Skip to main content
Join the official 2020 Python Developers SurveyStart the survey!

Some command lines utilities to interact with files stored in AWS S3 incl. versioned ones.

Project description

If you use AWS S3, then this can be handy tool for you.

This package offers few command line utilities, which allow a bit more, then scripts provided by boto.

Some special features:

  • List versions in defined time period, see versioning
  • Fetch versions specified in CSV file (list-file)
  • Generate temporary url links for set of keys in buckets


$ pip install

Other methods (easy_install, work too.

Quick start

We want to fetch versions of feed in bucket mybucket named my/versioned/feed.xml

  1. Be sure, you have your BOTO credentials configured. You shall have file in form:

    aws_access_key_id = <your access key>
    aws_secret_access_key = <your secret key>

    somewhere, and have set variable BOTO_CONFIG to value of complete path to this file. For more see BotoConfig.


    There are alternative methods of specifying AWS credentials, described later on.

  2. create csv file for given feed and time period:

    $ s3lsvers -from 2012-05-24T00:15 -to 2012-05-24T01:15 -list-file list.csv mybucket my/versioned/feed.xml

    You shall then find file list.csv on your disk.

  3. Review records in list.csv and delete all lines with version, which are not of your interest.

  4. Using list.csv, ask s3getvers to fetch all versions specified in the file. Be sure to run it on empty directory:

    $ s3getvers mybucket list.csv

    You will see, how is each version downloaded and saved to your current directory.

  5. Finally, you can try generating temorary url to your feed (showing latest of versions):

    $ s3tmpgen 2014-09-30T00:00:00Z mybucket my/versioned/feed.xml

    Note, that the url does not include VersionId, so it will always point to the most up todate version (in case the key happens to be on versioned bucket).

Provided commands


List versions of some feed. Could output into CSV file (-list-file) and/or html chart (-html-file).:

$ s3lsvers -h
usage: s3lsvers [-h] [-from None] [-to None] [-list-file None] [-html-file None] [-version-id None] [-profile-name None] [-aws-access-key-id None] [-aws-secret-access-key None] bucket_key

Lists all versions of given key, possibly filtering by from - to range
    for version last_modified time.
    Allows to put the listing into csv file and or into html chart.

        Listing shows:
            "file name". Can repeat if the file has more versions

            unique identifier for given version on given bucket.  Has form of
            string and not a number. identifiers are "random", do not expect
            that they are sorten alphabetically.

            size of file in bytes

            ISO 8601 formated time of file modification,
            e.g. `2011-06-22T03:05:09.000Z`

            difference between last_modified or given version
            and preceding version. It is sort of current
            update interval for that version.

        Sample use:
        Lists to the screen all versions of file keyname in the
        bucketname bucket::

            $ s3lsvers bucketname/keyname

        Lists all versions younger then given time (from given time till now)::

            $ s3lsvers -from 2011-07-19T12:00:00 bucketname/keyname

        Lists all versions older then given time
        (from very first version till given date)::

            $ s3lsvers -to 2011-07-19T12:00:00 bucketname/keyname

        Lists all versions in period betwen from and to time::

            $ s3lsvers -from 2010-01-01 -to 2011-07-19T12:00:00             bucketname/keyname

        Lists all versions and writes them into csv file named versions.csv::

            $ s3lsvers -list-file versions.csv bucketname/keyname

        Lists all versions and writes them into html chart file
        named chart.html::

            $ s3lsvers -html-file chart.html bucketname/keyname

        Prints to screen, writes to csv, creates html chart and this all
        for versions in given time period.::

            $ s3lsvers -from 2010-01-01 -to 2011-07-19T12:00:00             -list-file versions.csv -html-file chart.html bucketname/keyname

        Using bucket/key_name aliases in .s3lsvers file

        Instead of using long bucket and key names on command line, you may define aliases.

        Aliases are specified in file .s3lsvers, which may be located in currect directory, home directory or /etc/s3lsvers"

        Content of .s3lsvers files may look like this::

            #.s3lsversrc - definition of some preconfigured bucket/key values

            #values left to ":" must not contain "/" to prevent confusion with real bucket names
            plcsr: %(pl-base)s/region/pl/ConsumerServiceReady.xml
            pldfs: %(pl-base)s/region/pl/DataFusionService.xml
            czcsr: %(cz-base)s/region/cz/ConsumerServiceReady.xml
            czdfs: %(cz-base)s/region/cz/DataFusionService.xml
            skcsr: %(sk-base)s/region/sk/ConsumerServiceReady.xml
            skdfs: %(sk-base)s/region/sk/DataFusionService.xml
            skes: %(sk-base)s/region/sk/EventService.xml
            sksr: %(sk-base)s/region/sk/SummaryReports.xml

        The format follows SafeConfigParser rules:

positional arguments:
  bucket_key            bucket_name/key_name for the key to list, or key alias defined in .s3lsvers file

optional arguments:
  -h, --help            show this help message and exit
  -from None, --from-time None
                        Modification time of oldest expected version as ISO 8601 format. Can be truncated. (default: goes to the oldest version)
  -to None, --to-time None
                        Modification time of youngest expected version as ISO 8601 format. Can be truncated. (default: goes to the latest version)
  -list-file None       Name of file, where is result written in csv format. If set, the file is always overwritten.
  -html-file None       Name of file, where is result written in html format (as a chart). If set, the file is always overwritten.
  -version-id None      Optional version-id. If specified, listing does not start from the freshest version, but starts searching from given VERSION_ID and continues searching older and older versions. This could speed up listng in
                        case, you need rather older files and you know VERSION_ID which came somehow later then is the time scope you are going to list.
  -profile-name None    Name of boto profile to use for credentials
  -aws-access-key-id None
                        Your AWS Access Key ID
  -aws-secret-access-key None
                        Your AWS Secret Access Key


