{ "info": { "author": "qiaowang", "author_email": "adawq0@gmail.com", "bugtrack_url": null, "classifiers": [ "License :: OSI Approved :: MIT License", "Operating System :: OS Independent", "Programming Language :: Python :: 3" ], "description": "# commented out code\n\nThis library can detect both inline or block commented out code.\n\nIt sends rpc to the server which uses LSTM model to predict commented out code.\nThe server runs on my school data center, and will be migrated to a private machine.\n\nOnly support c/c++ now. More language support is about to come\n\n# How to use\n\n## sample code\n\n```\ntext = '''\n void DropoutLayer::updateB(){\n\n int num = this->numUnit;\n //cout<<\"Error!!!!!!!!!!!!!!!!in DropoutLayer!!!!!!!\"<z == NULL){ // embeddings\n // gradB += dE_dy, because y = b\n iXpY( num , this->dE_dy, gradBiases + bidx );\n\n return;\n }\n\n if (fprime != dummy){\n\n // dy_dz = f', evaluated at y\n ( * this->fprime)(this->y, this->dy_dz, num);\n // dE_dz = dE_dy .* dy_dz\n\n pointwise_dot(this->dE_dy, this->dy_dz, this->dE_dz, num);\n //cout<<\"dropout backward\"<dE_dz[i] *= this->indicator[i];\n\n }\n\n }// else if fprime == softmaxprime{\n // do nothing, because we assume dE_dz is given by softmax\n\n //}\n\n //\t\tReLUPrime(float * y, float * dy_dz, int n);\n }\n'''\nprint(client.search(text, 'cpp')) # return OrderedDict object, key is line_number, value is the commented code itself\n\n# output\n# OrderedDict([(4, 'cout<<\"Error!!!!!!!!!!!!!!!!in DropoutLayer!!!!!!!\"<