# A non regression test for the transform stack Feature: transformstack Scenario Outline: mtransformstack When I launch videostitch-cmd with transformstack/<ptv>/<ptv>.ptv and "-d 0 -f 0 -l 0" Then I expect the command to succeed When I compare transformstack/<ptv>/<ptv>_output-0.jpg with transformstack/<ptv>/reference-1.jpg Then I expect the comparison error to be less than 0.03 Examples: | ptv | | circular | | elmo | @slow Scenario Outline: projections When I launch videostitch-cmd with transformstack/regression/<in>_<out>.ptv and "-d 0 -l 0" Then I expect the command to succeed When I compare transformstack/regression/<in>_<out>-1.png with transformstack/regression/<in>_<out>-2-reference.png Then I expect the comparison error to be less than 0.03 Examples: | in | out | | cf | cf | | cf | erect | | cf | ff | | cf | rect | | cf | stereo | | erect | cf | | erect | erect | | erect | ff | | erect | rect | | erect | stereo | | ff | cf | | ff | erect | | ff | ff | | ff | rect | | ff | stereo | | rect | cf | | rect | erect | | rect | ff | | rect | rect | | rect | stereo | @slow Scenario Outline: pano parameters When I launch videostitch-cmd with transformstack/regression/<in>_<out>_<param>_<value1>_<value2>.ptv and "-d 0 -l 0" Then I expect the command to succeed When I compare transformstack/regression/<in>_<out>_<param>_<value1>_<value2>-1.png with transformstack/regression/<in>_<out>_<param>_<value1>_<value2>-2-reference.png Then I expect the comparison error to be less than 0.03 Examples: | in | out | param | value1 | value2 | | ff | erect | size | 64 | 128-v2 | | ff | erect | size | 256 | 512 | | ff | erect | size | 1024 | 2048 | | ff | erect | size | 2048 | 4096 | | ff | stereo | size | 128 | 1024 | | ff | stereo | size | 1024 | 1024 | | ff | erect | hfov | 2 | 0-v2 | | ff | erect | hfov | 8 | 0-v2 | | ff | erect | hfov | 45 | 0-v2 | | ff | erect | hfov | 90 | 0 | | ff | erect | hfov | 150 | 0 | | ff | erect | hfov | 220 | 0 | | ff | erect | hfov | 360 | 0 | | ff | stereo | hfov | 2 | 0-v2 | | ff | stereo | hfov | 8 | 0-v2 | | ff | stereo | hfov | 45 | 0-v2 | | ff | stereo | hfov | 90 | 0 | | ff | stereo | hfov | 150 | 0 | | ff | stereo | hfov | 220 | 0 | | ff | stereo | hfov | 350 | 0 | | ff | erect | crop | l500 | 0 | | ff | erect | pad | t500 | 0 | | ff | erect | pad | b500 | 0 | | ff | erect | pad | t100b100 | 0 | | ff | erect | yaw | 45 | 0 | | ff | erect | yaw | 90 | 0 | | ff | erect | yaw | 270 | 0 | | ff | erect | pitch | 45 | 0 | | ff | erect | pitch | 90 | 0 | | ff | erect | pitch | 270 | 0 | | ff | erect | roll | 45 | 0 | | ff | erect | roll | 90 | 0 | | ff | erect | roll | 270 | 0 | # | ff | erect | crop | r500 | 0 | # | ff | erect | crop | b500 | 0 | # | ff | erect | crop | t500 | 0 | # | ff | erect | crop | l200r200t200b200 | 0 | Scenario Outline: input parameters When I launch videostitch-cmd with transformstack/inputs/test_input_<setting>.ptv and "-d 0 -l 0" Then I expect the command to succeed When I compare transformstack/inputs/input_<setting>-1.png with transformstack/inputs/input_<setting>-1-reference-1.png Then I expect the comparison error to be less than 0.03 Examples: | setting | # VSA-4791 #| crop | | dist | | hfov | | roll | | width_height | Scenario Outline: projections board When I launch videostitch-cmd with transformstack/regression/<in>_<out>_board.ptv and "-d 0 -l 0" Then I expect the command to succeed When I compare transformstack/regression/<in>_<out>_board-1.png with transformstack/regression/<in>_<out>_board-1-reference-1.png Then I expect the comparison error to be less than 0.0004 Examples: | in | out | | erect | erect | Scenario Outline: Stitching identity accuracy When I launch videostitch-cmd with input_formats/<test>-<format>.ptv and "-d 0 -f 0 -l 0" Then I expect the command to succeed When I compare input_formats/<format>-out-0.<format> with input_formats/input.<format> Then I expect the comparison error to be less than <expected_error> Examples: | test | format | expected_error | | identity | jpg | 0.014 | | identity | png | 0.013 | | flip-unflip | png | 0.015 |