Centralize vision_opencv repository

[testkit]

When do ROS2 cv_bridge porting, the question about where to host it has been raised and discussed at #12, the final decision is to manage the code in https://github.com/ros-perception/vision_opencv ros2 branch.
Sorry to bring this question to the table again, since we run into a conflict issue when try to add cv_bridge to coming bouncy release. In vision_opencv, there are two key packages, cv_bridge and image_geometry, but they are maintained in ros-perception/vision_opencv and ros2/vision_opencv respectively, which brings difficulties to add these two packages into bloom release system.
For coming bouncy release, we'll copy ROS2 cv_bridge implementation to this project, but it's not an idea way to maintain the code in both repositories moving forward.

There are two options to get it resolved, I list them as below with the order of my preference, pls. comment.

• manage cv_bridge and image_geometry in this project, bouncy release from this project as well
• manage cv_bridge and image_geometry in ros-perception/vision_opencv ros2 branch, bouncy release from that project

@dirk-thomas @clalancette @gaoethan @mikaelarguedas @vrabaud.

[mikaelarguedas]

Thanks @testkit for bringing it up!

We discussed it internally a few weeks back and concluded that it should be centralized on https://github.com/ros-perception/vision_opencv as the ROS 2 fork was only temporary.
Ideally we would also keep the goal of "keeping the diff with the ROS1 version as small as possible", as this will allow for a lower maintenance burden and make patches trivially portable between the ROS1 and ROS2 versions.
This can be done in multiple ways:

• submit all changes not specific to ROS 2 (and backward compatible) to the ROS 1 branch
• avoid unnecessary changes if the ROS 1 maintainer is not willing to take them upstream (style changes for example).

I started the work of rebasing our changes of image_geometry on top of ros-perception/vision_opencv ros2 branch a few weeks back. I have not been able to use them on our CI system yet as:

• cv_bridge has compilation warnings
• the conversions.py test from cv_bridge fails on my machine

I opened ros-perception#219 with the image_geometry changes we currently use in ROS2.

Once that is merged, we can modify the ROS 2 CI to use the ros2 branch from https://github.com/ros-perception/vision_opencv and just skip the packages that we currently dont use and that are failings tests / having compiler warnings

[testkit]

@mikaelarguedas, that's great to have this issue considered and the solution is on the way already!
@gaoethan will help to check the two issues you mentioned above for cv_bridge.

[gaoethan]

@mikaelarguedas thanks for your quick support, regarding the test of conversions.py you mentioned

the conversions.py test from cv_bridge fails on my machine

Generally, it contains 4 test cases in it and it takes relatively long time than the other to finish, because it's normal with my trial, could u please give your error log for more details ? thanks !

[mikaelarguedas]

@gaoethan below the console output of the test.
Tested on Ubuntu bionic after installing manually python3-numpy and python3-opencv