$ s3getvers -h
usage: s3getvers [-h] [-output-version-id-names] [-no-decompression] [-profile-name None] [-aws-access-key-id None] [-aws-secret-access-key None] bucket_name csv_version_file

Fetch file versions as listed in provided csv file

    Typical csv file (as by default produced by s3lsvers) is:


    and has columns:
    :key_name: name of the feed (not containing the bucket name itself)
    :version_id: string, identifying unique version. Any following columns can contain anything.
    :size: size in bytes. This column is not used and can be missing.
    :last_modified: date, when the version was posted. This column is not used and can be missing.

    Typical use (assuming, above csv file is available under name verlist.csv)::

        $ s3getvers yourbucketname verlist.csv

    What will create following files in current directory:

    + my/versioned/feed.xml.2012-05-23T20_45_10.xml
    + my/versioned/feed.xml.2012-05-23T20_44_31.xml
    + my/versioned/feed.xml.2012-05-23T20_43_29.xml

    Even though these files are gzipped on server, they will be decompressed on local disk.

positional arguments:
  bucket_name           bucket name (default: None)
  csv_version_file      name of CSV file with version_id

optional arguments:
  -h, --help            show this help message and exit
                        Resulting file names shall use version_id to become distinguished (default is to use timestamp of file creation)
  -no-decompression     Keeps the files as they come, do not decompress, if they come compressed
  -profile-name None    Name of boto profile to use for credentials
  -aws-access-key-id None
                        Your AWS Access Key ID
  -aws-secret-access-key None
                        Your AWS Secret Access Key


$ s3tmpgen -h
usage: s3tmpgen [-h] [-profile-name None] [-aws-access-key-id None] [-aws-secret-access-key None] [-validate-bucket] [-validate-key] expire_dt bucket_name [key_names [key_names ...]]

Generate temporary url for accessing content of AWS S3 key with defined expiration date-time.

    Urls are printed one per line to stdout.

    For missing key names empty line is printed and error goes to stderr.

    Note: if the bucket is versioned, tmp url will serve the latest version
    at the moment of request (version_id is not part of generated url).

positional arguments:
  expire_dt             ISO formatted time of expiration, full seconds, 'Z' is obligatory, e.g. '2014-02-14T21:47:16Z'
  bucket_name           name of bucket
  key_names             key names to generate tmpurl for

optional arguments:
  -h, --help            show this help message and exit
  -profile-name None    Name of boto profile to use for credentials
  -aws-access-key-id None
                        Your AWS Access Key ID
  -aws-secret-access-key None
                        Your AWS Secret Access Key
  -validate-bucket      Make sure, the bucket really exists
  -validate-key         Make sure, the key really exists

Configuring AWS S3 credentials

There are multiple methods for specifying AWS credentials

  • access key and secret key on command line
  • access key and secret key set as environmental variables AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY
  • boto config file in default location
  • boto config file in alternate location pointed to by variable BOTO_CONFIG
  • using profile name of boto config section

Details are described at BotoConfig.

The most stright-forward way of setting these variables is to create ini file in boto format as follows:

aws_access_key_id = <your_access_key_here>
aws_secret_access_key = <your_secret_key_here>

Name of the file may be arbitrary if you then ensure, environmental variable BOTO_CONFIG has value with complete path to that file.

In case, you use multiple AWS identities, add profile sections to boto ini file:

aws_access_key_id = <your default access key>
aws_secret_access_key = <your default secret key>

[profile jekyl]
aws_access_key_id = <jekyl access key for this profile>
aws_secret_access_key = <jekyl secret key for this profile>

[profile hyde]
aws_access_key_id = <hyde access key for this profile>
aws_secret_access_key = <hyde secret key for this profile>

and use switch -profile when calling the commands.


This work is built on top of boto module, great Python library for accessing AWS services created by Mitch Garnaat .

Copyright © 2011, Jan Vlcinsky

Copyright © 2012-2014, TamTam Research s.r.o.

All rights reserved.


Release date: 2014-09-08

  • corrected README.rst and NEWS.txt not to break rendering on pypi


Release date: 2014-09-08

  • s3tmpget now does not point to version_id for versioned buckets


Release date: ???

  • Corrected NEWS typos
  • MANIFEST cleaned to get rid of some installation complains


Release date: 2014-02-18

  • Added options to specify AWS credentials via boto profile_name (boto>=2.25.0) or explicit id and secret key
  • s3tmpgen - complete rewrite, simplified and finally it works
  • Removed buildout installation method.


Release date: 2012-06-15

  • command s3tmpgen - generating temporary urls for selected keys in buckets


Release date: 2012-05-28

  • command s3lsvers - to list key versions
  • command s3getvers - to fetch versions listed in csv file

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for, version
Filename, size File type Python version Upload date Hashes
Filename, size (18.3 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page