ERROR bad callback: <PyCall.jlwrap jl_cb> when trying to use with sensor_msgs/Image

[nstiurca]

I tried modifying the example code to use sensor_msgs/Image, but I am getting errors when trying to run, and the callback isn't called. Any thoughts?

---

Source code:

#!/usr/bin/env julia

using RobotOS
@rosimport sensor_msgs.msg.Image
rostypegen()
using sensor_msgs.msg

function callback(img::Image, pub_obj::Publisher{Image})
    println(img.header)
    publish(pub_obj, img)
end

function loop(pub_obj)
    loop_rate = Rate(5.0)
    while ! is_shutdown()
#         npt = Point(rand(), rand(), 0.0)
#         publish(pub_obj, npt)
        rossleep(loop_rate)
    end
end

function main()
    init_node("inpaint_depth")
    pub = Publisher{Image}("depth_inpainted", queue_size=10)
    sub = Subscriber{Image}("depth", callback, (pub,), queue_size=10)
    loop(pub)
end

if ! isinteractive()
    main()
end

---

Error:

$ ./inpaint_depth.jl depth:=/camera/depth_registered/image_raw
[ERROR] [WallTime: 1447366470.036321] bad callback: <PyCall.jlwrap jl_cb>
Traceback (most recent call last):
  File "/opt/ros/indigo/lib/python2.7/dist-packages/rospy/topics.py", line 711, in _invoke_callback
    cb(msg)
TypeError: an integer is required

[ERROR] [WallTime: 1447366470.307730] bad callback: <PyCall.jlwrap jl_cb>
Traceback (most recent call last):
  File "/opt/ros/indigo/lib/python2.7/dist-packages/rospy/topics.py", line 711, in _invoke_callback
    cb(msg)
TypeError: an integer is required

[ERROR] [WallTime: 1447366470.869406] bad callback: <PyCall.jlwrap jl_cb>
Traceback (most recent call last):
  File "/opt/ros/indigo/lib/python2.7/dist-packages/rospy/topics.py", line 711, in _invoke_callback
    cb(msg)
TypeError: an integer is required

< repeats with every incoming Image >

[nstiurca]

In case it's relevant, I'm using Julia v0.4, Python v2.7, and ROS Indigo.

[jdlangs]

Thanks for the report. Could you make a small bag file of the image messages that are failing so I can be sure I'm exactly reproducing the problem?

[nstiurca]

https://www.assembla.com/spaces/cis700-snap/documents/be0R-MIAGr5zGHdmr6CpXy/download/be0R-MIAGr5zGHdmr6CpXy

[nstiurca]

$ rosbag info /tmp/2015-11-12-20-43-40.bag 
path:         /tmp/2015-11-12-20-43-40.bag
version:      2.0
duration:     5.3s
start:        Nov 12 2015 20:43:41.27 (1447379021.27)
end:          Nov 12 2015 20:43:46.57 (1447379026.57)
size:         1.9 MB
messages:     55
compression:  bz2 [28/28 chunks; 5.98%]
uncompressed: 32.2 MB @   6.1 MB/s
compressed:    1.9 MB @ 372.1 KB/s (5.98%)
types:        sensor_msgs/Image [060021388200f6f0f447d0fcd9c64743]
topics:       /camera/depth_registered/image_raw   55 msgs    : sensor_msgs/Image

[jdlangs]

Thanks, I'll see if I have some time to try a fix by this weekend.

[jdlangs]

It turned out to be an upstream issue with PyCall. Specifically, it didn't know how to transparently convert a python string (rospy's byte storage method) to Julia's Vector{UInt8}. That is fixed now. Try running Pkg.checkout(PyCall) and see if it works.

I'll also add some Image tests since apparently I didn't do a thorough enough job in manual testing before.

[nstiurca]

Much better, but still a strange error:

$ julia inpaint_depth.jl depth:=/camera/depth_registered/image_raw
std_msgs.msg[ERROR] [WallTime: 1447462026.962798] bad callback: <PyCall.jlwrap jl_cb>
Traceback (most recent call last):
  File "/opt/ros/indigo/lib/python2.7/dist-packages/rospy/topics.py", line 711, in _invoke_callback
    cb(msg)
MemoryError: Julia exception: OutOfMemoryError()

