Skip to main content

dce client for epipearl

Project description

===============================
epiclient
===============================

DCE client for `epipearl
<https://github.com/harvard-dce/epipearl>`_

epiclient wraps `epipearl <https://github.com/harvard-dce/epipearl>`_
HTTP API and web ui calls. epiclient creates channels and recorders and
configures them, as well as general configurations for an epiphan-pearl
device. There is also configurations for the epiphan-pearl to work as an
`Opencast Matterhorn <http://opencast.org/matterhorn>`_ Capture Agent,
and these require a custom firmware from DCE that includes `mhpearl
<https://bitbucket.org/hudcede/mhpearl>`_

This software should be considered alpha, therefore likely to change/break in
the near future.



*******************************************************
Install
*******************************************************

.. code-block:: bash

pip install epiclient


*******************************************************
Overview
*******************************************************

To configure an epiphan-pearl device to work as a Capture Agent, there are
general configs (like ntp server, touchscreen, permanent logs), and channels
and recorders that need to be created and proper configured.

epiclient has a CLI that goes through the configuration steps:

* clear channels
* set source deinterlacing
* config ntp server
* config touchscreen settings
* enable permanent logs
* create and config DCE standard channels
* create and config DCE standard recorders
* config mhpearl

Other than the CLI, epiclient is a python module that wraps `epipearl
<https://github.com/harvard-dce/epipearl>`_
calls to an epiphan-pearl in a
DCE-tailored way. Learn how you can use it by checking
section Testing_.



*******************************************************
epiclient CLI
*******************************************************

To check how to use epiclient CLI, after Install_:

.. code-block:: bash

epiclient --help
Usage: epiclient [OPTIONS] DEVICE_NAME

Console script for epiclient

Options:
--step
[validate|clear_channels|config_deinterlacing|create_channels|config_layout|config_encodings|config_recorder_channels|config_recorder_settings|config_live|config_general|config_mhpearl]
when absent, execute all steps
--verbose / --quiet
--help Show this message and exit.


The CLI requires some environment variables to be defined in a file
named ${HOME}/.epiclient.env.

Check an example in `epiclient/example.env`.

.. code-block:: bash

# edit example.env
cat epiclient/example.env
# ca status board url, user/pwd
CA_STATS_URL=http://ca-status.dceapp.net/ca_stats/ca_stats.json
CA_STATS_USER=user
CA_STATS_PASSWD=password

# user/pwd for the epiphan-pearl being configured
EPIPEARL_USER=user
EPIPEARL_PASSWD=password

# digest user/pwd for opencast cluster epiphan-pearl points to
MH_DIGEST_USER=user
MH_DIGEST_PASSWD=password

# akamai user/pwd (live streaming)
RTMP_USER=user
RTPM_PASSWD=password

# s3 bucket that hosts ca json settings
CA_SETTINGS_BUCKET=capture-agent-settings

# set as ${HOME}/.epiclient.env
cp epiclient/example.env ${HOME}/.epiclient.env

The CLI also needs AWS credentials configuration; check `aws configuration
<https://boto3.readthedocs.io/en/latest/guide/quickstart.html#configuration>`_

The CLI takes a device name as input, for example, `dev-epiphan006`. Each
device has a json config that follows a json-schema specified in
``epiclient/json_schema/role_settings_schema.json``. An example of what a json
input look like can be found in ``tests/json/primary_sample_ca_settings.json``.

These json configs are pulled from s3, and the bucket name can be configured in
the .epiclient.env file as `CA_SETTINGS_BUCKET`.



Steps
=======================================================

Some steps are required in specific order (e.g: ``config_recorder_channels``
after ``create_channels``), but in general the steps are independent.

Note that, if you ``create_channels``, the channel IDs change and the json
input is updated in s3.

Also, when you ``create_channels``, it does not check if there is already a
channel with the same name. That's why there is a ``clear_channels`` command:
make sure no previous DCE channels exist and prevent creating many
channels with the same name. This is to guarantee the defaults are set or unset
as expected.

validate
-------------------------------------------------------

This step always runs to check that the json input conforms to the json-schema
spec'd in ``epiphan/json_schema/role_settings_schema.json``.


clear_channels
-------------------------------------------------------

As said, to make sure that there won't be channels or recorders with the same
name, clear all channels first. This operation removes channels named after DCE
(*dce_pr*, *dce_pn*, *dce_prpn*, *dce_live*, *dce_live_lowbr*).

