Add opencast#18, script to simulate ingestion of multiple types of me…

…diapackages

simulate-ingests/README.md

@@ -0,0 +1,12 @@
+# Opencast Simulate Ingests
+Simulate ingesting of recordings of using multiple mediapackage profiles.
+
+## Requirements
+
+* depends on python-jinja2, python-requests
+* appropriate presentation and presenter tracks in that match the mediapackage profiles, 
+see mediapackages/tracks/README.md for suggestions on how to locate suitable data.
+
+## Configuration
+
+Copy simulation-example.properties to simulation.properties and edit according to the comments.

simulate-ingests/__init__.py

@@ -0,0 +1 @@
+__author__ = 'andrew wilson, james perrin'

simulate-ingests/mediapackages/templates/episode.xml

@@ -0,0 +1,9 @@
+<?xml version='1.0' encoding='UTF-8'?>
+<dublincore xmlns="http://www.opencastproject.org/xsd/1.0/dublincore/" xmlns:dcterms="http://purl.org/dc/terms/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance/">
+    <dcterms:created>{{ created }}</dcterms:created>
+    <dcterms:identifier>{{ identifier }}</dcterms:identifier>
+    <dcterms:title>{{ title }}</dcterms:title>
+    <dcterms:isPartOf>{{ is_part_of }}</dcterms:isPartOf>
+    <dcterms:source>{{ source }}</dcterms:source>
+    <dcterms:spatial>{{ location }}</dcterms:spatial>
+</dublincore>

simulate-ingests/mediapackages/templates/manifest.xml

@@ -0,0 +1,48 @@
+<?xml version='1.0' encoding='UTF-8'?>
+<mediapackage xmlns="http://mediapackage.opencastproject.org" duration="{{ duration }}" id="{{ identifier }}" start="{{ created }}">
+    <media>
+        {%- if has_presentation_video %}
+        <track id="track-0" type="presentation/source">
+            <mimetype>video/{{ video_mime }}</mimetype>
+            <url>presentation.{{ video_ext }}</url>
+            <duration>{{ duration }}</duration>
+        </track>
+        {% endif -%}
+        {%- if has_presenter_video %}
+        <track id="track-1" type="presenter/source">
+            <mimetype>video/{{ video_mime }}</mimetype>
+            <url>presenter.{{ video_ext }}</url>
+            <duration>{{ duration }}</duration>
+        </track>
+        {% endif -%}
+        {%- if has_presentation_audio %}
+        <track id="track-2" type="presentation/source">
+            <mimetype>audio/{{ audio_mime }}</mimetype>
+            <url>presentation.{{ audio_ext }}</url>
+            <duration>{{ duration }}</duration>
+        </track>
+        {% endif -%}
+        {%- if has_presenter_audio %}
+        <track id="track-3" type="presenter/source">
+            <mimetype>audio/{{ audio_mime }}</mimetype>
+            <url>presenter.{{ audio_ext }}</url>
+            <duration>{{ duration }}</duration>
+        </track>
+        {% endif %}
+    </media>
+    <metadata>
+        <catalog id="catalog-0" type="dublincore/episode">
+            <mimetype>text/xml</mimetype>
+            <url>episode.xml</url>
+        </catalog>
+        <catalog id="catalog-1" type="dublincore/series">
+            <mimetype>text/xml</mimetype>
+            <url>series.xml</url>
+        </catalog>
+    </metadata>
+    <attachments>
+        <attachment id="org.opencastproject.capture.agent.properties">
+            <url>org.opencastproject.capture.agent.properties</url>
+        </attachment>
+    </attachments>
+</mediapackage>

simulate-ingests/mediapackages/templates/org.opencastproject.capture.agent.properties

@@ -0,0 +1,7 @@
+event.location={{ location }}
+event.series={{ series_id }}
+event.title={{ title }}
+org.opencastproject.workflow.definition={{ workflow }} 
+org.opencastproject.workflow.config.emailAddresses={{ email }}
+org.opencastproject.workflow.config.publishToOaiPmh={{ publish }}
+org.opencastproject.workflow.config.editRecording={{ edit }}

simulate-ingests/mediapackages/templates/series.xml

@@ -0,0 +1,7 @@
+<dublincore xmlns="http://www.opencastproject.org/xsd/1.0/dublincore/" xmlns:dcterms="http://purl.org/dc/terms/" xmlns:oc="http://www.opencastproject.org/matterhorn/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
+<dcterms:title>{{ title }}</dcterms:title>
+<dcterms:created xsi:type="dcterms:W3CDTF">{{ created }}</dcterms:created>
+<dcterms:source>{{ source }}</dcterms:source>
+<oc:annotation>false</oc:annotation>
+<dcterms:identifier>{{ identifier }}</dcterms:identifier>
+</dublincore>

simulate-ingests/mediapackages/tracks/README.md

@@ -0,0 +1,46 @@
+# Mediapackage Tracks Directory
+Video and audio tracks should be present here that match the tracks listed in the mediapackage 
+profiles. The filenames must have the form:
+
+    [presenation|presenter]-<type>-<duration>.[avi|mp3]
+
+e.g
+
+    presenter-single-1.avi
+
+## Filtering Sample Mediapackage from Database
+
+Analysing the mediapackage xml in the archive is expensive, therefore create a temporary table (#tablename) for 
+selecting prospective tracks:
+
+    /* create temp table */
+    with xmlnamespaces (default 'http://mediapackage.opencastproject.org')
+    select
+      id,
+      mp.value('(/mediapackage/media/track/@type)[1]', 'varchar(256)') as track_1,
+      left(mp.value('(/mediapackage/media/track/mimetype)[1]', 'varchar(256)'), 5) as mime_1,
+      mp.value('(/mediapackage/media/track/@type)[2]', 'varchar(256)') as track_2,
+      left(mp.value('(/mediapackage/media/track/mimetype)[2]', 'varchar(256)'), 5) as mime_2,
+      mp.value('(/mediapackage/media/track/@type)[3]', 'varchar(256)') as track_3,
+      left(mp.value('(/mediapackage/media/track/mimetype)[3]', 'varchar(256)'), 5) as mime_3,
+      mp.value('(/mediapackage/@duration)[1]', 'int')/60000 as duration
+      into #mp_data
+      from
+      /* Cast the xml strings to UTF-16 encoded xml fields */
+      (select id, cast(replace(cast(mediapackage_xml as nvarchar(max)), 'UTF-8', 'UTF-16') as xml) as mp
+        from dbo.mh_archive_episode
+        /* Pre filter the data */
+        where modification_date > '2017-01-10 00:00:00'
+        and version = 0 and deleted = 0) as mpxml;
+        
+Once the temporary table #mp_data has been created queries can be run against it get ids of prospective example 
+mediapackages:
+
+    select id
+      from #mp_data
+      where
+        duration=55
+        and track_1='presenter/source' and mime_1='video'
+        and track_2='presenter/source' and mime_2='audio'
+        and track_3 is null and mime_3 is null;
+