The Astropy project supports and fosters the development of open-source and openly-developed Python packages that provide commonly-needed functionality to the astronomical community. A key element of the Astropy project is the core package Astropy, which serves as the foundation for more specialized projects and packages. In this article, we provide an overview of the organization of the Astropy project and summarize key features in the core package as of the recent major release, version 2.0. We then describe the project infrastructure designed to facilitate and support development for a broader ecosystem of inter-operable packages. We conclude with a future outlook of planned new features and directions for the broader Astropy project.

A molecular hydrogen absorber at a lookback time of 12.4 billion years, corresponding to 10$\%$ of the age of the universe today, is analyzed to put a constraint on a varying proton--electron mass ratio, $\mu$. A high resolution spectrum of the J1443$+$2724 quasar, which was observed with the Very Large Telescope, is used to create an accurate model of 89 Lyman and Werner band transitions whose relative frequencies are sensitive to $\mu$, yielding a limit on the relative deviation from the current laboratory value of $\Delta\mu/\mu=(-9.5\pm5.4_{\textrm{stat}} \pm 5.3_{\textrm{sys}})\times 10^{-6}$.

We present a new `supercalibration' technique for measuring systematic distortions in the wavelength scales of high resolution spectrographs. By comparing spectra of `solar twin' stars or asteroids with a reference laboratory solar spectrum, distortions in the standard thorium--argon calibration can be tracked with $\sim$10 m s$^{-1}$ precision over the entire optical wavelength range on scales of both echelle orders ($\sim$50--100 \AA) and entire spectrographs arms ($\sim$1000--3000 \AA). Using archival spectra from the past 20 years we have probed the supercalibration history of the VLT--UVES and Keck--HIRES spectrographs. We find that systematic errors in their wavelength scales are ubiquitous and substantial, with long-range distortions varying between typically $\pm$200 m s$^{-1}$ per 1000 \AA. We apply a simple model of these distortions to simulated spectra that characterize the large UVES and HIRES quasar samples which previously indicated possible evidence for cosmological variations in the fine-structure constant, $\alpha$. The spurious deviations in $\alpha$ produced by the model closely match important aspects of the VLT--UVES quasar results at all redshifts and partially explain the HIRES results, though not self-consistently at all redshifts. That is, the apparent ubiquity, size and general characteristics of the distortions are capable of significantly weakening the evidence for variations in $\alpha$ from quasar absorption lines.

Large statistical samples of quasar spectra have previously indicated possible cosmological variations in the fine-structure constant, $\alpha$. A smaller sample of higher signal-to-noise ratio spectra, with dedicated calibration, would allow a detailed test of this evidence. Towards that end, we observed equatorial quasar HS 1549$+$1919 with three telescopes: the Very Large Telescope, Keck and, for the first time in such analyses, Subaru. By directly comparing these spectra to each other, and by `supercalibrating' them using asteroid and iodine-cell tests, we detected and removed long-range distortions of the quasar spectra's wavelength scales which would have caused significant systematic errors in our $\alpha$ measurements. For each telescope we measure the relative deviation in $\alpha$ from the current laboratory value, $\Delta\alpha/\alpha$, in 3 absorption systems at redshifts $z_{\mathrm{abs}}=1.143$, 1.342, and 1.802. The nine measurements of $\Delta\alpha/\alpha$ are all consistent with zero at the 2-$\sigma$ level, with 1-$\sigma$ statistical (systematic) uncertainties 5.6--24 (1.8--7.0) parts per million (ppm). They are also consistent with each other at the 1-$\sigma$ level, allowing us to form a combined value for each telescope and, finally, a single value for this line of sight: $\Delta\alpha/\alpha=-5.4 \pm 3.3_{\mathrm{stat}} \pm 1.5_{\mathrm{sys}}$ ppm, consistent with both zero and previous, large samples. We also average all Large Programme results measuring $\Delta\alpha/\alpha=-0.6 \pm 1.9_{\mathrm{stat}} \pm 0.9_{\mathrm{sys}}$ ppm. Our results demonstrate the robustness and reliability at the 3 ppm level afforded by supercalibration techniques and direct comparison of spectra from different telescopes.