Only ``create_channels`` step works after running this operation because all
subsequent steps assume the channels are created


config_deinterlacing
-------------------------------------------------------

Source signals are usually set to be deinterlaced.

This step is independent.


create_channels
-------------------------------------------------------

Creates the DCE channels and recorders, and names them as mentioned before
*dce_pr*, *dce_pn*, *dce_prpn*, *dce_live*, *dce_live_lowbr*.

As mentioned, this step updates the s3 json config.


config_layout
-------------------------------------------------------

Configures the source layout for each channel. Note that the live channels
have the presenter and presentation combined, thus a different framesize than
presenter-only or presentation-only.

This step depends on ``create_channels``.


config_encodings
-------------------------------------------------------

Basic settings for encodings in each channel.

This step depends on ``create_channels``.


config_recorder_channels
-------------------------------------------------------

The recorder, *dce_prpn*, is the channel that is recorded as a media file and
then ingested into `Opencast Matterhorn <http://opencast.org/matterhorn>`_.
This step configures which channels are to be combined and saved into file
(*dce_pr* and *dce_pn*).

This step depends on ``create_channels``.


config_recorder_settings
-------------------------------------------------------

Configures the max size of recorded media files: per duration or file size.

This step depends on ``create_channels``.


config_live
-------------------------------------------------------

Sets the server to which live stream a channel.

This step depends on ``create_channels``.


config_general
-------------------------------------------------------

Here, the ntp server, the touchscreen settings, and the permanent log are
enabled.

This step is independent.


config_mhpearl
-------------------------------------------------------

This operation configures settings for
`mhpearl <https://bitbucket.org/hudcede/mhpearl>`_.

This step is independent.


ommit --step option
-------------------------------------------------------

When the `--step` option is absent, all steps are run and you should have a
configured capture agent at the end of the process. The CLI logs steps as they
are done so you have an idea of what's going on.


epiclient JSON Input
=======================================================

Here is an example of json input with comments for clarification. As mentioned,
these configs are stored in a s3 bucket.

.. code-block:: json

{
# name of capture agent that identifies it; only alpha-num and
# underscores allowed
"ca_name_id": "my_dev_epiphan",

# capture agent url for web ui; hostname preferred instead of IP
"ca_url": "http://epiphan.pearl.url",

# capture card id is specific to each device, if not present,
# epiclient queries the device and saves this info generating a new json
"ca_capture_card_id": "D12345678",

# similar to capture card id; epiclient queries device if not present
"ca_serial_number": "ED123456",

# valid values 'primary' or 'secondary'
# currently, 'experimental' devices are not automatically configurable
"role": "primary",

# opencast cluster name this capture agent must point to
"cluster_name_id": "cluster-name",

# if the opencast cluster is a dev, stage or prod cluster
"cluster_env": "dev",

# opencast admin node url, to pull scheduling from
"mh_admin_url": "http://cluster.dev.url.edu",

# firmware version, for now it's not mandatory
"firmware_version": "epiphan_firmware_version",

# mhpearl version, for now it's not mandatory
"mhpearl_version": "some_version",

# mhpearl settings file_search_range in secs
"mhpearl_file_search_range": 58,

# mhpearl settings update_frequency in secs
"mhpearl_update_frequency": 123,

# mhpearl settings, which ca_name pull scheduling of
"mh_ca_name": "my-dev-epiphan",

# classroom id; only alpha_num and underscores allowed
"location_name_id": "lab",

# presenter connector available in the classroom this device
# is installed
"pr_vconnector": "sdi",
"pr_vinput": "a",

# presentation connector available in the classroom this device
# is installed
"pn_vconnector": "hdmi",
"pn_vinput": "b",

# default for all epiphan-pearls is to have deinterlacing ON
"source_deinterlacing": "on",

"date_and_time": {
# ntp server, ask sysops for correct value
"ntp_server": "0.nz.pool.ntp.org",

