Skip to content

Commit 88ffe25

Browse files
jrobblejrobble
authored
Update release notes for R6.2.0. (#128)
1 parent 7b1865c commit 88ffe25

5 files changed

Lines changed: 253 additions & 24 deletions

File tree

‎docs/docs/Release-Notes.md‎

Lines changed: 105 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,112 @@
11
> **NOTICE:** This software (or technical data) was produced for the U.S. Government under contract, and is subject to the Rights in Data-General Clause 52.227-14, Alt. IV (DEC 2007). Copyright 2021 The MITRE Corporation. All Rights Reserved.
22
3+
# OpenMPF 6.2.x
4+
5+
<h2>6.2.0: May 2021</h2>
6+
7+
<h3>Tesseract OCR Text Detection Component Support for Videos</h3>
8+
9+
- The component can now process videos in addition to images and PDFs. Each video frame is processed sequentially.
10+
The `MAX_PARALLEL_SCRIPT_THREADS` property determines how many threads to use to process each frame, one thread per
11+
language or script.
12+
- Note that for videos without much text, it may be faster to disable threading by
13+
setting `MAX_PARALLEL_SCRIPT_THREADS=1`. This will allow the component to reuse TessAPI instances instead of creating
14+
new ones for every frame. Please refer to the Known Issues section.
15+
- Resolved issues: [#1285](https://github.com/openmpf/openmpf/issues/1285)
16+
17+
<h3>Updates</h3>
18+
19+
- [[#1086](https://github.com/openmpf/openmpf/issues/1086)] Added support for `COULD_NOT_OPEN_MEDIA`
20+
and `COULD_NOT_READ_MEDIA` error types
21+
- [[#1159](https://github.com/openmpf/openmpf/issues/1159)] Split `IssueCodes.REMOTE_STORAGE`
22+
into `REMOTE_STORAGE_DOWNLOAD` and `REMOTE_STORAGE_UPLOAD`
23+
- [[#1250](https://github.com/openmpf/openmpf/issues/1250)] Modified `/rest/jobs/{id}` to include the job's media
24+
- [[#1312](https://github.com/openmpf/openmpf/issues/1312)] Created `NETWORK_ERROR` error code for when a component
25+
can't connect to an external server. Updated Python HTTP retry code to return `NETWORK_ERROR`. This affects the Azure
26+
components.
27+
28+
<h3>Known Issues</h3>
29+
30+
- [[#1008](https://github.com/openmpf/openmpf/issues/1008)] Use global TessAPI instances with parallel processing
31+
332
# OpenMPF 6.1.x
433

34+
<h2>6.1.6: May 2021</h2>
35+
36+
<h3>Handle Variable Frame Rate Videos</h3>
37+
38+
- The Workflow Manager will attempt to detect if a video is constant frame rate (CFR) or variable frame rate (VFR)
39+
during media inspection. If no determination can be made, it will default to VFR behavior. If CFR, the JSON output
40+
object will have a `HAS_CONSTANT_FRAME_RATE=true` property in the `mediaMetadata` field.
41+
- When `MPFVideoCapture` handles a CFR video it will use OpenCV to set the frame position, unless the position is within
42+
16 frames of the current position, in which case it will iteratively use OpenCV `grab()` to advance to the desired
43+
frame.
44+
- When `MPFVideoCapture` handles a VFR video it will always iteratively use OpenCV `grab()` to advance to the desired
45+
frame because setting the frame position directly has been shown to not work correctly on VFR videos.
46+
- When a video is split into multiple segments, `MPFVideoCapture` must iteratively use `grab()` to advance from frame 0
47+
to the start of the segment. This introduces performance overhead. To mitigate this we recommend using larger video
48+
segments than those used for CFR videos.
49+
- In addition to the existing `TARGET_SEGMENT_LENGTH` and `MIN_SEGMENT_LENGTH` job
50+
properties (`detection.segment.target.length` and `detection.segment.minimum.length` system properties) for CFR
51+
videos, the Workflow Manager now supports the `VFR_TARGET_SEGMENT_LENGTH` and `VFR_MIN_SEGMENT_LENGTH` job
52+
properties (`detection.vfr.segment.target.length` and `detection.vfr.segment.minimum.length` system properties) for
53+
VFR videos.
54+
- Note that the timestamps associated with tracks and detections in a VFR video may be wrong. Please refer to the Known
55+
Issues section.
56+
- Resolved issues: [#1307](https://github.com/openmpf/openmpf/issues/1307)
57+
58+
<h3>Updates</h3>
59+
60+
- [[#1287](https://github.com/openmpf/openmpf/issues/1287)] Updated Tika Text Detection Component to break up large
61+
chunks of text. The component now generates tracks with both a `PAGE_NUM` property and `SECTION_NUM` property. Please
62+
refer to
63+
the [README](https://github.com/openmpf/openmpf-components/blob/master/java/TikaTextDetection/README.md#overview).
64+
65+
<h3>Known Issues</h3>
66+
67+
- [[#1313](https://github.com/openmpf/openmpf/issues/1313)] Incorrect JSON output object timestamps for variable frame
68+
rate videos
69+
- [[#1317](https://github.com/openmpf/openmpf/issues/1317)] Tika Text Detection component generates first PDF track
70+
at `PAGE_NUM` 2
71+
72+
<h2>6.1.5: April 2021</h2>
73+
74+
<h3>Updates</h3>
75+
76+
- [[#1300](https://github.com/openmpf/openmpf/issues/1300)] Parallelized S3 artifact upload. Use
77+
the `detection.artifact.extraction.parallel.upload.count` system property to configure the number of parallel uploads.
78+
79+
<h2>6.1.4: April 2021</h2>
80+
81+
<h3>Updates</h3>
82+
83+
- [[#1299](https://github.com/openmpf/openmpf/issues/1299)] Improved artifact extraction performance when there is no
84+
rotation or flip
85+
86+
<h2>6.1.3: April 2021</h2>
87+
88+
<h3>Updates</h3>
89+
90+
- [[#1295](https://github.com/openmpf/openmpf/issues/1295)] Improved artifact extraction and markup JNI memory
91+
utilization
92+
- [[#1297](https://github.com/openmpf/openmpf/issues/1297)] Limited Workflow Manager IO threads to a reasonable number
93+
94+
<h3>Bug Fixes</h3>
95+
96+
- [[#1296](https://github.com/openmpf/openmpf/issues/1296)] Fixed ActiveMQ job priorities
97+
98+
<h2>6.1.2: April 2021</h2>
99+
100+
<h3>Updates</h3>
101+
102+
- [[#1294](https://github.com/openmpf/openmpf/issues/1294)] Limited ffmpeg threads to a reasonable number
103+
104+
<h2>6.1.1: April 2021</h2>
105+
106+
<h3>Bug Fixes</h3>
107+
108+
- [[#1292](https://github.com/openmpf/openmpf/issues/1292)] Don't skip artifact extraction for failed media
109+
5110
<h2>6.1.0: April 2021</h2>
6111

