Issue with output when topics specified with preceding "/"

[nlitz88]

Hi there!

I have a rosbag with a few image topics that I am trying to export. I was able to build your Docker image, get a workspace built with your package, and then run the utility to export from a ros bag with the bag_to_image launch file. However, there seems to be a problem with the output filepath of each image file.

Specifically, I believe the issue is that this node assumes that each of the messages you are trying to export is "stamped"--I.e., contains a header with a timestamp. For my particular use case, these image messages don't have their timestamps set properly (yet).

My particular use case and output looks like this, for reference:

ros2 launch ros2_bag_to_image bag_to_image.xml input/path:=/home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2/workzone_boundary_detection_pipeline_images_2_0.db3 input/topics:="['/result_image', '/boundary_image']" output/path:=/home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png/

And the output that follows looks like this:

[INFO] [launch]: All log files can be found below /home/nlitz88/.ros/log/2024-04-22-20-21-43-930541-b683bef1f499-481
[INFO] [launch]: Default logging verbosity is set to INFO
[INFO] [bag_to_image_node-1]: process started with pid [482]
[bag_to_image_node-1] [INFO] [1713831704.112200246] [bag_to_image_node]: Using bag format: 'cdr'
[bag_to_image_node-1] [INFO] [1713831704.112279615] [bag_to_image_node]: Using bag storage_id: 'sqlite3'
[bag_to_image_node-1] [INFO] [1713831704.112400941] [bag_to_image_node]: Saving Images PNGs to:/home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png/
[bag_to_image_node-1] [INFO] [1713831704.112415748] [bag_to_image_node]: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2/workzone_boundary_detection_pipeline_images_2_0.db3
[bag_to_image_node-1] [INFO] [1713831704.112421842] [bag_to_image_node]: sqlite3
[bag_to_image_node-1] [INFO] [1713831704.112427711] [bag_to_image_node]: cdr
[bag_to_image_node-1] [INFO] [1713831704.112433155] [bag_to_image_node]: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png/
[bag_to_image_node-1] [INFO] [1713831704.112444970] [bag_to_image_node]: /result_image
[bag_to_image_node-1] [INFO] [1713831704.112878970] [bag_to_image_node]: /boundary_image
[bag_to_image_node-1] created: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png/result_image
[bag_to_image_node-1] created: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png/boundary_image
[bag_to_image_node-1] [INFO] [1713831704.115028613] [rosbag2_storage]: Opened database '/home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2/workzone_boundary_detection_pipeline_images_2_0.db3' for READ_ONLY.
[bag_to_image_node-1] [INFO] [1713831704.129542774] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.137748402] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.145742231] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.152676010] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.159884524] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.167009065] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.174369170] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.181280828] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.188460978] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.195304223] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.201408062] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.208343054] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.215605250] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.222662456] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.229948645] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.236998884] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.244346854] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.251422649] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.258960605] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.266249477] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.273581719] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//result_image_0.0.png
[bag_to_image_node-1] [INFO] [1713831704.280489216] [bag_to_image_node]: Image saved to: /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png//boundary_image_0.0.png

...and so on until it finishes. However, in the output directory, I am only left with this:

nlitz88@b683bef1f499:/workspaces/isaac_ros-dev/18744/rosbag_export_ws$ ls -l /home/nlitz88/Documents/workzone_boundary_detection_pipeline_images_2_png/
total 36
drwxr-xr-x 2 nlitz88 nlitz88  4096 Apr 22 20:19 boundary_image
-rw-r--r-- 1 nlitz88 nlitz88  5546 Apr 22 20:21 boundary_image_0.0.png
drwxr-xr-x 2 nlitz88 nlitz88  4096 Apr 22 20:19 result_image
-rw-r--r-- 1 nlitz88 nlitz88 19612 Apr 22 20:21 result_image_0.0.png

where the directories boundary_image and result_image are empty.

I added a simple workaround/fix in place of this code where I just create a static counter variable and name the images accordingly like this:

static size_t message_count = 0;
fname = output_path_ + "/" + target_topic_name + "_" +
boost::lexical_cast<std::string>(message_count) + ".png";
message_count += 1;

Might not be worth "fixing" depending on your needs--particularly because my image messages should have valid timestamp values--but this is an edge case where the source of these messages isn't doing something right and this is all I have at the moment. Just wanted to leave this here in case someone else runs into it.

[amc-nu]

I'm sorry for not getting back to you sooner. I haven't read the message until now.
Can you check the update code by @IshitaTakeshi on #4
He did a great job standardizing the file naming.

[nlitz88]

Thanks for getting back.

I think #4 definitely fixes the output formatting, but I still don't think it would resolve the issue I was encountering above. The updates he lists in #4

• The number of digits in the timestamp is now fixed for consistency.
• Output files are now stored in separate directories based on the topic.

If I am remembering correctly, my problem was that all my messages on a given topic didn't have valid timestamps. That is, I'm pretty sure my image messages didn't have a header field. So, even with the formatting update in #4 , I still think all the output files would get saved with the same suffix "0.0" (where I'm assuming those variables are what got initialized to zero).

To resolve the issue above, maybe a simple solution would be to add a command line argument that allows the user to specify whether or not the topic they're running this on has messages with headers. If not, then maybe use a counter to incrementally name each message like that little snippet I posted above might.

TBH, though: Sorry I didn't re-run the tool with those changes. I don't think I even have that data anymore 😆