OpenI
/
graphkit-learn
Not watched
1
0
Fork 0
Code
Releases 2 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain
You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
71 Commits
4 Branches
345 MB
158 Download
Tree: 68772ba4b4
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '68772ba4b4'
${ noResults }
graphkit-learn/notebooks/run_spkernel.py

run_spkernel.py 9.8 kB
Raw Normal View History

1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. upgrate spkernel, now supporting symbolic and non-symbolic node labels, directed and undirected graphs, and allowing user-defined node kernels. 2. model_selection_precomputed can now save all results as human readable text. 3. modify pygraph.utils.utils.floydTransformation and pygraph.utils.graphdataset.get_dataset_attributes.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. upgrate spkernel, now supporting symbolic and non-symbolic node labels, directed and undirected graphs, and allowing user-defined node kernels. 2. model_selection_precomputed can now save all results as human readable text. 3. modify pygraph.utils.utils.floydTransformation and pygraph.utils.graphdataset.get_dataset_attributes.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
​
​
​
​
​
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
​
​
​
​
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
​
​
​
1. upgrate spkernel, now supporting symbolic and non-symbolic node labels, directed and undirected graphs, and allowing user-defined node kernels. 2. model_selection_precomputed can now save all results as human readable text. 3. modify pygraph.utils.utils.floydTransformation and pygraph.utils.graphdataset.get_dataset_attributes.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
1. [IMPORTANT] correct an error in dataset split section in function cross_validation_pre_computed. DON NOT use previous versions. They are all wrong. 2. save gram matrices and relative data when using function cross_validation_pre_computed, before cross validation step, in case that something goes wrong with CV. Parameter read_gm_from_file can be used to choose whether to read gram matrices from file. 3. add some test code to check if a gram matrix is symmetric and positive semi-definite.
6 years ago
​
​
​
1. add parallel computing scheme to spkernel and model_selection_precomputed. 2. modify model_selection_precomputed so that all results are written into memory and then to a file at last section of code, in case that on cpu/disk seperated systems the IO takes too much time. 3. correct utils.floyd_warshall_numpy function. DONNOT use the last version.
7 years ago
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163 
  1. # %load_ext line_profiler

  2. # %matplotlib inline

  3. import functools

  4. from libs import *

  5. import multiprocessing

  6. from sklearn.metrics.pairwise import rbf_kernel

  7. 

  8. from pygraph.kernels.spKernel import spkernel, spkernel_do

  9. from pygraph.utils.kernels import deltakernel, kernelproduct

  10. from pygraph.utils.model_selection_precomputed import trial_do

  11. 

  12. dslist = [

  13. #    {'name': 'Acyclic', 'dataset': '../datasets/acyclic/dataset_bps.ds',

  14. #        'task': 'regression'},  # node symb

  15. #    {'name': 'Alkane', 'dataset': '../datasets/Alkane/dataset.ds', 'task': 'regression',

  16. #             'dataset_y': '../datasets/Alkane/dataset_boiling_point_names.txt', },  # contains single node graph, node symb

  17. #    {'name': 'MAO', 'dataset': '../datasets/MAO/dataset.ds', },  # node/edge symb

  18. #    {'name': 'PAH', 'dataset': '../datasets/PAH/dataset.ds', },  # unlabeled

  19. #    {'name': 'MUTAG', 'dataset': '../datasets/MUTAG/MUTAG.mat',

  20. #             'extra_params': {'am_sp_al_nl_el': [0, 0, 3, 1, 2]}},  # node/edge symb

  21.     {'name': 'Letter-med', 'dataset': '../datasets/Letter-med/Letter-med_A.txt'},

  22. #    # node symb/nsymb

  23. #    {'name': 'ENZYMES', 'dataset': '../datasets/ENZYMES_txt/ENZYMES_A_sparse.txt'},

  24. #    # node/edge symb

  25. #    {'name': 'Mutagenicity', 'dataset': '../datasets/Mutagenicity/Mutagenicity_A.txt'},

  26. #    {'name': 'D&D', 'dataset': '../datasets/D&D/DD.mat',

  27. #     'extra_params': {'am_sp_al_nl_el': [0, 1, 2, 1, -1]}},  # node symb

  28. 