Rovibronic molecular hydrogen (H$_2$) transitions at redshift $z_{\rm abs} \simeq 2.659$ towards the background quasar B0642$-$5038 are examined for a possible cosmological variation in the proton-to-electron mass ratio, $\mu$. We utilise an archival spectrum from the Very Large Telescope/Ultraviolet and Visual Echelle Spectrograph with a signal-to-noise ratio of $\sim$35 per 2.5-km$\,$s$^{-1}$ pixel at the observed H$_2$ wavelengths (335--410 nm). Some 111 H$_2$ transitions in the Lyman and Werner bands have been identified in the damped Lyman $\alpha$ system for which a kinetic gas temperature of $\sim$84 K and a molecular fraction $\log f = -2.18\pm0.08$ is determined. The H$_2$ absorption lines are included in a comprehensive fitting method, which allows us to extract a constraint on a variation of the proton-electron mass ratio, $\Delta\mu/\mu$, from all transitions at once. We obtain $\Delta\mu/\mu = (17.1 \pm 4.5_{\rm stat} \pm3.7_{\rm sys})\times10^{-6}$. However, we find evidence that this measurement has been affected by wavelength miscalibration errors recently identified in UVES. A correction based on observations of objects with solar-like spectra gives a smaller $\Delta\mu/\mu$ value and contributes to a larger systematic uncertainty: $\Delta\mu/\mu = (12.7 \pm 4.5_{\rm stat} \pm4.2_{\rm sys})\times10^{-6}$.

Absorption-line systems detected in high resolution quasar spectra can be used to compare the value of dimensionless fundamental constants such as the fine-structure constant, alpha, and the proton-to-electron mass ratio, mu = m_p/m_e, as measured in remote regions of the Universe to their value today on Earth. In recent years, some evidence has emerged of small temporal and also spatial variations in alpha on cosmological scales which may reach a fractional level of 10 ppm . We are conducting a Large Programme of observations with VLT UVES to explore these variations. We here provide a general overview of the Large Programme and report on the first results for these two constants, discussed in detail in Molaro et al. and Rahmani et al. A stringent bound for Delta(alpha)/Alpha is obtained for the absorber at_abs = 1.6919 towards HE 2217-2818. The absorption profile is complex with several very narrow features, and is modeled with 32 velocity components. The relative variation in alpha in this system is +1.3+-2.4_{stat}+-1.0_{sys} ppm if Al II lambda 1670AA and three Fe II transitions are used, and +1.1+-2.6_{stat} ppm in a lightly different analysis with only Fe II transitions used. The expectation at this sky position of the recently-reported dipolar variation of alpha is (3.2--5.4)+-1.7 ppm depending on dipole model. This constraint of Delta(alpha)/alpha at face value is not supporting this expectation but is not inconsistent with it at the 3 sigma level. For the proton-to-electron mass ratio the analysis of the H_2 absorption lines of the z_{abs}~2.4018 damped Ly alpha system towards HE 0027- 1836 provides Delta(mu)/mu = (-7.6 +- 8.1_{stat} +- 6.3_{sys}) ppm which is also consistent with a null variation. (abridged)

We present an accurate analysis of the H2 absorption lines from the zabs ~ 2.4018 damped Ly{\alpha} system towards HE 0027-1836 observed with the Very Large Telescope Ultraviolet and Visual Echelle Spectrograph (VLT/UVES) as a part of the European Southern Observatory Large Programme "The UVES large programme for testing fundamental physics" to constrain the variation of proton-to-electron mass ratio, {\mu} = mp/me. We perform cross-correlation analysis between 19 individual exposures taken over three years and the combined spectrum to check the wavelength calibration stability. We notice the presence of a possible wavelength dependent velocity drift especially in the data taken in 2012. We use available asteroids spectra taken with UVES close to our observations to confirm and quantify this effect. We consider single and two component Voigt profiles to model the observed H2 absorption profiles. We use both linear regression analysis and Voigt profile fitting where {\Delta}{\mu}/{\mu} is explicitly considered as an additional fitting parameter. The two component model is marginally favored by the statistical indicators and we get {\Delta}{\mu}/{\mu} = (-2.5 +/- 8.1(stat) +/- 6.2(sys)) ppm. When we apply the correction to the wavelength dependent velocity drift we find {\Delta}{\mu}/{\mu} = (-7.6 +/- 8.1(stat) +/- 6.3(sys)) ppm. It will be important to check the extent to which the velocity drift we notice in this study is present in UVES data used for previous {\Delta}{\mu}/{\mu} measurements.