7112
<h3>OpenMPF Command Line Runner</h3>

‎docs/site/Release-Notes/index.html‎

Lines changed: 119 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -64,6 +64,9 @@
6464

6565
<ul>
6666

67+
<li class="toctree-l3"><a href="#openmpf-62x">OpenMPF 6.2.x</a></li>
68+
69+
6770
<li class="toctree-l3"><a href="#openmpf-61x">OpenMPF 6.1.x</a></li>
6871

6972

@@ -300,7 +303,123 @@
300303
<blockquote>
301304
<p><strong>NOTICE:</strong> This software (or technical data) was produced for the U.S. Government under contract, and is subject to the Rights in Data-General Clause 52.227-14, Alt. IV (DEC 2007). Copyright 2021 The MITRE Corporation. All Rights Reserved.</p>
302305
</blockquote>
306+
<h1 id="openmpf-62x">OpenMPF 6.2.x</h1>
307+
<h2>6.2.0: May 2021</h2>
308+
309+
<h3>Tesseract OCR Text Detection Component Support for Videos</h3>
310+
311+
<ul>
312+
<li>The component can now process videos in addition to images and PDFs. Each video frame is processed sequentially.
313+
The <code>MAX_PARALLEL_SCRIPT_THREADS</code> property determines how many threads to use to process each frame, one thread per
314+
language or script.</li>
315+
<li>Note that for videos without much text, it may be faster to disable threading by
316+
setting <code>MAX_PARALLEL_SCRIPT_THREADS=1</code>. This will allow the component to reuse TessAPI instances instead of creating
317+
new ones for every frame. Please refer to the Known Issues section.</li>
318+
<li>Resolved issues: <a href="https://github.com/openmpf/openmpf/issues/1285">#1285</a></li>
319+
</ul>
320+
<h3>Updates</h3>
321+
322+
<ul>
323+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1086">#1086</a>] Added support for <code>COULD_NOT_OPEN_MEDIA</code>
324+
and <code>COULD_NOT_READ_MEDIA</code> error types</li>
325+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1159">#1159</a>] Split <code>IssueCodes.REMOTE_STORAGE</code>
326+
into <code>REMOTE_STORAGE_DOWNLOAD</code> and <code>REMOTE_STORAGE_UPLOAD</code></li>
327+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1250">#1250</a>] Modified <code>/rest/jobs/{id}</code> to include the job's media</li>
328+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1312">#1312</a>] Created <code>NETWORK_ERROR</code> error code for when a component
329+
can't connect to an external server. Updated Python HTTP retry code to return <code>NETWORK_ERROR</code>. This affects the Azure
330+
components.</li>
331+
</ul>
332+
<h3>Known Issues</h3>
333+
334+
<ul>
335+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1008">#1008</a>] Use global TessAPI instances with parallel processing</li>
336+
</ul>
303337
<h1 id="openmpf-61x">OpenMPF 6.1.x</h1>
338+
<h2>6.1.6: May 2021</h2>
339+
340+
<h3>Handle Variable Frame Rate Videos</h3>
341+
342+
<ul>
343+
<li>The Workflow Manager will attempt to detect if a video is constant frame rate (CFR) or variable frame rate (VFR)
344+
during media inspection. If no determination can be made, it will default to VFR behavior. If CFR, the JSON output
345+
object will have a <code>HAS_CONSTANT_FRAME_RATE=true</code> property in the <code>mediaMetadata</code> field.</li>
346+
<li>When <code>MPFVideoCapture</code> handles a CFR video it will use OpenCV to set the frame position, unless the position is within
347+
16 frames of the current position, in which case it will iteratively use OpenCV <code>grab()</code> to advance to the desired
348+
frame.</li>
349+
<li>When <code>MPFVideoCapture</code> handles a VFR video it will always iteratively use OpenCV <code>grab()</code> to advance to the desired
350+
frame because setting the frame position directly has been shown to not work correctly on VFR videos.</li>
351+
<li>When a video is split into multiple segments, <code>MPFVideoCapture</code> must iteratively use <code>grab()</code> to advance from frame 0
352+