10: Test command: /usr/bin/python3 "-u" "/root/ros2_ws/install/ament_cmake_test/share/ament_cmake_test/cmake/run_test.py" "/root/ros2_overlay_ws/build/cv_bridge/test_results/cv_bridge/conversions.py.xunit.xml" "--package-name" "cv_bridge" "--output-file" "/root/ros2_overlay_ws/build/cv_bridge/ament_cmake_pytest/conversions.py.txt" "--command" "/usr/bin/python3" "-u" "-m" "pytest" "/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/test/conversions.py" "-o" "cache_dir=/root/ros2_overlay_ws/build/cv_bridge/test/ament_cmake_pytest/conversions.py/.cache" "--junit-xml=/root/ros2_overlay_ws/build/cv_bridge/test_results/cv_bridge/conversions.py.xunit.xml" "--junit-prefix=cv_bridge"
10: Test timeout computed to be: 600
10: -- run_test.py: invoking following command in '/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge':
10:  - /usr/bin/python3 -u -m pytest /root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/test/conversions.py -o cache_dir=/root/ros2_overlay_ws/build/cv_bridge/test/ament_cmake_pytest/conversions.py/.cache --junit-xml=/root/ros2_overlay_ws/build/cv_bridge/test_results/cv_bridge/conversions.py.xunit.xml --junit-prefix=cv_bridge
10: ============================= test session starts ==============================
10: platform linux -- Python 3.6.5, pytest-3.6.0, py-1.5.3, pluggy-0.6.0
10: rootdir: /root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge, inifile:
10: plugins: rerunfailures-4.1, repeat-0.4.1, cov-2.5.1
10: collected 4 items
10: 
10: test/conversions.py .F..                                                 [100%]
10: 
10: =================================== FAILURES ===================================
10: ______________ TestConversions.test_encode_decode_cv2_compressed _______________
10: 
10: self = <conversions.TestConversions testMethod=test_encode_decode_cv2_compressed>
10: 
10:     def test_encode_decode_cv2_compressed(self):
10:         import numpy as np
10:         # from:
10:         # http://docs.opencv.org/2.4/modules/highgui/doc/reading_and_writing_images_and_video.html#Mat
10:         # imread(const string& filename, int flags)
10:         formats = ['jpg', 'jpeg', 'jpe', 'png', 'bmp', 'dib', 'ppm', 'pgm', 'pbm',
10:                    'jp2', 'sr', 'ras', 'tif', 'tiff']  # this formats rviz is not support
10:     
10:         cvb_en = CvBridge()
10:         cvb_de = CvBridge()
10:     
10:         for w in range(100, 800, 100):
10:             for h in range(100, 800, 100):
10:                 for f in formats:
10:                     for channels in ([], 1, 3):
10:                         if channels == []:
10:                             original = np.uint8(np.random.randint(0, 255, size=(h, w)))
10:                         else:
10:                             original = np.uint8(np.random.randint(0, 255, size=(h, w, channels)))
10: >                       compress_rosmsg = cvb_en.cv2_to_compressed_imgmsg(original, f)
10: 
10: test/conversions.py:78: 
10: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
10: 
10: self = <cv_bridge.core.CvBridge object at 0x7f9e8e836d68>
10: cvim = array([[215, 148, 142, ..., 116, 162, 242],
10:        [242,  60, 244, ...,  84, 114, 155],
10:        [ 50, 234,  48, ..., 16..., 147, 249, 159],
10:        [186, 233, 250, ..., 170, 116, 158],
10:        [157,  38,  84, ..., 184, 144, 137]], dtype=uint8)
10: dst_format = 'jp2'
10: 
10:     def cv2_to_compressed_imgmsg(self, cvim, dst_format='jpg'):
10:         """
10:             Convert an OpenCV :cpp:type:`cv::Mat` type to a ROS sensor_msgs::CompressedImage message.
10:     
10:             :param cvim:      An OpenCV :cpp:type:`cv::Mat`
10:             :param dst_format:  The format of the image data, one of the following strings:
10:     
10:             http://docs.opencv.org/2.4/modules/highgui/doc/reading_and_writing_images_and_video.html
10:             http://docs.opencv.org/2.4/modules/highgui/doc/reading_and_writing_images_and_video.html#Mat
10:             * imread(const string& filename, int flags)
10:                * bmp, dib
10:                * jpeg, jpg, jpe
10:                * jp2
10:                * png
10:                * pbm, pgm, ppm
10:                * sr, ras
10:                * tiff, tif
10:     
10:             :rtype:           A sensor_msgs.msg.CompressedImage message
10:             :raises CvBridgeError: when the ``cvim`` has a type that is incompatible with ``format``
10:     
10:     
10:             This function returns a sensor_msgs::Image message on success,
10:             or raises :exc:`cv_bridge.CvBridgeError` on failure.
10:             """
10:         import cv2
10:         import numpy as np
10:         if not isinstance(cvim, (np.ndarray, np.generic)):
10:             raise TypeError('Your input type is not a numpy array')
10:         cmprs_img_msg = sensor_msgs.msg.CompressedImage()
10:         cmprs_img_msg.format = dst_format
10:         ext_format = '.' + dst_format
10:         try:
10: >           cmprs_img_msg.data = np.array(cv2.imencode(ext_format, cvim)[1]).tostring()
10: E           cv2.error: /build/opencv-zcaJjh/opencv-3.2.0+dfsg/modules/imgcodecs/src/loadsave.cpp:682: error: (-2) could not find encoder for the specified extension in function imencode
10: 
10: ../../../../install/cv_bridge/lib/python3.6/site-packages/cv_bridge/core.py:240: error
10: ----------------------------- Captured stderr call -----------------------------
10: OpenCV Error: Unspecified error (could not find encoder for the specified extension) in imencode, file /build/opencv-zcaJjh/opencv-3.2.0+dfsg/modules/imgcodecs/src/loadsave.cpp, line 682
10:  generated xml file: /root/ros2_overlay_ws/build/cv_bridge/test_results/cv_bridge/conversions.py.xunit.xml 
10: ===================== 1 failed, 3 passed in 147.82 seconds =====================
10: -- run_test.py: return code 1
10: -- run_test.py: verify result file '/root/ros2_overlay_ws/build/cv_bridge/test_results/cv_bridge/conversions.py.xunit.xml'
10/11 Test #10: conversions.py ...................***Failed  148.77 sec

Build warnings:

CMake Warning at /usr/share/cmake-3.10/Modules/FindBoost.cmake:1626 (message):
  No header defined for python3; skipping header check
Call Stack (most recent call first):
  CMakeLists.txt:20 (find_package)


/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/cv_bridge.cpp: In function ‘cv_bridge::CvImageConstPtr cv_bridge::cvtColorForDisplay(const CvImageConstPtr&, const string&, cv_bridge::CvtColorForDisplayOptions)’:
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/cv_bridge.cpp:631:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t j = 0; j < source->image.rows; ++j) {
                        ~~^~~~~~~~~~~~~~~~~~~~
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/cv_bridge.cpp:632:28: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
       for (size_t i = 0; i < source->image.cols; ++i) {
                          ~~^~~~~~~~~~~~~~~~~~~~
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/cv_bridge.cpp:684:30: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
         for (size_t j = 0; j < source->image.rows; ++j) {
                            ~~^~~~~~~~~~~~~~~~~~~~
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/cv_bridge.cpp:685:32: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
           for (size_t i = 0; i < source->image.cols; ++i) {
                              ~~^~~~~~~~~~~~~~~~~~~~
In file included from /root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module.cpp:50:0:
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module.hpp: In function ‘int do_numpy_import()’:
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module.hpp:39:3: warning: converting to non-pointer type ‘int’ from NULL [-Wconversion-null]
   import_array();
   ^~~~~~~~~~~~
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module.hpp:40:1: warning: control reaches end of non-void function [-Wreturn-type]
 }
 ^
In file included from /root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module_opencv3.cpp:18:0:
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module.hpp: In function ‘int do_numpy_import()’:
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module.hpp:39:3: warning: converting to non-pointer type ‘int’ from NULL [-Wconversion-null]
   import_array();
   ^~~~~~~~~~~~
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module.hpp:40:1: warning: control reaches end of non-void function [-Wreturn-type]
 }
 ^
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module_opencv3.cpp: At global scope:
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/src/module_opencv3.cpp:97:19: warning: ‘PyObject* failmsgp(const char*, ...)’ defined but not used [-Wunused-function]
 static PyObject * failmsgp(const char * fmt, ...)
                   ^~~~~~~~
In file included from /root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/test/utest.cpp:16:0:
/root/ros2_ws/install/gtest_vendor/src/gtest_vendor/include/gtest/gtest.h: In instantiation of ‘testing::AssertionResult testing::internal::CmpHelperEQ(const char*, const char*, const T1&, const T2&) [with T1 = unsigned int; T2 = int]’:
/root/ros2_ws/install/gtest_vendor/src/gtest_vendor/include/gtest/gtest.h:1429:23:   required from ‘static testing::AssertionResult testing::internal::EqHelper<lhs_is_null_literal>::Compare(const char*, const char*, const T1&, const T2&) [with T1 = unsigned int; T2 = int; bool lhs_is_null_literal = false]’
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/test/utest.cpp:32:3:   required from here
/root/ros2_ws/install/gtest_vendor/src/gtest_vendor/include/gtest/gtest.h:1401:11: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
   if (lhs == rhs) {
       ~~~~^~~~~~
/root/ros2_ws/install/gtest_vendor/src/gtest_vendor/include/gtest/gtest.h: In instantiation of ‘testing::AssertionResult testing::internal::CmpHelperEQ(const char*, const char*, const T1&, const T2&) [with T1 = int; T2 = long unsigned int]’:
/root/ros2_ws/install/gtest_vendor/src/gtest_vendor/include/gtest/gtest.h:1429:23:   required from ‘static testing::AssertionResult testing::internal::EqHelper<lhs_is_null_literal>::Compare(const char*, const char*, const T1&, const T2&) [with T1 = int; T2 = long unsigned int; bool lhs_is_null_literal = false]’
/root/ros2_overlay_ws/src/ros2/vision_opencv/cv_bridge/test/utest.cpp:106:3:   required from here
/root/ros2_ws/install/gtest_vendor/src/gtest_vendor/include/gtest/gtest.h:1401:11: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]

[mikaelarguedas]

PRs to migrate ROS2 CI to use the https://github.com/ros-perception/vision_opencv repository instead of the ros2 fork:
ros2/ci#180
ros2/turtlebot2_demo#86

@testkit @gaoethan please speak up if there is any reason to hold this, otherwise I'll go ahead and merge these tomorrow

[gaoethan]

@mikaelarguedas thanks for your detail. I create an extra issue ros-perception#221 to track and fix the warning and test failure for you to avoid messy topic here, next I'll investigate and fix it there accordingly.

[mikaelarguedas]

@testkit @gaoethan I'm closing this ticket as I believe the original question has been answered.
If there is any item we forgot to address, please feel free to comment here and we can reopen it