What’s new in Marvin

2.3.0 (unreleased)

  • A new Marvin Image Tool for interfacing with the MaNGA optical images.

2.2.6 (August 2018)

Attention

This is a critical bugfix release that corrects a problem that could affect your science results. Please update as soon as possible and check whether your analysis is impacted by this bug.

This version fixes a critical bug when retrieving the spaxels associated with a bin. It also simplifies the library namespace allowing for easier access to the most used Tools.

Spaxels associated with a bin

In version 2.2 we introduced the concept of Bin as a collection of spaxels that belong to the same binning unit. As part of the API, one can use spaxels attribute to access a list of the spaxels that are included in the bin. The bug now fixed caused a list of incorrect spaxels to be associated with the bin, due to an inversion in the (x, y) order of the spaxels. For example, before 2.2.6 one would get

>>> cube = Cube('8485-1901')
>>> maps = cube.getMaps('HYB10')
>>> bb = maps[22, 14]
>>> bb.spaxels
[<Marvin Spaxel (x=21, y=13, loaded=False),
 <Marvin Spaxel (x=21, y=14, loaded=False),
 <Marvin Spaxel (x=22, y=13, loaded=False),
 <Marvin Spaxel (x=22, y=14, loaded=False)]

where the x and y values should be

[<Marvin Spaxel (x=13, y=21, loaded=False),
 <Marvin Spaxel (x=14, y=21, loaded=False),
 <Marvin Spaxel (x=13, y=22, loaded=False),
 <Marvin Spaxel (x=14, y=22, loaded=False)]

Simplifying the namespace

Prior to 2.2.6 accessing different Tools classes was inconvenient since one would need to import them independently (e.g., from marvin.tools.cube import Cube, from marvin.tools.maps import Maps, etc.) This version makes access easier by exposing all the Tools from the marvin.tools namespace so that you can now do

import marvin
cube = marvin.tools.Cube('8485-1901')
maps = marvin.tools.Maps('7443-12701')

Stellar Sigma Correction

For MPL-6, we now raise an explicit error when attempting to apply the correction to stellar_sigma, using the inst_sigma_correction method. The error message now suggests to upgrade to MPL-7 data. For the web display of the stellar_sigma and emline_gsigma maps, we now apply the sigma correction automatically. The corrected map is indicated via Corrected: stellar_sigma map title.


2.2 (January 2018)

Marvin 2.2.0 brings significant improvements in the way you interact with MaNGA data. Try the Jupyter Notebook for a small sample.

  • MPL-6 compatible
  • New DRP, DAP, and Query Datamodels
  • Cubes, Maps, and ModelCubes now use Astropy Quantities, i.e. encapsulating a measurement with its associated parameters (e.g., unit, mask, or inverse variance)
  • Improved Bin class
  • Fuzzy Searching and Tab Completion
  • New access to DAPall data on Maps and ModelCubes
  • Scatter and Histogram Plotting
  • Improved Query Results Handling and Integrated Plotting
  • New MaskBit class

2.1.4 (August 2017)

  • Refactored the Query Page in Marvin Web: Adds more intuitive parameters naming in dropdown. Adds Guided Marvin Query Builder, using Jquery Query Builder. See the Search page section of Web Docs.
  • Adds Galaxy Postage Stamp view of the result set from a Marvin Query in the Web
  • Adds Rate Limiting for the Marvin API. Adopts a limit of 200 requests/min on all routes and 60/min for queries.
  • Adds new query_params object in Marvin Tools for improved navigation and selection of available query parameters. See updated documentation for Queries and Query Params
  • Adds ability for creating custom maps (using custom values and masks) with Marvin Plotting framework. See updated Plotting Tutorial
  • New Sidebar in Marvin Documentation for easier navigation.
  • New Marvin Getting Started Page.
  • New Marvin Exercises for showcasing utilization of Marvin in science workflows
  • Numerous bug fixes. See Changelog for full account of all Github Issues closed.

