Releases · mbari-org/annosaurus

04 Feb 21:40

0.6.1

d22bc97

annosaurus 0.6.1

This is a bug fix release. The ZeroMQ publisher was posting duplicate messages, at least in some circumstances. We now filter those out using observable.distinct where the filter compares the annotations observationUuid and observationTimestamp to determine if a message is a duplicate.

Assets 2

03 Feb 23:43

hohonuuli

0.6.0

b151f63

annosaurus 0.6.0

Added ZeroMQ notifcations that publish when an annotation has been created or updated. This is targeted at MBARI's ship deployments where ML apps can listen for user input from VARS. To deploy, modify your application.conf or pass the following environment parameters on start up to your Docker container:

MESSAGING_ZEROMQ_PORT: Defines the port to publish messages to. Default is 5563
MESSAGING_ZEROMQ_ENABLE: true/false. If you want ZMQ running set it to true. Default is false
MESSAGING_ZEROMQ_TOPIC: The ZMQ topic to publish to. The default is "vars".

The messages will be JSON. For real-time operations, the minimum JSON will be formatted like so:

{
  "observation_uuid": "49e0cd44-8915-4b6c-ad68-400ef3d41b06",
  "observer": "brian",
  "observation_timestamp": "2020-02-03T23:25:45.503055Z",
  "concept": "Grimpoteuthis bathynectes",
  "recorded_timestamp": "2020-02-03T23:25:45.503055Z",
  "video_reference_uuid": "7cb320dc-0da5-4448-a26f-79177651735c",
  "associations": [],
  "image_references": [] 
}

Here's example ZMQ subscribers:

psenvsub.java

import org.zeromq.SocketType;
import org.zeromq.ZMQ;
import org.zeromq.ZMQ.Socket;
import org.zeromq.ZContext;

/**
 * Pubsub envelope subscriber
 */
public class psenvsub {

    int port = 5563;         // MESSAGING_ZEROMQ_PORT:
    String topic = "vars"; // MESSAGING_ZEROMQ_TOPIC

    public static void main(String[] args) {
        // Prepare our context and subscriber
        try (ZContext context = new ZContext()) {
            Socket subscriber = context.createSocket(SocketType.SUB);
            subscriber.connect("tcp://localhost:" + port);
            subscriber.subscribe(topic.getBytes(ZMQ.CHARSET));

            while (!Thread.currentThread().isInterrupted()) {
                // Read envelope with address
                String address = subscriber.recvStr();
                // Read message contents
                String contents = subscriber.recvStr();
                System.out.println(address + " : " + contents);
            }
        }
    }
}

psenvsub.c

//  Pubsub envelope subscriber

#include "zhelpers.h"

int main (void)
{
    //  Prepare our context and subscriber
    void *context = zmq_ctx_new ();
    void *subscriber = zmq_socket (context, ZMQ_SUB);
    zmq_connect (subscriber, "tcp://localhost:5563");
    zmq_setsockopt (subscriber, ZMQ_SUBSCRIBE, "vars", 1);

    while (1) {
        //  Read envelope with address
        char *address = s_recv (subscriber);
        //  Read message contents
        char *contents = s_recv (subscriber);
        printf ("[%s] %s\n", address, contents);
        free (address);
        free (contents);
    }
    //  We never get here, but clean up anyhow
    zmq_close (subscriber);
    zmq_ctx_destroy (context);
    return 0;
}

psenvsub.py

"""
   Pubsub envelope subscriber
"""
import zmq

def main():
    """ main method """

    # Prepare our context and publisher
    context    = zmq.Context()
    subscriber = context.socket(zmq.SUB)
    subscriber.connect("tcp://localhost:5563")
    subscriber.setsockopt(zmq.SUBSCRIBE, b"vars")

    while True:
        # Read envelope with address
        [address, contents] = subscriber.recv_multipart()
        print("[%s] %s" % (address, contents))

    # We never get here but clean up anyhow
    subscriber.close()
    context.term()

if __name__ == "__main__":
    main()

Assets 2

30 Oct 16:25

hohonuuli

0.5.1

5e5497b

annosaurus 0.5.1

Improved bulk create/delete methods.

Create

Allows users to bulk insert annotations into annosaurus as JSON. We've tested this up to about 5000 annotations at once. Note that annosaurus will handle large bulk inserts, but be aware most clients will timeout while waiting for the insert to complete. (The results is that the annotations do get inserted though).

