teltek / galicaster Goto Github PK

See:

• http://jderose.blogspot.com.es/2012/04/how-to-create-weak-reference-to.html
• kivy/kivy#70

We have noticed this after discovering that the number of file descriptors open by Galicaster (in /proc/PID/fd) steadily grows with each recording.

Modify the medatadata editor in order to select the visibility of known metadata (allow editting, just viewing or hide it) and make means to add simple custom metadata.

And change logging to use a timedrotatingfilehandler

See:

http://docs.python.org/2/library/logging.handlers.html#sysloghandler

Galicaster doesnot show the dialogboxes when compiz is disabled.

Traceback:

Traceback (most recent call last):
  File "rr.py", line 10, in <module>
    serializer.save_in_zip(mp, location, None)
  File "/usr/share/galicaster/galicaster/mediapackage/serializer.py", line 102, in save_system_zip
    m = open(tmp_file,'w')
IOError: [Errno 13] Permission denied: 'manifest.xml'

The temp file in save_system_zip should create in a temp directory.

def save_system_zip(mp, loc, use_namespace=True):

    tmp_file = 'manifest.xml'
    m = open(tmp_file,'w')
    m.write(set_manifest(mp, use_namespace))
    m.close()

Option to translate Galicaster GUI to other langues.
as suggestion: Use Transifex for translations

I'm using the following test profile to record my webcam and
audiotest and encode them as mp4 and aac.

[data]
name = MyEncode

[track1]
name = WebCam
caps = image/jpeg,framerate=24/1,width=1280,height=720
active = True
location = /dev/video0
file = camera.mp4
device = v4l2
flavor = presenter
videoencoder = videorate ! videoscale add-borders=true ! video/x-raw-yuv,width=320,height=240,framerate=10/1 ! x264enc bitrate=1200 pass=5 quantizer=22
muxer = mp4mux faststart=true

[track2]
amplification = 1.0
name = Noise
vumeter = True
pattern = pink-noise
audioencoder = faac outputformat=1 bitrate=128000
volume = 0.3
player = True
frequency = 440
location = default
file = sound.aac
device = audiotest
flavor = presenter
active = True

The issue I'm seeing it that the zip file ingested by MH contains a truncated version of the camera.mp4 file. If I use faststart the file is actual 0bytes in length, if I don't then it is is a few kbytes shorter. It appears as though galicaster is not waiting for the gstreamer to finish writing the file before zipping and uploading it to MH. MH is then failing at the inspection operation.

Using faststart=true requires the muxer needs to know the size of video track so it can write the headers out first, it does not write anything to the camera.mp4 file until the recording has completed (it does write data to /tmp/qtmux####). This means that if the recording fails you will not have anything on disk to try to salvage (it may not be possible to read the track without a header anyway).

I report this on list with RC2 but still appears to be in 1.3 release

The audio dialog is quite annoying sometimes, like when you want to close a recording.

The idea is to separate this feature from the admin enable/disable option.
If visibility is on tracks are selectable for the MH scheduler, otherwise it records the current profile.

When you plug and unplug the VGA signal, you can see in the stderr:

(run_galicaster.py:7953): GStreamer-CRITICAL **: 
Trying to dispose element device, but it is in PAUSED instead of the NULL state.
You need to explicitly set elements to the NULL state before
dropping the final reference, to allow them to clean up.
This problem may also be caused by a refcounting bug in the
application or some element.

When a manual recording is active, Galicaster still reports itself as "Idle"

See https://bug811261.bugzilla.mozilla.org/attachment.cgi?id=693331

It would be really helpful to have kind of monitoring tool for remote Galicaster units. (to quickly check Remote Host Status, Downtime reports, Remote Stop/Start/Restart ..)