  29.     #     {'name': 'COIL-DEL', 'dataset': '../datasets/COIL-DEL/COIL-DEL_A.txt'}, # edge symb, node nsymb

  30.     # # #     {'name': 'BZR', 'dataset': '../datasets/BZR_txt/BZR_A_sparse.txt'}, # node symb/nsymb

  31.     # # #     {'name': 'COX2', 'dataset': '../datasets/COX2_txt/COX2_A_sparse.txt'}, # node symb/nsymb

  32.     #     {'name': 'Fingerprint', 'dataset': '../datasets/Fingerprint/Fingerprint_A.txt'},

  33.     #

  34.     # #     {'name': 'DHFR', 'dataset': '../datasets/DHFR_txt/DHFR_A_sparse.txt'}, # node symb/nsymb

  35.     # #     {'name': 'SYNTHETIC', 'dataset': '../datasets/SYNTHETIC_txt/SYNTHETIC_A_sparse.txt'}, # node symb/nsymb

  36.     # #     {'name': 'MSRC9', 'dataset': '../datasets/MSRC_9_txt/MSRC_9_A.txt'}, # node symb

  37.     # #     {'name': 'MSRC21', 'dataset': '../datasets/MSRC_21_txt/MSRC_21_A.txt'}, # node symb

  38.     # #     {'name': 'FIRSTMM_DB', 'dataset': '../datasets/FIRSTMM_DB/FIRSTMM_DB_A.txt'}, # node symb/nsymb ,edge nsymb

  39. 

  40.     # #     {'name': 'PROTEINS', 'dataset': '../datasets/PROTEINS_txt/PROTEINS_A_sparse.txt'}, # node symb/nsymb

  41.     # #     {'name': 'PROTEINS_full', 'dataset': '../datasets/PROTEINS_full_txt/PROTEINS_full_A_sparse.txt'}, # node symb/nsymb

  42.     # #     {'name': 'AIDS', 'dataset': '../datasets/AIDS/AIDS_A.txt'}, # node symb/nsymb, edge symb

  43.     #     {'name': 'NCI1', 'dataset': '../datasets/NCI1/NCI1.mat',

  44.     #         'extra_params': {'am_sp_al_nl_el': [1, 1, 2, 0, -1]}}, # node symb

  45.     #     {'name': 'NCI109', 'dataset': '../datasets/NCI109/NCI109.mat',

  46.     #         'extra_params': {'am_sp_al_nl_el': [1, 1, 2, 0, -1]}}, # node symb

  47.     #     {'name': 'NCI-HIV', 'dataset': '../datasets/NCI-HIV/AIDO99SD.sdf',

  48.     #         'dataset_y': '../datasets/NCI-HIV/aids_conc_may04.txt',}, # node/edge symb

  49. 

  50.     #     # not working below

  51.     #     {'name': 'PTC_FM', 'dataset': '../datasets/PTC/Train/FM.ds',},

  52.     #     {'name': 'PTC_FR', 'dataset': '../datasets/PTC/Train/FR.ds',},

  53.     #     {'name': 'PTC_MM', 'dataset': '../datasets/PTC/Train/MM.ds',},

  54.     #     {'name': 'PTC_MR', 'dataset': '../datasets/PTC/Train/MR.ds',},

  55. ]

  56. estimator = spkernel

  57. mixkernel = functools.partial(kernelproduct, deltakernel, rbf_kernel)

  58. param_grid_precomputed = {'node_kernels': [

  59.     {'symb': deltakernel, 'nsymb': rbf_kernel, 'mix': mixkernel}]}

  60. param_grid = [{'C': np.logspace(-10, 10, num=41, base=10)},

  61.               {'alpha': np.logspace(-10, 10, num=41, base=10)}]

  62. 

