groups = get_build_groups('v01')
for key in v01hdf.keys():
if key in ['catalog','quasars']+redo_groups+skip_groups:
continue
else:
#v01hdf.copy(key, hdf) # ONE STOP SHOPPING
grp = hdf.create_group(key)
# Copy spectra
v01hdf.copy(key+'/spec', hdf[key])
# Modify v01 meta and add
if key == 'BOSS_DR12':
meta = boss.add_coflag(v01hdf)
else:
meta = Table(v01hdf[key+'/meta'].value)
meta.rename_column('GRATING', 'DISPERSER')
hdf[key+'/meta'] = meta
for akey in v01hdf[key+'/meta'].attrs.keys():
hdf[key+'/meta'].attrs[akey] = v01hdf[key+'/meta'].attrs[akey]
# SSA info
old_groups[key].add_ssa(hdf, key)
skip_myers = False
if skip_myers:
warnings.warn("NEED TO INCLUDE MYERS!")
else:
myers.add_to_hdf(hdf)
# Setup groups
old_groups = get_build_groups('v01')
pair_groups = []
group_dict = igmsp_v01.qcat.group_dict
# Set/Check keys (and set idkey internally for other checks)
idkey = 'IGM_ID'
_, tkeys = sdbbu.start_maindb(idkey)
mkeys = list(maindb.keys())
for key in tkeys:
assert key in mkeys
# Loop over the old groups to update (as needed)
new_IDs = False
for gname in redo_groups:
print("Working to replace meta/spec for group: {:s}".format(gname))
# Meta
meta = old_groups[gname].grab_meta()
# Group flag
flag_g = group_dict[gname]
# IDs
if new_IDs:
pdb.set_trace() # NOT READY FOR THIS
#maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
# first=(flag_g==1), close_pairs=(gname in pair_groups))
else:
_, _, ids = sdbbu.set_new_ids(maindb, meta, idkey)
# Spectra
old_groups[gname].hdf5_adddata(hdf, gname, meta)
old_groups[gname].add_ssa(hdf, gname)
meta_only = False
new_groups = get_build_groups(version)
# Loop over the new groups
for gname in new_groups:
print("Working on group: {:s}".format(gname))
# Meta
meta = new_groups[gname].grab_meta()
# Survey flag
flag_g = sdbbu.add_to_group_dict(gname, group_dict, skip_for_debug=True)
# IDs
debug= False
#if gname == 'XQ-100':
# debug = True
maindb = sdbbu.add_ids(maindb, meta, flag_g, tkeys, idkey,
first=(flag_g==1), close_pairs=(gname in pair_groups),
debug=debug)
# Spectra
if not meta_only:
new_groups[gname].hdf5_adddata(hdf, gname, meta)
new_groups[gname].add_ssa(hdf, gname)
# Check for duplicates -- There is 1 pair in SDSS (i.e. 2 duplicates)
if not sdbbu.chk_for_duplicates(maindb, dup_lim=2):
raise ValueError("Failed duplicates")
# Check stacking
if not sdbbu.chk_vstack(hdf):
print("Meta data will not stack using specdb.utils.clean_vstack")
print("Proceed to write at your own risk..")
pdb.set_trace()
# Finish
zpri = v01hdf['catalog'].attrs['Z_PRIORITY']
sdbbuwrite_hdf(hdf, str('igmspec'), maindb, zpri,
group_dict, version, Publisher=str(publisher))
print("Wrote {:s} DB file".format(outfil))
print("Update DB info in specdb.defs.dbase_info !!")