Absorption line systems detected in quasar spectra can be used to compare the value of the fine-structure constant, {\alpha}, measured today on Earth with its value in distant galaxies. In recent years, some evidence has emerged of small temporal and also spatial variations of {\alpha} on cosmological scales which may reach a fractional level of ~ 10 ppm (parts per million). To test these claims we are conducting a Large Program with the VLT UVES . We are obtaining high-resolution (R ~ 60000 and high signal-to-noise ratio (S/N ~ 100) UVES spectra calibrated specifically for this purpose. Here we analyse the first complete quasar spectrum from this Program, that of HE 2217-2818. We apply the Many Multiplet method to measure {\alpha} in 5 absorption systems towards this quasar: zabs = 0.7866, 0.9424, 1.5558, 1.6279 and 1.6919. The most precise result is obtained for the absorber at zabs = 1.6919 where 3 Fe II transitions and Al II {\lambda}1670 have high S/N and provide a wide range of sensitivities to {\alpha}. The absorption profile is complex, with several very narrow features, and requires 32 velocity components to be fitted to the data. Our final result for the relative variation in {\alpha} in this system is Delta{\alpha}/{\alpha} = +1.3 +/- 2.4stat +/- 1.0sys ppm. This is one of the tightest current bounds on {\alpha} variation from an individual absorber. The absorbers towards quasar HE 2217-2818 reveal no evidence for variation in {\alpha} at the 3 ppm precision level (1{\sigma} confidence). If the recently reported 10 ppm dipolar variation of {\alpha} across the sky were correct, the expectation at this sky position is (3.2-5.4) +/-1.7 ppm depending on dipole model used . Our constraint of Delta{\alpha}/{\alpha}=+1.3+/-2.4stat +/-1.0sys ppm is not inconsistent with this expectation.

We attempt to measure possible miscalibration of the wavelength scale of the VLT-UVES spectrograph. We take spectra of QSO HE0515-4414 through the UVES iodine cell which contains thousands of well-calibrated iodine lines and compare these lines to the wavelength scale from the standard thorium-argon pipeline calibration. Analyzing three exposures of this z = 1.71 QSO, we find two distinct types of calibration shifts needed to correct the Th/Ar wavelength scale. First, there is an overall average velocity shift of between 100 m/s and 500 m/s depending upon the exposure. Second, within a given exposure, we find intra-order velocity distortions of 100 m/s up to more than 200 m/s. These calibration errors are similar to, but smaller than, those found earlier in the Keck HIRES spectrometer. We discuss the possible origins of these two types of miscalibration. We also explore the implications of these calibration errors on the systematic error in measurements of the relative change in alpha (current value - past value) / current value, the change in the fine-structure constant derived from accurate measurement of the relative redshifts of absorption lines in QSO absorption systems. The overall average, exposure-dependent shifts should be less relevant for fine-structure work, but the intra-order shifts have the potential to affect these results. Using either our measured calibration offsets or a Gaussian model with sigma of around 90 m/s, Monte Carlo mock experiments find errors in the relative change in alpha of between 1e-6 Nsys^(-1/2) and 3e-6 Nsys^(-1/2), where Nsys is the number of systems used and the range is due to dependence on how many metallic absorption lines in each system are compared.

We report on an attempt to accurately wavelength calibrate four nights of data taken with the Keck HIRES spectrograph on QSO PHL957, for the purpose of determining whether the fine structure constant was different in the past. Using new software and techniques, we measured the redshifts of various Ni II, Fe II, Si II, etc. lines in a damped Ly-alpha system at z=2.309. Roughly half the data was taken through the Keck iodine cell which contains thousands of well calibrated iodine lines. Using these iodine exposures to calibrate the normal Th-Ar Keck data pipeline output we found absolute wavelength offsets of 500 m/s to 1000 m/s with drifts of more than 500 m/s over a single night, and drifts of nearly 2000 m/s over several nights. These offsets correspond to an absolute redshift of uncertainty of about Delta z=10^{-5} (Delta lambda= 0.02 Ang), with daily drifts of around Delta z=5x10^{-6} (Delta lambda =0.01 Ang), and multiday drifts of nearly Delta z=2x10^{-5} (0.04 Ang). The causes of the wavelength offsets are not known, but since claimed shifts in the fine structure constant would result in velocity shifts of less than 100 m/s, this level of systematic uncertainty makes may make it difficult to use Keck HIRES data to constrain the change in the fine structure constant. Using our calibrated data, we applied both our own fitting software and standard fitting software to measure (Delta alpha)/alpha, but discovered that we could obtain results ranging from significant detection of either sign, to strong null limits, depending upon which sets of lines and which fitting method was used. We thus speculate that the discrepant results on (Delta alpha)/alpha reported in the literature may be due to random fluctuations coming from under-estimated systematic errors in wavelength calibration and fitting procedure.