  63. for ds in dslist:

  64.     print()

  65.     print(ds['name'])

  66.     model_selection_for_precomputed_kernel(

  67.         ds['dataset'],

  68.         estimator,

  69.         param_grid_precomputed,

  70.         (param_grid[1] if ('task' in ds and ds['task']

  71.                            == 'regression') else param_grid[0]),

  72.         (ds['task'] if 'task' in ds else 'classification'),

  73.         NUM_TRIALS=30,

  74.         datafile_y=(ds['dataset_y'] if 'dataset_y' in ds else None),

  75.         extra_params=(ds['extra_params'] if 'extra_params' in ds else None),

  76.         ds_name=ds['name'],

  77.         n_jobs=multiprocessing.cpu_count(),

  78.         read_gm_from_file=False)

  79. 

  80. #     %lprun -f trial_do -f spkernel -f spkernel_do -f model_selection_for_precomputed_kernel \

  81. #         model_selection_for_precomputed_kernel( \

  82. #             ds['dataset'], \

  83. #             estimator, \

  84. #             param_grid_precomputed, \

  85. #             (param_grid[1] if ('task' in ds and ds['task'] == 'regression') else param_grid[0]), \

  86. #             (ds['task'] if 'task' in ds else 'classification'), \

  87. #             NUM_TRIALS=30, \

  88. #             datafile_y=(ds['dataset_y'] if 'dataset_y' in ds else None), \

  89. #             extra_params=(ds['extra_params'] if 'extra_params' in ds else None), \

  90. #             ds_name=ds['name'], \

  91. #             n_jobs=multiprocessing.cpu_count())

  92.     print()

  93. 

  94. # import functools

  95. # from libs import *

  96. # from pygraph.kernels.spKernel import spkernel

  97. # from pygraph.utils.kernels import deltakernel, kernelsum

  98. # from sklearn.metrics.pairwise import rbf_kernel

  99. 

  100. # dslist = [

  101. #     {'name': 'Acyclic', 'dataset': '../datasets/acyclic/dataset_bps.ds', 'task': 'regression'}, # node symb

  102. # #     {'name': 'COIL-DEL', 'dataset': '../datasets/COIL-DEL/COIL-DEL_A.txt'}, # edge symb, node nsymb

  103. # #    {'name': 'PAH', 'dataset': '../datasets/PAH/dataset.ds',}, # unlabeled

  104. # #    {'name': 'MAO', 'dataset': '../datasets/MAO/dataset.ds',}, # node/edge symb

  105. # #    {'name': 'MUTAG', 'dataset': '../datasets/MUTAG/MUTAG.mat',

  106. # #        'extra_params': {'am_sp_al_nl_el': [0, 0, 3, 1, 2]}}, # node/edge symb

  107. # #    {'name': 'Alkane', 'dataset': '../datasets/Alkane/dataset.ds', 'task': 'regression',

  108. # #        'dataset_y': '../datasets/Alkane/dataset_boiling_point_names.txt',}, # contains single node graph, node symb

  109. # #     {'name': 'BZR', 'dataset': '../datasets/BZR_txt/BZR_A_sparse.txt'}, # node symb/nsymb

  110. # #     {'name': 'COX2', 'dataset': '../datasets/COX2_txt/COX2_A_sparse.txt'}, # node symb/nsymb

  111. # #    {'name': 'Mutagenicity', 'dataset': '../datasets/Mutagenicity/Mutagenicity_A.txt'}, # node/edge symb

  112. # #    {'name': 'ENZYMES', 'dataset': '../datasets/ENZYMES_txt/ENZYMES_A_sparse.txt'}, # node symb/nsymb

  113. # #     {'name': 'Fingerprint', 'dataset': '../datasets/Fingerprint/Fingerprint_A.txt'},

  114. # #    {'name': 'Letter-med', 'dataset': '../datasets/Letter-med/Letter-med_A.txt'},

  115. # #     {'name': 'DHFR', 'dataset': '../datasets/DHFR_txt/DHFR_A_sparse.txt'}, # node symb/nsymb

  116. # #     {'name': 'SYNTHETIC', 'dataset': '../datasets/SYNTHETIC_txt/SYNTHETIC_A_sparse.txt'}, # node symb/nsymb

  117. # #     {'name': 'MSRC9', 'dataset': '../datasets/MSRC_9_txt/MSRC_9_A.txt'}, # node symb

  118. # #     {'name': 'MSRC21', 'dataset': '../datasets/MSRC_21_txt/MSRC_21_A.txt'}, # node symb

  119. # #     {'name': 'FIRSTMM_DB', 'dataset': '../datasets/FIRSTMM_DB/FIRSTMM_DB_A.txt'}, # node symb/nsymb ,edge nsymb

  120. 