Something similar:

• https://github.com/polimediaupv/dashboard

My firewire card is:

06:01.0 FireWire (IEEE 1394): VIA Technologies, Inc. VT6306/7/8 [Fire II(M)] IEEE 1394 OHCI Controller (rev 46) (prog-if 10 [OHCI])
Subsystem: Allied Telesis, Inc IEEE 1394 4port DCST 1394-3+1B
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Latency: 32 (8000ns max), Cache Line Size: 64 bytes
Interrupt: pin A routed to IRQ 17
Region 0: Memory at fb300000 (32-bit, non-prefetchable) [size=2K]
Region 1: I/O ports at d000 [size=128]
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1- D2+ AuxCurrent=0mA PME(D0-,D1-,D2+,D3hot+,D3cold+)
Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
Kernel driver in use: firewire_ohci
Kernel modules: firewire-ohci

Using GStreamer the card works out.

gst-launch-0.10 -m --gst-debug=2 \
\(\( dv1394src ! queue ! dvdemux name=demux ! \
level message=true interval=100000000 ! volume ! alsasink sync=false \
demux. ! queue ! ffdec_dvvideo ! ffmpegcolorspace ! \
xvimagesink  qos=false async=false sync=false \)\)

You can reproduce the bug using https://gist.github.com/dde3276a341494b75115

Mediapackage Id not set correctly in galicaster

1.- Program a new episode in matterhorn
2.- Galicaster records that episode
Galicaster greates a new folder named gc_{agentname}_{workflowid}

Check that galicaster is using the workflow id for folder names... In
that folfer the manifest.xml
is something like this:

Galicaster is putting in the mediapackage ID thw workflowId and not the
mediapackageId that must be something like
2c32c4d5-7216-4ea3-9749-5820d1edde3a

The same happen in episode.xml

The dublincore tag is "isPartOf" not "ispartof".

See MH-7831 issue for more info.

Let Galicaster run 24/7, but only activate it's access to the devices on demand ("button to activate") or during recordings?

When recording experience an error and the error dialog is not closed, the next schedule doesn't reload the profile.

Pointed out by Tobias Schiebeck from Manchester

The Media Manager is completely reloaded in some situations. This provokes a slow access to the Media Manager when the data is accessed through shared storage.

Also, when a reference to a Mediapackage object is stored - in operations for instance - there could be inconsistencies. Before data is modified the object should be reloaded.

Consequently, storing the MP could be a bad practice, better storing the identifier only.

An undocumented bug that has been resolved on 1.2.2.
Serializing datetime metadata using toString instead of isoformat cause this bug. This could provoke a fail on ingest if the MH core doesn't have a previous episode made on schedule.

This feature will provide a marking system that could be used to point important moments or intervals in a recording.
Marks can be used by users to define trimming or edition position. Also, operations can create mark sequences regarding slides changes, trimming point, audio silence etc.

Recordings can be hidden into a second level. Once there they can be deleted on a single operation.
This feature is specially useful on studio environments where multiple takes are made of a single recording and only one will be used.

Setting "No series" to a MediaPackage with an assigned series cause a duplication of the episode.xml catalog.

The bug is probably caused by a miscall of the funciont addDublincoreAsString from mediapackage.mediapackage.py

This operation will export a Mediapackage to a given mounted position: a NAS or NTFS folder, a mounted USB device, a shared.

Paths will support templates.

I am using a mp4mux as a muxer in Galicaster, but I want to pass the parameter 'moov-recovery-file'

Example: mp4mux moov-recovery-file=${galicaster-working-folder}/file.mrf

But I Galicaster does not recognise ${galicaster-working-folder}.
¿Maybe you can add an aux.replace('${galicaster-working-folder}', self.options['path']) or something similar to the bins?

If I use a mp4mux, the moov file is written at the end of the file. If Galicaster breaks while is recording, I can not play the video recorded, so I need to create a moov recovery file in case Galicaster breaks.

Due to a bug in GTK, dialogs with scroll bars fail to be modal. Therefore, Scrolledwindow widgets should be replaced for Frames and create the scroll effect with buttons.

Alternativaly, other scrollbars can be installed.

Where you can write the pipeline in the conf file.

When using Epiphan devices with the newer drivers, quite often tracks got a duration longer than expected. The behaviour is simillar to non-live sources which provides buffers continously instead of send them synced with the clock.

