aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorAaron LI <aaronly.me@outlook.com>2017-02-11 14:15:39 +0800
committerAaron LI <aaronly.me@outlook.com>2017-02-11 14:15:39 +0800
commit93681a53a89243f057c295136084e5e719d7b602 (patch)
tree73ff9c7b346864dd2cc27d81a1edfadac2a0d0e6
parent32ac48643f5760a8f6d912154405786e2cec40d4 (diff)
downloadatoolbox-93681a53a89243f057c295136084e5e719d7b602.tar.bz2
astro/query_simbad.py: Significant update to sync with 'query_ned.py'
-rwxr-xr-xastro/query_simbad.py116
1 files changed, 57 insertions, 59 deletions
diff --git a/astro/query_simbad.py b/astro/query_simbad.py
index 4e7ccd7..c0b7ba3 100755
--- a/astro/query_simbad.py
+++ b/astro/query_simbad.py
@@ -1,5 +1,7 @@
#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
+#
+# Copyright (c) 2016-2017 Weitian LI <liweitianux@live.com>
+# MIT license
#
# NOTE:
# * SimbadClass
@@ -7,7 +9,10 @@
# * All available VOTable fields:
# http://simbad.u-strasbg.fr/simbad/sim-help?Page=sim-fscript#VotableFields
#
-# ChangeLog:
+# Change logs:
+# 2017-02-11:
+# * Add argument "--brief" to not print header
+# * Sync with 'query_ned.py'
# 2016-01-14:
# * Add 'z_value'
#
@@ -22,31 +27,30 @@ Query SIMBAD with the provided name or coordinate.
http://simbad.u-strasbg.fr/simbad/
"""
-__version__ = "0.1.1"
-__date__ = "2016-01-14"
-
-
import sys
import argparse
import csv
+from collections import OrderedDict
+
from astroquery.simbad import Simbad
-from astropy import coordinates
-import astropy.units as u
+# from astropy import coordinates
+# import astropy.units as u
-## Simbad configurations
+# Simbad configurations
Simbad.ROW_LIMIT = 30
Simbad.TIMEOUT = 20
-## Add query items/fields
+# Add query items/fields:
# otype: standard name of the object type
# rv_value: Radial velocity value. Eventually translated from a redshift
# z_value: Redshift value. Eventually translated from a radial velocity
# rvz_qual: Quality code (A: best, .., E: worst)
# rvz_type: stored type of velocity: 'v'=radial velocity, 'z'=redshift
Simbad.reset_votable_fields()
-Simbad.add_votable_fields('otype', 'rv_value', 'z_value', 'rvz_qual', 'rvz_type')
+Simbad.add_votable_fields('otype', 'rv_value', 'z_value',
+ 'rvz_qual', 'rvz_type')
def query_name(name, verbose=False):
@@ -74,62 +78,56 @@ def query_name(name, verbose=False):
if verbose:
print('*** %s: not found ***' % name, file=sys.stderr)
#
- results = {
- 'main_id': main_id,
- 'otype': otype,
- 'rv': rv,
- 'z': z,
- 'rvz_qual': rvz_qual,
- 'rvz_type': rvz_type,
- }
+ results = OrderedDict([
+ ("Name", name),
+ ("SIMBAD_ID", main_id),
+ ("Type", otype),
+ ("RV", rv),
+ ("z", z),
+ ("RV/z_Quality", rvz_qual),
+ ("RV/z_Type", rvz_type),
+ ])
return results
def main():
parser = argparse.ArgumentParser(
- description="Query SIMBAD ...")
- parser.add_argument("-V", "--version", action="version",
- version="%(prog)s " + "%s (%s)" % (__version__, __date__))
- parser.add_argument("infile",
- help="file contains list of names; one per line")
- parser.add_argument("outfile",
- help="output with queryed data, empty if not found; CSV format")
+ description="Query SIMBAD database by source name")
parser.add_argument("-v", "--verbose", dest="verbose",
- action="store_true", help="show verbose information")
+ action="store_true",
+ help="show verbose information")
+ parser.add_argument("-b", "--brief", dest="brief",
+ action="store_true",
+ help="be brief and do not print header")
+ parser.add_argument("-i", "--input", dest="input", required=True,
+ help="source names to be queried (sep by comma); " +
+ "or a file contains the names (one per line)")
+ parser.add_argument("-o", "--output", dest="output", default=sys.stdout,
+ help="output CSV file with queried data")
args = parser.parse_args()
- name_list = []
- main_id_list = []
- otype_list = []
- rv_list = []
- z_list = []
- rvz_qual_list = []
- rvz_type_list = []
-
- with open(args.infile) as f:
- for name in f:
- name = str.strip(name)
- name_list.append(name)
- qr = query_name(name, verbose=args.verbose)
- main_id_list.append(qr['main_id'])
- otype_list.append(qr['otype'])
- rv_list.append(qr['rv'])
- z_list.append(qr['z'])
- rvz_qual_list.append(qr['rvz_qual'])
- rvz_type_list.append(qr['rvz_type'])
-
- with open(args.outfile, 'w') as of:
- writer = csv.writer(of)
- writer.writerow([ "Name", "SIMBAD_ID", "Type",
- "RV", "z", "RV/z_Quality", "RV/z_Type" ])
- for i in range(len(name_list)):
- writer.writerow([ name_list[i],
- main_id_list[i],
- otype_list[i],
- rv_list[i],
- z_list[i],
- rvz_qual_list[i],
- rvz_type_list[i] ])
+ try:
+ names = map(str.strip, open(args.input).readlines())
+ except FileNotFoundError:
+ names = map(str.strip, args.input.split(","))
+
+ results_list = []
+
+ for name in names:
+ qr = query_name(name, verbose=args.verbose)
+ results_list.append(qr)
+
+ try:
+ of = open(args.output, "w")
+ except TypeError:
+ of = args.output
+ writer = csv.writer(of)
+ if not args.brief:
+ writer.writerow(results_list[0].keys())
+ for res in results_list:
+ writer.writerow(res.values())
+ if of is not sys.stdout:
+ of.close()
if __name__ == "__main__":