Introduction to sdss_access

SDSS Access provides a convenient way of navigating local and remote filesystem paths from the Science Archive Server (SAS). sdss_access uses the SDSS Tree product for all path look-ups.


SDSS Access works with abstract filepaths to data products, as a way of easily generating paths to specific data products. An abstract filepath is a generalized path that represents the path to the data product or file_species, and can be resolved to an example of any individual file of that “file_species”. A abstract path is defined by:

  • path_name: a reference name to the data product or file_species

  • path_template: a string filepath starting with an environment variable, and using {} templating for keyword variable replacement

These paths are defined in the tree product as path_name = path_template. See Defining New Paths for more info. For example, the abstract path for the MaNGA cube data products is

# path_name = path_template
mangacube = '$MANGA_SPECTRO_REDUX/{drpver}/{plate}/stack/manga-{plate}-{ifu}-{wave}CUBE.fits.gz'

The variable names within the {} are specified at runtime to create a path to a specific file on disk.

Path Generation

You can generate full paths to local files easily with Path.full.

# import the path
from sdss_access import Path
path = Path(release='dr17')

# generate a file system path
path.full('mangacube', drpver='v3_1_1', plate='8485', ifu='1901', wave='LOG')

Note that this only generates a path. The file may not actually exist locally. If you want to generate a URL path to the file on the SAS at Utah, you can use Path.url.

# generate a http path to the file
path.url('mangacube', drpver='v3_1_1', plate='8485', ifu='1901', wave='LOG')

You can also pass in the full path directly as a string in cases. In those cases, the first argument passed in must be an empty string.

# pass in the full path directly to path.url
full = path.full('mangacube', drpver='v3_1_1', plate='8485', ifu='1901', wave='LOG')
path.url('', full=full)

Path Names

The syntax for all paths defined in sdss_access, for most methods, is (name, **kwargs). Each path is defined by a name and several keyword arguments, indicated in the template filepath by {keyword_name}. For example, the path to a MaNGA data cube has name mangacube and path keywords, plate, drpver, ifu, and wave, defined in the path $MANGA_SPECTRO_REDUX/{drpver}/{plate}/stack/manga-{plate}-{ifu}-{wave}CUBE.fits.gz. All paths are defined inside the SDSS tree product, within a [PATHS] section in the environment configuration files, e.g. data/sdsswork.cfg or data/dr15.cfg. Within sdss_access, all paths are available as a dictionary, path.templates:

from sdss_access.path import Path
path = Path(release='dr17')

# show the dictionary of available paths

To look up what path names are available, you can use Path.lookup_names.

# look up the available path names
['BOSSLyaDR_cat', ..., 'mangacube', ..., 'xdqso_index']

To look up what keywords are needed for a given path, you can use Path.lookup_keys.

# look up the keyword arguments needed to define a MaNGA cube path
['plate', 'drpver', 'ifu', 'wave']

The full list of paths can also be found here. To create a new path, see Add Paths into the Tree and follow the instructions.

To check if a path exists locally, use the exists method. To check if the file exists remotely on the SAS, pass in the remote keyword argument

# check for local path existence
path.exists('mangacube', drpver='v3_1_1', plate='8485', ifu='1901', wave='LOG')

# check for remote path existence on the SAS
path.exists('mangacube', drpver='v3_1_1', plate='8485', ifu='1901', wave='LOG', remote=True)

Required Keywords

All the keyword variables defined in a path_template, and returned by Path.lookup_keys, are required. Not specifying all the keywords will result in an error raised.

>>> path = Path(release='dr17')

>>> # see the required keys
>>> path.lookup_keys('mangacube')
['plate', 'drpver', 'wave', 'ifu']

>>> path.full('mangacube', drpver='v3_1_1', plate='8485', ifu='1901')
KeyError: "Missing required keyword arguments: ['wave']"

Environment Paths

By default, when instantiating a new Path, it will automatically load the tree environment from any currently loaded module file, identified with any TREE_VER environment variable. Otherwise it loads the sdsswork environment, and all paths relevant to that environment.

>>> # load the default environment / paths
>>> from sdss_access.path import Path
>>> path = Path()
>>> path
<Path(release="sdsswork", public=False, n_paths=233)

To access paths from a different environment, you can change environments by passing in the release keyword argument. The release acts as an indicator for both a valid data release, e.g. “DR15”, and a valid environment configuration, e.g. “sdsswork”.

>>> # load the SDSS-V environment and paths
>>> from sdss_access.path import Path
>>> path = Path(release='sdsswork')
>>> path
<Path(release="sdsswork", public=False, n_paths=233)

>>> # switch to the environment for public data release DR17
>>> path = Path(release='DR17')
>>> path
<Path(release="dr17", public=True, n_paths=420)

When reloading a new tree environment configuration, sdss_access automatically updates the Python session os.environ with the relevant environment variables for the given release/configuration. You can preserve your original os.environ by setting the preserve_envvars keyword to True. This will preserve your original environment in its entirety.