Endpoint

/anno/v1/annotations/bulk

Example

POST /anno/v1/annotations/bulk
Content-Type: application/json
Authorization: Bearer <jwt access token>

[
  {
    "observation_uuid": "77d71526-0f5a-4f5b-8f00-964454aaf53e",
    "concept": "Pandalus platyceros",
    "observer": "brian",
    "observation_timestamp": "2017-08-22T20:38:48.219Z",
    "video_reference_uuid": "b6be1826-7413-48cc-8b74-324477eed114",
    "imaged_moment_uuid": "52cf5b45-427a-4d01-a175-e89387da419f",
    "elapsed_time_millis": 0,
    "associations": [],
    "image_references": [
      {
        "uuid": "e9e7a01f-f675-4f08-9434-1eaf0f462c3d",
        "description": "uncompressed image",
        "url": "http://somserver.org/images/i2MAP_2020170622T173833Z-b6be1826.png",
        "format": "image/png",
        "last_updated_time": "2017-09-11T14:37:09Z"
      },
      {
        "uuid": "ec4715f3-705f-4ece-b966-05c814ab6666",
        "description": "uncompressed image",
        "url": "http://somserver.org//images/i2MAP_2020170622T173833Z-b6be1826.png",
        "format": "image/png",
        "last_updated_time": "2017-09-12T15:21:52Z"
      }
    ]
  }
]

Returns

It returns the annotations, as json, after they are inserted. Note that you do not need to define any uuids, such as observation_uuid or uuid (other than the video_reference_uuid which indicates what video the annotations are associated with), however, the returned annotations will have the uuid keys generated by the database.

Delete

Deletes all annotations associated with a given video (defined by it's video_reference_uuid}.

Endpoint

/anno/v1/fast/videoreference/{video_reference_uuid}

Example

DELETE /anno/v1/fast/videoreference/b6be1826-7413-48cc-8b74-324477eed114
Authorization: Bearer <jwt access token>

Returns

Counts of all the objects deleted

{
  "video_reference_uuid": "ef78c89e-079e-4e78-9645-56cfb2c23e3e",
  "ancillary_data_count": 0,
  "image_reference_count": 108,
  "association_count": 10,
  "observation_count": 500,
  "imaged_moment_count": 499
}

Assets 2

04 Oct 17:05

hohonuuli

0.5.0

45e3be8

annosaurus 0.5.0

We're working on adding methods to support exporting of datasets to be referenced by publications. Added an endpoint http://localhost:8082/anno/v1/fast/details/{link_name}/{link_value} that allow finding annotations tagged with a specific detail/association. For example:

http://ione.mbari.org:8100/anno/v1/fast/details/comment/Nature20190609559

Assets 2

18 Sep 23:55

hohonuuli

0.4.0

0a11fa5

annosaurus 0.4.0

Add a /windowrequest endpoint to find annotations in a set of videos near an existing one. The expected workflow to use this endpoint is as follows:

1. Fetch videos of interest

In general, these are videos from the same camera deployment. You can retrieve these from vampire-squid using http://myserver.org/vam/v1/media/concurrent/{video_reference_uuid}. This will return an array of media objects that overlap with the one you provided (via it's video_reference_uuid). For example:

[
  {
    "video_sequence_uuid": "28d1c03e-723e-4d78-9705-9b973a768c66",
    "video_reference_uuid": "e3005de7-2c0a-4a3f-80ad-79e5e79ee4b8",
    "video_uuid": "db0c3d6a-9849-4ff4-bc18-fd134337d114",
    "video_sequence_name": "Ventana 3009",
    "camera_id": "Ventana",
    "video_name": "V3009-01",
    "uri": "urn:tid:mbari.org:V3009-01",
    "start_timestamp": "2007-05-25T15:28:39Z",
    "duration_millis": 3600000,
    "container": "tape",
    "width": 1920,
    "height": 1080,
    "frame_rate": 29.97,
    "size_bytes": 0,
    "description": "Tape loaded from VARS on 2019-01-09T16:30:23.971Z. 86/86 valid dates"
  }
]

2. Use windowrequest

You will need:

The video_reference_uuids of the videos you want to query from
The imaged_moment_uuid of the existing annotation, it's recordedTimestamp will be the center of the time window
the window width in milliseconds

POST /anno/v1/imagedmoments/windowrequest
Content-Type: application/json

{
  video_reference_uuids: [
    "e3005de7-2c0a-4a3f-80ad-79e5e79ee4b8",
    "9be0daeb-d303-46a9-b3fa-699d7d40591e"
  ],
  imaged_moment_uuid: "0557b59b-60d5-44d0-acb3-71bf4acddf0a",
  window: 10000
}

Assets 2

10 Sep 18:12

hohonuuli

0.3.2

33e7011

annosaurus 0.3.2

Added a fast endpoint for fetching images for a given video_reference_uuid. The endpoint is:

http://my.server.net/anno/v1/fast/images/videoreference/<video_reference_uuid>

It will return JSON image data structured likes so:

[
  {
    "image_reference_uuid": "63dfcfa7-ea7c-4809-8435-89d3bae4bd64",
    "format": "image/jpg",
    "width": 1920,
    "height": 1080,
    "url": "http://search.mbari.org/ARCHIVE/frameGrabs/Ventana/images/3812/00_50_10_19.jpg",
    "description": "compressed image with overlay",
    "video_reference_uuid": "c101ecb4-d22d-4f5c-aa9f-1d1048643086",
    "imaged_moment_uuid": "57be066a-6498-4858-ad5f-07ed4254d8ea",
    "timecode": "00:50:10:19",
    "recorded_timestamp": "2014-11-04T17:05:31Z"
  },
  {
    "image_reference_uuid": "3faed261-d002-418d-9921-fcf163a0191d",
    "format": "image/png",
    "width": 1920,
    "height": 1080,
    "url": "http://search.mbari.org/ARCHIVE/frameGrabs/Ventana/images/3812/00_50_10_19.png",
    "description": "source image",
    "video_reference_uuid": "c101ecb4-d22d-4f5c-aa9f-1d1048643086",
    "imaged_moment_uuid": "57be066a-6498-4858-ad5f-07ed4254d8ea",
    "timecode": "00:50:10:19",
    "recorded_timestamp": "2014-11-04T17:05:31Z"
  }
]

Assets 2

31 Jul 17:35

hohonuuli

0.2.14

1d20b6e

annosaurus 0.2.14

Added an endpoint for fast look up of imaged_moment_uuids by concept with images:

Example:

GET http://localhost:8082/anno/v1/fast/imagedmoments/concept/images/Nanomia%20bijuga

This call will return a list of imaged_moment_uuids that contain both an observation matching the given concept and an image. Useful for machine learning tools looking for images matching a given term. The results will be JSON like:

[
  "a6d073f5-2e8d-4d3b-a39f-0124aec5c1f7",
  "f47645d5-f5a8-41f1-87d2-024f2560d7b9",
  "ab7e53bc-dafe-4571-af65-0268be422b19",
  "e49f67eb-ac45-462a-8aed-02865f68dbb8",
  "137b6826-cd82-45ee-8f6e-0291769709f2",
  "bcd8aaa0-64fa-4970-9d38-032b5bc9573d",
  "636cb0c7-8fa9-4143-92ee-041ca70274e1",
  "1fdee1e2-7a3a-48ec-824d-0441a9fcc8b2",
  "3be58b2a-3fe6-4bc8-a078-04cf0d456ae8",
  "aa6c6b47-175a-4655-bae9-0612dca91d94",
  "2707eda5-7080-40e3-b85e-073541c087a1",
  "e8a0b3a0-8932-489f-84d4-0794eb6bc4da",
  "c1f853a3-297f-4652-9d01-07c5b2bdd616"
]

Assets 2

30 Jul 20:19

hohonuuli

0.2.13

1c6bef0

annosaurus 0.2.13

This is a bug fix addressing a minor issue introduced in 0.2.11. A call to /fast/concept/:concept?limit=xx would not always return the correct number of rows. This has been fixed.

Assets 2

29 Jul 21:50

hohonuuli

0.2.12

4603cd4

annosaurus 0.2.12

Added JavaMelody for monitoring. It's linked to the path /monitoring. For example http://localhost:8082/anno/monitoring

Assets 2

29 Jul 19:02

hohonuuli

0.2.11

41a7e2f

annosaurus 0.2.11

Added fast endpoints for fetching only annotations with images.

Example queries

Description	URL Example
Find by concept with images	http://localhost:8082/anno/v1/fast/concept/images/Nanomia
Find by concept with images with limit and offset	http://localhost:8082/anno/v1/fast/concept/images/Nanomia?limit=20&offset=50
Count by concept (includes annotations without images)	http://localhost:8082/anno/v1/observations/concept/count/Nanomia
Count by concept with images	http://localhost:8082/anno/v1/observations/concept/images/count/Nanomia

Retrieving counts

Counting with or without images returns the same data structure. It contains the concept you requested and the count for that concept.

{"concept":"Nanomia", "count":"62"}

Assets 2

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

psenvsub.java

psenvsub.c

psenvsub.py

Create

Endpoint

Example

Returns

Delete

Endpoint

Example

Returns

annosaurus 0.4.0

1. Fetch videos of interest

2. Use windowrequest

Example queries

Retrieving counts

Releases: mbari-org/annosaurus

annosaurus 0.6.1

annosaurus 0.6.0

psenvsub.java

psenvsub.c

psenvsub.py

annosaurus 0.5.1

Create

Endpoint

Example

Returns

Delete

Endpoint

Example

Returns

annosaurus 0.5.0

annosaurus 0.4.0

annosaurus 0.4.0

1. Fetch videos of interest

2. Use windowrequest

annosaurus 0.3.2

annosaurus 0.2.14

annosaurus 0.2.13

annosaurus 0.2.12

annosaurus 0.2.11

Example queries

Retrieving counts