If a device can be probed, reload the profile when signal input is recovered.

Using Blackmagic cards but not exclusivelly, sometimes the preview area is shown in an external video area.
The problem is caused by a desynchronization between GUI creation and device capturing.

Once activated Galicaster is only accessible by registered users on the Matterhorn LDAP. Parameters could be fetched and included on the Mediapackage.

This refactorization will provide:

• Preserved queues. both immediate and nightly.
• Hability to cancel or execute pending operations
• Customazible operations, including the possibility of having multiple operations of the same typ
• Modularity to develop new operations
• Big parametrization allowing UI interactivity to define values.
• Means to track operations parameters and status (spent time, log)

Multiple paramaters will be accessible to create a highly customized layout including: branding, margin, scaling, cropping, positioning ...

In combination with the customizable operations, multiple predefined composition can be set on the same application.

Useful for sound cards like: RME Hammerfall DSP + Digiface

NOTE: Check if the XML namespaces brakes the backwards compability

The XML will be like:

<?xml version="1.0" encoding="utf-8"?>
<mediapackage xmlns="http://mediapackage.opencastproject.org" 
      duration="1995" id="f13e2862-3dc0-4da1-9e0a-4c79271f2735" start="2012-08-02T14:20:11">
  <media>
...

It would be nice if one could click one of the video areas in Media Manager and look that video fullscreen. Second click would return to the normal side-by-side playback screen.

If any file is deleted from the repository (either intentionally or unintentionally) from outside Galicaster, then Galicaster fails to start, printing an exception similar to the following:

Traceback (most recent call last):
  File "./run_galicaster.py", line 43, in <module>
    sys.exit(main(sys.argv)) 
  File "./run_galicaster.py", line 34, in main
    gc = core.Main()
  File "/home/sinho/workspace/local/Galicaster_orig/galicaster/core/core.py", line 48, in __init__
    self.load_modules()
  File "/home/sinho/workspace/local/Galicaster_orig/galicaster/core/core.py", line 70, in load_modules
    self.listing  = ListingClassUI()
  File "/home/sinho/workspace/local/Galicaster_orig/galicaster/classui/listing.py", line 67, in __init__
    self.populate_treeview(self.repository.list().values())
  File "/home/sinho/workspace/local/Galicaster_orig/galicaster/classui/listing.py", line 105, in populate_treeview
    self.insert_data_in_list(self.lista, mp)
  File "/home/sinho/workspace/local/Galicaster_orig/galicaster/classui/listing.py", line 91, in insert_data_in_list
    long(mp.getSize()),
  File "/home/sinho/workspace/local/Galicaster_orig/galicaster/mediapackage/mediapackage.py", line 651, in getSize
    size += path.getsize(f.uri)
  File "/usr/lib/python2.7/genericpath.py", line 49, in getsize
    return os.stat(filename).st_size
OSError: [Errno 2] Non hai tal ficheiro ou directorio: '/home/sinho/Repository/gc_GCminho_20130318T11h05m42/SCREEN.mp4'

Scheduled events are not working in my installation.
We use a MH 1.3 and we have the legacy = True in [basic] configuration

The calendar.ical file is empty
$cat Repository/attach/calendar.ical
BEGIN:VCALENDAR
PRODID:Opencast Matterhorn Calendar File 0.5
VERSION:2.0
CALSCALE:GREGORIAN
END:VCALENDAR

Galicaster has to execute the script when:

• Galicaster starts
• A profile changes.

Despite framedrop being very very unfrequent, preventing it will maximize synchronism and it won't suppose big changes on the pipeline

Mediapackages will be selectable in groups and Operations can be assigned to them.

The scenario is a large number of capture agents recording scheduled lectures which all start and finish at the same time. If ingest is set to immediate all the capture agents try to push their files to the ingest server at the same time, resulting in an overloaded server. Yes, load balancing and retrying can mitigate this to some extent but rather than periodically hammering the ingest servers and network, ingest on each capture agent should be configured to occur and some random time in a given duration. eg all capture agent start to ingest in a 15 minute window after the end of a lecture.