>>> # load the SDSS-V environment but preserve your original os.environ
>>> path = Path(release='sdsswork', preserve_envvars=True)

Alternatively, you can preserve a subset of enviroment variables from your original os.environ by passing in a list of environment variables.

>>> # preserve only a single environment variable
>>> path = Path(release='sdsswork', preserve_envvars=['ROBOSTRATEGY_DATA'])

If you wish to permanently preserve your locally set environment variables, you can set the preserve_envvars parameter to true in a custom tree YAML configuration file located at ~/.config/sdss/sdss_access.yml. For example

preserve_envvars: true

Alternatively, you can permanently set a subset of environment variables to preserve by defining a list.


Extracting Keywords from Filepaths

You can extract the keyword variables from a specific filepath, by using the Path.extract method and specifying the path_name reference, and the full filepath. For the extraction to work, the path to the file must match the SAS directory structure, and have the relevant environment variable defined from the path_template.

>>> # set a path to a file
>>> filepath = '/Users/Brian/Work/sdss/sas/dr17/manga/spectro/redux/v3_1_1/8485/stack/manga-8485-1901-LOGCUBE.fits.gz'

>>> # extract the keywords
>>> path = Path(release='dr17')
>>> path.extract('mangacube', filepath)
{'drpver': 'v3_1_1', 'plate': '8485', 'ifu': '1901', 'wave': 'LOG'}

Downloading Files

You can download files from the SAS and place them in your local SAS. sdss_access expects a local SAS filesystem that mimics the real SAS at Utah. If you do not already have a SAS_BASE_DIR set, one will be defined in your home directory, as a new sas directory.

sdss_access requires valid authentication to download proprietary data. See SDSS Authentication for more information.

sdss_access has four classes designed to facilitate access to SAS data.

  • Access - class that automatically decides between RsyncAccess and CurlAccess based on the operating system.

  • HttpAccess - uses the urllib package to download data using a direct http request

  • RsyncAccess - uses rsync to download data. Available for Linux and MacOS.

  • CurlAccess - uses curl to download data. This is the only available method for use on Windows machines.

Note that all remote access classes, after instantiation, must call the Access.remote method before adding paths to ensure successful downloading of data.

Using the HttpAccess class.

from sdss_access import HttpAccess
http_access = HttpAccess(release='DR17', verbose=True)

# set to use remote

# get the file
http_access.get('mangacube', drpver='v3_1_1', plate='8485', ifu='1901', wave='LOG')

Using the RsyncAccess class. RsyncAccess is generally much faster then HttpAccess as it spreads multiple file downloads across multiple continuous rsync download streams.

# import the rsync class
from sdss_access import RsyncAccess
rsync = RsyncAccess(release='DR17')

# sets a remote mode to the real SAS

# add all the file(s) you want to download
# let's download all DR17 MaNGA cubes for plate 8485
rsync.add('mangacube', drpver='v3_1_1', plate='8485', ifu='*', wave='LOG')

# set the stream tasks

# start the download(s)

Using the CurlAccess class. CurlAccess behaves exactly the same way as RsyncAccess. After importing and instantiating a CurlAccess object, all methods and behavior are the same as in the RsyncAccess class.

# import the curl class
from sdss_access import CurlAccess
curl = CurlAccess(release='DR17')

Using the Access class. Depending on your operating system, posix or not, Access will either create itself using RsyncAccess or CurlAccess, and behave as either object. Via Acccess, Windows machines will always use CurlAccess, while Linux or Macs will automatically utilize RsyncAccess.

# import the access class
from sdss_access import Access
access = Access(release='DR17')

# the access mode is automatically set to rsync.
>>> <Access(access_mode="rsync", using="")>

# the class now behaves exactly like RsyncAccess.
# download a MaNGA cube
access.add('mangacube', drpver='v3_1_1', plate='8485', ifu='1901')

In all all cases, successful sdss_access downloads will return a code of 0. Any other number indicates that a problem occurred. If no verbose message is displayed, you may need to check the sdss_access_XX.log and sdss_access_XX.err files within the temporary directory.

Downloading with Resolved Paths

If you already have a list of resolved filepaths or urls that you wish to use sdss_access to download, you can add them using the .add_file method, instead of the .add method. The .add method takes as input a path_name and set of path template keyword arguments, while .add_file takes either a fully resolved url, filepath, or location. The input_type keyword specifies the type of path input.

from sdss_access import RsyncAccess
rsync = RsyncAccess(release='DR17')

# add a url to the stream for download
url = 'rsync://'
rsync.add_file(f, input_type='url')

# add a file to the stream for download
path = '/Users/Brian/Work/sdss/sas/dr17/manga/spectro/redux/v3_1_1/8485/stack/manga-8485-1902-LOGCUBE.fits.gz'
rsync.add_file(path, input_type='filepath')

Accessing SDSS-V Products

With SDSS-V, the usage of sdss_access remains the same. The only difference is SDSS-V products are now delivered by the “” server instead of “”. When specifying release="sdss5", you may notice the new server location, e.g.