# timezone is usually 'US/Eastern'
"timezone": "US/Alaska"
},
"touchscreen": {
# after timeout, touchscreen turns off
"episcreen_timeout": 579
},
"maintenance": {
# default is to have permanent logs ON
"permanent_logs": "on"
},
# these are common encoding settings for all channels
"channel_encodings": {
"autoframesize": "",
"codec": "h.264",
"vprofile": "100",
"vencpreset": "5",
"vkeyframeinterval": 1,
"fpslimit": 30,
"audio": "on",
"audiochannels": "1",
"audiopreset": "libfaac;44100"
},
"channels": {
"dce_pr": {
# this value, ``channel_id``, changes when creating a new
# channel; epiclient will update this value and save it in a
# new json file, backing up the previous json with a timestamp
"channel_id": "1",
"encodings": {
"framesize": "1280x720",
"vbitrate": 9000,
"audiobitrate": 160
}
},
"dce_pn": {
"channel_id": "2",
"encodings": {
"framesize": "1920x1080",
"vbitrate": 9000,
"audiobitrate": 160
}
},
"dce_live": {
"channel_id": "3",
"encodings": {
"framesize": "1920x1080",
"vbitrate": 4000,
"audiobitrate": 96
},

# akamai streaming entry point
# note that this differ for primary and secondary CAs
"rtmp_url": "rtmp://p.blobblob.i.streamingservice.org/EntryPoint",
"stream_name": "my-dev-epiphan"
},
"dce_live_lowbr": {
"channel_id": "4",
"encodings": {
"framesize": "960x270",
"vbitrate": 250,
"audiobitrate": 64
},

# akamai streaming entry point
# note that this differ for primary and secondary CAs
# this value is the same as **dce_live**
"rtmp_url": "rtmp://p.blobblob.i.streamingservice.org/EntryPoint",
"stream_name": "my-dev-epiphan"
}
},
"recorders": {
"dce_prpn": {

# this value, ``recorder_id``, changes when creating a new
# recorder; epiclient will update this value and save it in a
# new json file, backing up the previous json with a timestamp
"recorder_id": "1",
"timelimit": 360,
"sizelimit": 64000000,
"output_format": "avi"
}
}
}





*******************************************************
Testing
*******************************************************

During development, epiclient tests were executed using
`pytest <http://pytest.org/latest/>`_

To run tests from a local clone:

.. code-block:: bash

pip install -r requirements_dev.txt
py.test tests

# to run live tests
export EPI_URL=http//epiphan_pearl_address
export EPI_USER=admin_user
export EPI_PASSWD=password
py.test tests --runlive

Live tests will connect with the actual device at the given ``EPI_URL`` and
change its settings! You might have to tweak
``tests/json/primary_sample_ca_settings.json`` for that to work.


*******************************************************
License
*******************************************************

epiclient is licensed under the `Apache License, Version 2.0
<http://ww.apache.org/licenses/LICENSE-2.0>`_.


*******************************************************
Copyright
*******************************************************

2016~2017 President and Fellows of Harvard College



.. _epipearl: https://github.com/harvard-dce/epipearl
.. _`Opencast Matterhorn`: http://opencast.org/matterhorn
.. _mhpearl: https://bitbucket.org/hudcede/mhpearl
.. _pytest: http://pytest.org/latest/
.. _`Apache License, Version 2.0`: http://www.apache.org/licenses/LICENSE-2.0


=======
History
=======


0.1.3 (2016-08-03)
------------------

* clean_channels back to removing channel by name (instead of id)


0.1.2 (2016-08-02)
------------------

* Fix static files into package


0.1.1 (2016-08-02)
------------------

* Fix release version


0.1.0 (2016-06-21)
------------------

* First release on PyPI.

Project details


Download files

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

Source Distribution

epiclient-0.1.3.tar.gz (19.7 kB view details)

Uploaded Source

File details

Details for the file epiclient-0.1.3.tar.gz.

File metadata

  • Download URL: epiclient-0.1.3.tar.gz
  • Upload date:
  • Size: 19.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No

File hashes

Hashes for epiclient-0.1.3.tar.gz
Algorithm Hash digest
SHA256 36362fbeee5c8ff1c7d98c126a84459b5d6aec0bb72513fed06d71bddbd1b075
MD5 b71e97c52a6af134ce774d012a337dca
BLAKE2b-256 710960d820a31d9a66c17bfcb78f11218b3a1f60fdc5ef09d8f6c4fabee9ce26

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page