to the start of the segment. This introduces performance overhead. To mitigate this we recommend using larger video
353+
segments than those used for CFR videos.</li>
354+
<li>In addition to the existing <code>TARGET_SEGMENT_LENGTH</code> and <code>MIN_SEGMENT_LENGTH</code> job
355+
properties (<code>detection.segment.target.length</code> and <code>detection.segment.minimum.length</code> system properties) for CFR
356+
videos, the Workflow Manager now supports the <code>VFR_TARGET_SEGMENT_LENGTH</code> and <code>VFR_MIN_SEGMENT_LENGTH</code> job
357+
properties (<code>detection.vfr.segment.target.length</code> and <code>detection.vfr.segment.minimum.length</code> system properties) for
358+
VFR videos.</li>
359+
<li>Note that the timestamps associated with tracks and detections in a VFR video may be wrong. Please refer to the Known
360+
Issues section.</li>
361+
<li>Resolved issues: <a href="https://github.com/openmpf/openmpf/issues/1307">#1307</a></li>
362+
</ul>
363+
<h3>Updates</h3>
364+
365+
<ul>
366+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1287">#1287</a>] Updated Tika Text Detection Component to break up large
367+
chunks of text. The component now generates tracks with both a <code>PAGE_NUM</code> property and <code>SECTION_NUM</code> property. Please
368+
refer to
369+
the <a href="https://github.com/openmpf/openmpf-components/blob/master/java/TikaTextDetection/README.md#overview">README</a>.</li>
370+
</ul>
371+
<h3>Known Issues</h3>
372+
373+
<ul>
374+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1313">#1313</a>] Incorrect JSON output object timestamps for variable frame
375+
rate videos</li>
376+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1317">#1317</a>] Tika Text Detection component generates first PDF track
377+
at <code>PAGE_NUM</code> 2</li>
378+
</ul>
379+
<h2>6.1.5: April 2021</h2>
380+
381+
<h3>Updates</h3>
382+
383+
<ul>
384+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1300">#1300</a>] Parallelized S3 artifact upload. Use
385+
the <code>detection.artifact.extraction.parallel.upload.count</code> system property to configure the number of parallel uploads.</li>
386+
</ul>
387+
<h2>6.1.4: April 2021</h2>
388+
389+
<h3>Updates</h3>
390+
391+
<ul>
392+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1299">#1299</a>] Improved artifact extraction performance when there is no
393+
rotation or flip</li>
394+
</ul>
395+
<h2>6.1.3: April 2021</h2>
396+
397+
<h3>Updates</h3>
398+
399+
<ul>
400+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1295">#1295</a>] Improved artifact extraction and markup JNI memory
401+
utilization</li>
402+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1297">#1297</a>] Limited Workflow Manager IO threads to a reasonable number</li>
403+
</ul>
404+
<h3>Bug Fixes</h3>
405+
406+
<ul>
407+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1296">#1296</a>] Fixed ActiveMQ job priorities</li>
408+
</ul>
409+
<h2>6.1.2: April 2021</h2>
410+
411+
<h3>Updates</h3>
412+
413+
<ul>
414+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1294">#1294</a>] Limited ffmpeg threads to a reasonable number</li>
415+
</ul>
416+
<h2>6.1.1: April 2021</h2>
417+
418+
<h3>Bug Fixes</h3>
419+
420+
<ul>
421+
<li>[<a href="https://github.com/openmpf/openmpf/issues/1292">#1292</a>] Don't skip artifact extraction for failed media</li>
422+
</ul>
304423
<h2>6.1.0: April 2021</h2>
305424

306425
<h3>OpenMPF Command Line Runner</h3>

‎docs/site/index.html‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -437,5 +437,5 @@ <h1 id="overview">Overview</h1>
437437

438438
<!--
439439
MkDocs version : 0.16.0
440-
Build Date UTC : 2021-04-07 19:44:01
440+
Build Date UTC : 2021-05-13 17:40:17
441441
-->

‎docs/site/mkdocs/search_index.json‎

Lines changed: 6 additions & 1 deletion
Large diffs are not rendered by default.

0 commit comments

Comments
 (0)