>>> from sdss_access import Access
>>> access = Access()
>>> access
<Access(access_mode="rsync", using="")>

As with SDSS-IV, sdss_access requires valid authentication to download proprietary data for SDSS-V. See SDSS Authentication for more information. Here is an example accessing the robostrategy completeness files for SDSS-V.


The below example contains large data, ~8 GB, and may take a while to download.

from sdss_access import Access
access = Access()
access.add('rsCompleteness', observatory='apo', plan='epsilon-2-core-*')


As of version >= 3.0.0, and tree >= 4.0.0 the default config of “sdsswork” is for SDSS-V data products. In versions >2.0 - <3.0, the “sdsswork” config is for SDSS-V data products, and SDSS-V data products can be accessed using the “sdss5” config or release name.

Accessing Public Data Products

The default configuration of all sdss_access classes, i.e. Path, Access, RsyncAccess, etc. is to use the sdsswork environment configuration, for access to proprietary data or up-to-date filepaths. To specify paths, or download files, of products from public data releases, specify the release keyword. sdss_access will automatically set public=True when the input release contains DRXX. You can also explicitly set the public keyword.

# import the path and set it to use the DR17 release
from sdss_access.path import Path
path = Path(release='DR17')

# check if a public path

# generate a file system path
path.full('mangacube', drpver='v3_1_1', plate=8485, ifu=1901, wave='LOG')

# setup rsync access to download public data from DR17
rsync = RsyncAccess(public=True, release='DR17')

Accessing Paths to Data Files in SVN

sdss_access can also be used to dynamically build paths to data files contained within SVN software products, e.g. plugmap files, platelist files, or MaNGA pre-imaging or slitmap files. To learn how to define these paths, see Defining Paths to Data Files in SVN.

Once the paths are defined, you can access them as usual in sdss_access. When specifying the full local path, it uses the local path definition, and for urls, it uses the correct domain.

from sdss_access.path import Path

# load the paths for DR17
path = Path(release='DR17')
path.full('mangapreimg', designid=8405, designgrp='D0084XX', mangaid='1-42007')

path.url('mangapreimg', designid=8405, designgrp='D0084XX', mangaid='1-42007')

As always, paths generated by tree and sdss_access use the directory structure as it exists on the SDSS Science Archive Server (SAS). The same is true for paths defined for SVN data files, using the directory structure as hosted on or products installed with sdss_install. Sometimes this may conflict with locally installed and managed products. For example, the trunk version of the mangapreim SVN data product is installed locally.

module avail mangapreim

----------------------------------------------------------- /Users/Brian/Work/sdss/data/modulefiles ------------------------------------------------------------

However, the DR17 generated mangapreimg path will be the offical tag of the product for DR17, v2_9, which does not exist locally. You can always override the generated path to use your local module environment by setting the force_module keyword.

# load the paths for DR17
path = Path(release='DR17')
path.full('mangapreimg', designid=8405, designgrp='D0084XX', mangaid='1-42007')

# Override the path to use my local module
path.full('mangapreimg', designid=8405, designgrp='D0084XX', mangaid='1-42007', force_module=True)

If you want to always override paths with any local modules found, you can set the force_modules keyword on Path instantiation.

path = Path(release='DR17', force_modules=True)
path.full('mangapreimg', designid=8405, designgrp='D0084XX', mangaid='1-42007')

You can also set the force_modules parameter in your custom config file, ~/.config/sdss/sdss_access.yml to set it once permanently.

Notes for Windows Users

sdss_access downloads files into a directory defined by the SAS_BASE_DIR enviroment variable. If this path points to another drive other than the C drive, make sure that the new drive and paths have full write permissions available to curl. CurlAccess may not work properly until correct permissions are set up in your folder system.



sdss_access.path.path.Path([release, ...])

Class for construction of paths in general.

sdss_access.sync.access.Access([label, ...])

Class for providing Rsync or Curl access depending on posix

sdss_access.sync.http.HttpAccess([verbose, ...])

Class for providing HTTP access via urllib.request (python3) or urllib2 (python2) to SDSS SAS Paths

sdss_access.sync.rsync.RsyncAccess([label, ...])

Class for providing Rsync access to SDSS SAS Paths

sdss_access.sync.curl.CurlAccess([label, ...])

Class for providing Curl access to SDSS SAS Paths



Return the full local path of a given type of file.


Return the url of a given type of file.


Lookup what path names are available


Lookup the keyword arguments needed for a given path name

sdss_access.path.path.BasePath.extract(name, ...)

Extract keywords from an example path


Return the location of the relative sas path of a given type of file.

Return the name of a file of a given type.

sdss_access.path.path.BasePath.dir(filetype, ...)

Return the directory containing a file of a given type.

sdss_access.path.path.BasePath.any(filetype, ...)

Checks if the local directory contains any of the type of file


Expand a wildcard path locally


Returns random number of the given type of file, ...)

Returns random one of the given type of file


Configures remote access


Adds a filepath into the list of tasks to download


Sets the download streams


Start the download