std_msgs.msg.Header(0x0000002a,RobotOS.Time(1447379021,246340814),"camera_rgb_optical_frame")
std_msgs.msg.Header(0x0000002c,RobotOS.Time(1447379021,385057292),"camera_rgb_optical_frame")
std_msgs.msg.Header(0x0000002e,RobotOS.Time(1447379021,448288826),"camera_rgb_optical_frame")
std_msgs.msg.Header(0x00000030,RobotOS.Time(1447379021,613769898),"camera_rgb_optical_frame")
std_msgs.msg.Header(0x00000031,RobotOS.Time(1447379021,648030688),"camera_rgb_optical_frame")
std_msgs.msg.Header(0x00000032,RobotOS.Time(1447379021,748913319),"camera_rgb_optical_frame")
std_msgs.msg.Header(0x00000033,RobotOS.Time(1447379021,848617210),"camera_rgb_optical_frame")
std_msgs.msg.Header(0x00000035,RobotOS.Time(1447379021,984761540),"camera_rgb_optical_frame")

[nstiurca]

Ie, the first callback fails, but the rest are fine. Race condition?

[jdlangs]

I have not ever seen an OutOfMemoryError before. Is it always repeatable? Are you running it on a somewhat resource-limited system?

[nstiurca]

Always repeatable on a sturdy gaming laptop. Is it not reproducible for you? There might be some other package on my system needing updating to master branch or such.

For the record,

julia> versioninfo()
Julia Version 0.4.0
Commit 0ff703b* (2015-10-08 06:20 UTC)
Platform Info:
  System: Linux (x86_64-linux-gnu)
  CPU: Intel(R) Core(TM) i7-4710HQ CPU @ 2.50GHz
  WORD_SIZE: 64
  BLAS: libopenblas (NO_LAPACK NO_LAPACKE DYNAMIC_ARCH NO_AFFINITY Haswell)
  LAPACK: liblapack.so.3
  LIBM: libopenlibm
  LLVM: libLLVM-3.3

julia> Pkg.status()
11 required packages:
 - CUDA                          0.1.0
 - Color                         0.2.11
 - Debug                         0.1.6
 - FactCheck                     0.4.1
 - IJulia                        1.1.8
 - ImageView                     0.1.18
 - Images                        0.5.0
 - Jewel                         1.0.7
 - Mocha                         0.1.0+             opencl
 - Quaternions                   0.0.4
 - RobotOS                       0.3.5
42 additional packages:
 - BinDeps                       0.3.19
 - Blosc                         0.1.4
 - CLBLAS                        0.1.0+             deprecated-release-fix
 - Cairo                         0.2.31
 - Calculus                      0.1.14
 - ColorTypes                    0.2.0
 - ColorVectorSpace              0.1.0
 - Colors                        0.6.0
 - Compat                        0.7.7
 - Compose                       0.3.18
 - Conda                         0.1.8
 - DataStructures                0.3.13
 - Dates                         0.4.4
 - Docile                        0.5.19
 - DualNumbers                   0.1.5
 - FileIO                        0.0.3
 - FixedPointNumbers             0.1.1
 - Graphics                      0.1.3
 - HDF5                          0.5.6
 - IniFile                       0.2.4
 - Iterators                     0.1.9
 - JLD                           0.5.6
 - JSON                          0.5.0
 - JuliaParser                   0.6.3
 - LNR                           0.0.2
 - Lazy                          0.10.1
 - Logging                       0.2.0
 - MacroTools                    0.2.0
 - NaNMath                       0.1.1
 - Nettle                        0.2.0
 - OpenCL                        0.3.6+             master
 - PyCall                        1.2.0+             master
 - Reexport                      0.0.3
 - Requires                      0.2.1
 - SHA                           0.1.2
 - SIUnits                       0.0.6
 - TexExtensions                 0.0.3
 - Tk                            0.3.6
 - URIParser                     0.1.1
 - Winston                       0.11.13
 - ZMQ                           0.3.0
 - Zlib                          0.1.12

[jdlangs]

No, all the messages in the bag file are received fine by the callback on my end and I tried it on several different julia versions (0.4.0, 0.4.1, 0.5). There's probably a bug somewhere but it seems to be platform-specific. If you can reproduce it on another machine that would be good to know.

[nstiurca]

Ok, good to know. Since it seems to be specific to my setup, I will try to
look into this tonight.

-- Nicu Știurcă
//./ Live long and prosper.

On Mon, Nov 16, 2015 at 1:04 PM, Josh Langsfeld [email protected]
wrote:

No, all the messages in the bag file are received fine by the callback on
my end and I tried it on several different julia versions (0.4.0, 0.4.1,
0.5). There's probably a bug somewhere but it seems to be
platform-specific. If you can reproduce it on another machine that would be
good to know.

—
Reply to this email directly or view it on GitHub
#9 (comment).