René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Slow performance on python wrapper","articleBody":"Let's start by considering this snippet:\r\n\r\n    import time\r\n    from pathlib import Path\r\n\r\n    import tinyobjloader\r\n\r\n\r\n    def load_obj(filename):\r\n        path = Path(filename).resolve()\r\n        start = time.time()\r\n        reader = tinyobjloader.ObjReader()\r\n        if not reader.ParseFromFile(str(path)):\r\n            raise Exception(f\"Problem loading '{path}'\")\r\n        attrib = reader.GetAttrib()\r\n        start = time.time()\r\n        num_vertices = 0\r\n        num_normals = 0\r\n        num_texcoords = 0\r\n\r\n        for shape in reader.GetShapes():\r\n            num_vertices += len(attrib.vertices)\r\n            num_normals += len(attrib.normals)\r\n            num_texcoords += len(attrib.texcoords)\r\n\r\n            for f in shape.mesh.indices:\r\n                attrib.vertices[f.vertex_index * 3 + 0]\r\n                attrib.vertices[f.vertex_index * 3 + 1]\r\n                attrib.vertices[f.vertex_index * 3 + 2]\r\n                attrib.normals[f.normal_index * 3 + 0]\r\n                attrib.normals[f.normal_index * 3 + 1]\r\n                attrib.normals[f.normal_index * 3 + 2]\r\n                attrib.texcoords[f.texcoord_index * 2 + 0]\r\n                attrib.texcoords[f.texcoord_index * 2 + 1]\r\n\r\n        print(\r\n            f\"num_vertices={num_vertices} num_normals={num_normals} num_texcoords={num_texcoords}\"\r\n        )\r\n        print(time.time() - start)\r\n\r\n\r\n    if __name__ == \"__main__\":\r\n        load_obj(r\"mcve.obj\")\r\n\r\nAnd here's the link to both [mcve.obj](https://www.dropbox.com/s/vslkrmgcczem7bh/mcve.obj?dl=0) and [mcve.mtl](https://www.dropbox.com/s/46rhn31wngtub5y/mcve.mtl?dl=1), this object is a very small \u0026 trivial scene, that looks like this:\r\n\r\n![showcase](https://dl.dropboxusercontent.com/s/xi9dyv8roc7fjn6/blender_2020-05-26_19-41-34.png)\r\n\r\nProblem being... if you run that snippet you'll get this output:\r\n\r\n    num_vertices=36150 num_normals=38400 num_texcoords=27950\r\n    13.565775871276855\r\n\r\nThat's really slow \u0026 unusable for such a trivial scene. I wanted to san miguel's scene from [https://casual-effects.com/data/](https://casual-effects.com/data/) but after seeing these poor results on such a trivial scene I'm not even considering it till I figure out how to proceed here.\r\n\r\nSo yeah, the main question would be, how could we use the tinyobjloader effectively on python so we can test real-world scenes?\r\n\r\nThanks in advance.\r\n\r\nPs. I've got allocated a huge chunk of memory for interleaved data (vertices+normals+texcoords) so I'm not sure how difficult would be providing a wrapper to make a fast memcpy or similar... Main problem seems to be the whole overhead of wrapping c++ to  python","author":{"url":"https://github.com/brupelo","@type":"Person","name":"brupelo"},"datePublished":"2020-05-26T17:50:08.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":8},"url":"https://github.com/275/tinyobjloader/issues/275"}