The python package to deal with infrared LDR and Teff.
Project description
ir_ldr
The package ir_ldr
is for measuring the spectral line depth of the APOGEE and WINERED spectra, calculating the line depth ratio (LDR) and finally deriving the effective temperature (T_LDR).
The LDR-Teff relations inside this package are from Jian+19, Taniguchi+18 and Jian+20a (in prep.). Please also refer to Fukue+15.
This package relys on numpy
, pandas
, matplotlib
and scipy
; it is based on python 3.
Installation
pip install ir_ldr
Tutorial
The synthetic spectra of a dwarf star (Teff=5000 K, logg=4.5 dex and feh=0 dex; generated from MOOG) in ir_ldr/file/dwarf
for an example of T_LDR calculation.
# Load the linelist.
linelist = ir_ldr.load_linelist('yj', 'dwarf-j20a')
# Here we use all the orders of synthetic spectra.
for order in [43, 44, 45, 46, 47, 48, 52, 53, 54, 55, 56, 57]:
# Load the synthetic spectra
spec = pd.read_csv(ir_ldr.__path__[0] + '/file/example_spectra/dwarf/order{}.txt'.format(order),
sep=' +', skiprows=2, engine='python', names=['wav', 'residual'])
wav = spec['wav'].values
residual = spec['residual'].values
# Select the line pairs for a specific order
linelist_sub = linelist[linelist['order'] == order]
if len(linelist_sub) == 0:
continue
linelist_sub.reset_index(drop=True, inplace=True)
# Measure the line depth of low(1)- and high(2)-EP line.
# Here the signal to noise ratio for the target star and telluric standard are
# manually set as 300, but the S_N of synthetic spectra is much higher than that.
d1 = ir_ldr.depth_measure(wav, residual, linelist_sub['linewav1'], suffix=1, S_N=[300, 300])
d2 = ir_ldr.depth_measure(wav, residual, linelist_sub['linewav2'], suffix=2, S_N=[300, 300])
# Calculate the logLDR value.
lgLDR = ir_ldr.cal_ldr(d1, d2, type='lgLDR')
# Combine the Dataframes of one order.
record = ir_ldr.combine_df([linelist_sub, d1, d2, lgLDR])
if order == 43:
record_all = record
else:
record_all = pd.concat([record_all, record], sort=False)
# Calculate T_LDR
LDR = ir_ldr.ldr2tldr_winered_solar(record_all, df_output=True)
And the result (T_LDR, T_LDR_err)
is:
LDR[0:2]
>>> (5009.857201559249, 22.35966233607925)
# Note the T_LDR_err is not an accurate estimation here since the S_N is manually set.
Author
Mingjie Jian (ssaajianmingjie@gmail.com)
PhD student, Department of Astronomy, the University of Tokyo
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
File details
Details for the file ir_ldr_mingjie-0.1.4-py3-none-any.whl
.
File metadata
- Download URL: ir_ldr_mingjie-0.1.4-py3-none-any.whl
- Upload date:
- Size: 163.7 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/1.15.0 pkginfo/1.5.0.1 requests/2.14.2 setuptools/46.0.0 requests-toolbelt/0.8.0 tqdm/4.19.4 CPython/3.5.4
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 723e7ba3f02ef111e76ec282ee09137d23626fb243069875496aabafdd8be55b |
|
MD5 | 7d47c47afddd80791fa47ed03a04451e |
|
BLAKE2b-256 | 0013a1272092cd828fb95441b98bdae6a2f165444cee6964d667b6c7b12822a4 |