  121. # #     {'name': 'PROTEINS', 'dataset': '../datasets/PROTEINS_txt/PROTEINS_A_sparse.txt'}, # node symb/nsymb

  122. # #     {'name': 'PROTEINS_full', 'dataset': '../datasets/PROTEINS_full_txt/PROTEINS_full_A_sparse.txt'}, # node symb/nsymb

  123. # #    {'name': 'D&D', 'dataset': '../datasets/D&D/DD.mat',

  124. # #     'extra_params': {'am_sp_al_nl_el': [0, 1, 2, 1, -1]}}, # node symb

  125. # #     {'name': 'AIDS', 'dataset': '../datasets/AIDS/AIDS_A.txt'}, # node symb/nsymb, edge symb

  126. # #     {'name': 'NCI1', 'dataset': '../datasets/NCI1/NCI1.mat',

  127. # #         'extra_params': {'am_sp_al_nl_el': [1, 1, 2, 0, -1]}}, # node symb

  128. # #     {'name': 'NCI109', 'dataset': '../datasets/NCI109/NCI109.mat',

  129. # #         'extra_params': {'am_sp_al_nl_el': [1, 1, 2, 0, -1]}}, # node symb

  130. # #     {'name': 'NCI-HIV', 'dataset': '../datasets/NCI-HIV/AIDO99SD.sdf',

  131. # #         'dataset_y': '../datasets/NCI-HIV/aids_conc_may04.txt',}, # node/edge symb

  132. 

  133. # #     # not working below

  134. # #     {'name': 'PTC_FM', 'dataset': '../datasets/PTC/Train/FM.ds',},

  135. # #     {'name': 'PTC_FR', 'dataset': '../datasets/PTC/Train/FR.ds',},

  136. # #     {'name': 'PTC_MM', 'dataset': '../datasets/PTC/Train/MM.ds',},

  137. # #     {'name': 'PTC_MR', 'dataset': '../datasets/PTC/Train/MR.ds',},

  138. # ]

  139. # estimator = spkernel

  140. # mixkernel = functools.partial(kernelsum, deltakernel, rbf_kernel)

  141. # param_grid_precomputed = {'node_kernels': [{'symb': deltakernel, 'nsymb': rbf_kernel, 'mix': mixkernel}]}

  142. # param_grid = [{'C': np.logspace(-10, 10, num = 41, base = 10)},

  143. #               {'alpha': np.logspace(-10, 10, num = 41, base = 10)}]

  144. 

  145. # for ds in dslist:

  146. #     print()

  147. #     print(ds['name'])

  148. #     model_selection_for_precomputed_kernel(

  149. #         ds['dataset'], estimator, param_grid_precomputed,

  150. #         (param_grid[1] if ('task' in ds and ds['task'] == 'regression') else param_grid[0]),

  151. #         (ds['task'] if 'task' in ds else 'classification'), NUM_TRIALS=30,

  152. #         datafile_y=(ds['dataset_y'] if 'dataset_y' in ds else None),

  153. #         extra_params=(ds['extra_params'] if 'extra_params' in ds else None),

  154. #         ds_name=ds['name'])

  155. 

  156. # #     %lprun -f spkernel \

  157. # #         model_selection_for_precomputed_kernel( \

  158. # #             ds['dataset'], estimator, param_grid_precomputed, \

  159. # #             (param_grid[1] if ('task' in ds and ds['task'] == 'regression') else param_grid[0]), \

  160. # #             (ds['task'] if 'task' in ds else 'classification'), NUM_TRIALS=30, \

  161. # #             datafile_y=(ds['dataset_y'] if 'dataset_y' in ds else None), \

  162. # #             extra_params=(ds['extra_params'] if 'extra_params' in ds else None))

  163. #     print()

A Python package for graph kernels, graph edit distances and graph pre-image problem.