2.1.3 (May 2017)

  • Slicing in tool objects now behaves as in a Numpy array. That means that cube[i, j] returns the same result as cube.getSpaxel(x=j, y=i, xyorig='lower').
  • Now it is possible to query on absolute magnitude colours from NSA’s elpetro_absmag. Absolute magnitudes are now the default for plotting on the web.
  • The data file for the default colormap for Map.plot() (“linear_Lab”) is now included in pip version of Marvin and does not throw invalid FileNotFoundError if the data file is missing.
  • Query shortcuts are now only applied on full words, to avoid blind replacements. This fixes a bug that made parameters such as elpetro_absmag_r being replaced by elpetro_absmaelpetro_mag_g_r.
  • Refactored Map plotting methods into Utilities.
    • Map plotting now accepts user-defined value, ivar, and mask arrays (e.g., BPT masks).
    • It is possible to create multi-panel map plots.
    • All plotting code no longer overwrites matplotlib rcParams.
    • Map plotting has new default gray/hatching scheme for data quality (in tools and web):
      • gray: spaxels with NOCOV.
      • hatched: spaxels with bad data (UNRELIABLE and DONOTUSE) or S/N below some minimum value.
      • colored: good data.
    • Map plotting no longer masks spaxels near zero velocity contour because by default (in tools and web), there is no minimum signal-to-noise ratio for velocity plots.
  • New tutorials: tutorials/plotting and Lean Tutorial.

2.1 (February 2017)

  • Marvin is now minimally compliant with Python 3.5+
  • https://sas.sdss.org/marvin now points to Marvin 2 (instead of Marvin 1).
  • The NSA catalog information is now available via Cube.nsa in Marvin Cubes.
  • Marvin Web (marvin.web) now has a new NSA Display tab with interactive scatter, and box-and-whisker plots.
  • Marvin Web (marvin.web) has more python tips for working with Marvin Tools (marvin.tools) objects.
  • Marvin now uses Sentry to catch and send errors.
  • Marvin Maps now include the ability to make and plot a BPT Diagrams diagram.
  • Marvin Maps have updated plotting display and now include a new signal-to-noise (snr) attribute on each map.
  • Check out the Visual Guide to Marvin Tools.
  • Marvin Spaxels now include ra and dec as properties.
  • Streamlined list of query parameters both in the Web (marvin.web) and Tools (marvin.tools). Added new parameter ha_to_hb ratio.
  • Marvin has updated the Image Utilities functions for downloading, showing, and locating.
  • New check_marvin utility to provide some basic system checks with regards to Marvin
  • Marvin Web (marvin.web) now has a “Provide Feedback” button in the navbar that directly links to a New Issue in Github.
  • See Changelog for more.

2.0 Beta (November 2016)

  • Brand new painless installation (pip install sdss-marvin)
  • New Marvin Tools (Maps, Bin, ModelCube)
  • Pickling of Marvin Tools, Queries, and Results (i.e. local save and restore)
  • DAP Spaxel (“Zonal”) Queries
  • Dynamic DAP Map display in the web, with point-and-click spaxel
  • For MPL-5+, display of model fits in spectrum view in the web
  • Versions simplified from mpl, drp, dap down to release
  • API API Authentication

2.0 Alpha (June 2016)

Marvin 2.0 is a complete overhaul of Marvin 1.0, converting Marvin into a full suite of interaction tools.

Marvin 2.0 introduces two new modes of operations, Tools (marvin.tools) and API (marvin.api), to the Marvin environment, and introduces an extensive redesign of the Marvin web app.

The major improvements and additions in this release:

  • Data Access Modes: a new mode based navigation system that allows you to seamlessly interact with MaNGA data no matter where it is.
  • Tools (marvin.tools): a python package for accessing and interacting with MaNGA data, whether the files are in your computer or they need to be retrieved remotely via the API.
  • API (marvin.api): remotely grab the data you are looking for as JSONs to integrate directly into your local scripts
  • Query: a tool to harness the full statistical power of the MaNGA data set by querying the Marvin Databases.
  • A completely overhauled Web (marvin.web) interface, including:
    • A more powerful Search with an intuitive pseudo-natural language search capability.
    • A simple and clean Plate and Galaxy detail page.
    • Interactive spectrum selection from the galaxy image.
    • An image roulette if you are feeling lucky.