lazar 0.0.5 → 0.0.6

Sign up to get free protection for your applications and to get access to all the features.
Files changed (64) hide show
  1. checksums.yaml +4 -4
  2. data/.gitignore +10 -0
  3. data/.yardopts +4 -0
  4. data/Gemfile +2 -0
  5. data/LICENSE +674 -0
  6. data/README.md +44 -0
  7. data/Rakefile +1 -0
  8. data/VERSION +1 -0
  9. data/ext/lazar/Makefile +5 -0
  10. data/java/CdkDescriptorInfo.class +0 -0
  11. data/java/CdkDescriptorInfo.java +22 -0
  12. data/java/CdkDescriptors.class +0 -0
  13. data/java/CdkDescriptors.java +141 -0
  14. data/java/Jmol.jar +0 -0
  15. data/java/JoelibDescriptorInfo.class +0 -0
  16. data/java/JoelibDescriptorInfo.java +15 -0
  17. data/java/JoelibDescriptors.class +0 -0
  18. data/java/JoelibDescriptors.java +60 -0
  19. data/java/Rakefile +15 -0
  20. data/java/cdk-1.4.19.jar +0 -0
  21. data/java/joelib2.jar +0 -0
  22. data/java/log4j.jar +0 -0
  23. data/lazar.gemspec +28 -0
  24. data/lib/SMARTS_InteLigand.txt +983 -0
  25. data/mongoid.yml +8 -0
  26. data/test/data/CPDBAS_v5c_1547_29Apr2008part.sdf +13553 -0
  27. data/test/data/CPDBAS_v5d_cleaned/CPDBAS_v5d_20Nov2008_mouse_TD50.csv +436 -0
  28. data/test/data/CPDBAS_v5d_cleaned/CPDBAS_v5d_20Nov2008_rat_TD50.csv +568 -0
  29. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Hamster.csv +87 -0
  30. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mouse.csv +978 -0
  31. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_MultiCellCall.csv +1120 -0
  32. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_MultiCellCall_no_duplicates.csv +1113 -0
  33. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mutagenicity.csv +850 -0
  34. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mutagenicity_no_duplicates.csv +829 -0
  35. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Rat.csv +1198 -0
  36. data/test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_SingleCellCall.csv +1505 -0
  37. data/test/data/EPAFHM.csv +618 -0
  38. data/test/data/EPAFHM.medi.csv +100 -0
  39. data/test/data/EPAFHM.mini.csv +22 -0
  40. data/test/data/EPA_v4b_Fathead_Minnow_Acute_Toxicity_LC50_mmol.csv +581 -0
  41. data/test/data/FDA_v3b_Maximum_Recommended_Daily_Dose_mmol.csv +1217 -0
  42. data/test/data/ISSCAN-multi.csv +59 -0
  43. data/test/data/LOAEL_log_mg_corrected_smiles.csv +568 -0
  44. data/test/data/LOAEL_log_mmol_corrected_smiles.csv +568 -0
  45. data/test/data/acetaldehyde.sdf +14 -0
  46. data/test/data/boiling_points.ext.sdf +11460 -0
  47. data/test/data/cpdb_100.csv +101 -0
  48. data/test/data/hamster_carcinogenicity.csv +86 -0
  49. data/test/data/hamster_carcinogenicity.mini.bool_float.csv +11 -0
  50. data/test/data/hamster_carcinogenicity.mini.bool_int.csv +11 -0
  51. data/test/data/hamster_carcinogenicity.mini.bool_string.csv +11 -0
  52. data/test/data/hamster_carcinogenicity.mini.csv +11 -0
  53. data/test/data/hamster_carcinogenicity.ntriples +618 -0
  54. data/test/data/hamster_carcinogenicity.sdf +2805 -0
  55. data/test/data/hamster_carcinogenicity.xls +0 -0
  56. data/test/data/hamster_carcinogenicity.yaml +352 -0
  57. data/test/data/hamster_carcinogenicity_with_errors.csv +88 -0
  58. data/test/data/kazius.csv +4070 -0
  59. data/test/data/multi_cell_call.csv +1067 -0
  60. data/test/data/multi_cell_call_no_dup.csv +1057 -0
  61. data/test/data/multicolumn.csv +8 -0
  62. data/test/data/rat_feature_dataset.csv +1179 -0
  63. data/test/data/wrong_dataset.csv +8 -0
  64. metadata +108 -8
data/test/data/wrong_dataset.csv ADDED
@@ -0,0 +1,8 @@
1
+ SMILES,Wrong Dataset
2
+ Tost,0
3
+ Is,1
4
+ A,0
5
+ Wrong,1
6
+ Dataset,0
7
+ Entry,1
8
+ O[C@@H]8[C@@H](O)[C@@H]1O[C@H](CO)[C@H]8O[C@H]7O[C@H](CO)[C@@H](O[C@H]6O[C@H](CO)[C@@H](O[C@H]5O[C@H](CO)[C@@H](O[C@H]4O[C@H](CO)[C@@H](O[C@H]3O[C@H](CO)[C@@H](O[C@H]2O[C@H](CO)[C@@H](O1)[C@H](O)[C@H]2O)[C@H](O)[C@H]3O)[C@H](O)[C@H]4O)[C@H](O)[C@H]5O)[C, 0
metadata CHANGED
@@ -1,7 +1,7 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: lazar
3
3
  version: !ruby/object:Gem::Version
4
- version: 0.0.5
4
+ version: 0.0.6
5
5
  platform: ruby
6
6
  authors:
7
7
  - Christoph Helma, Martin Guetlein, Andreas Maunz, Micha Rautenberg, David Vorgrimmler,
@@ -89,7 +89,30 @@ extensions:
89
89
  - ext/lazar/extconf.rb
90
90
  extra_rdoc_files: []
91
91
  files:
92
+ - ".gitignore"
93
+ - ".yardopts"
94
+ - Gemfile
95
+ - LICENSE
96
+ - README.md
97
+ - Rakefile
98
+ - VERSION
99
+ - ext/lazar/Makefile
92
100
  - ext/lazar/extconf.rb
101
+ - java/CdkDescriptorInfo.class
102
+ - java/CdkDescriptorInfo.java
103
+ - java/CdkDescriptors.class
104
+ - java/CdkDescriptors.java
105
+ - java/Jmol.jar
106
+ - java/JoelibDescriptorInfo.class
107
+ - java/JoelibDescriptorInfo.java
108
+ - java/JoelibDescriptors.class
109
+ - java/JoelibDescriptors.java
110
+ - java/Rakefile
111
+ - java/cdk-1.4.19.jar
112
+ - java/joelib2.jar
113
+ - java/log4j.jar
114
+ - lazar.gemspec
115
+ - lib/SMARTS_InteLigand.txt
93
116
  - lib/algorithm.rb
94
117
  - lib/bbrc.rb
95
118
  - lib/classification.rb
@@ -109,8 +132,47 @@ files:
109
132
  - lib/similarity.rb
110
133
  - lib/unique_descriptors.rb
111
134
  - lib/validation.rb
135
+ - mongoid.yml
112
136
  - test/all.rb
113
137
  - test/compound.rb
138
+ - test/data/CPDBAS_v5c_1547_29Apr2008part.sdf
139
+ - test/data/CPDBAS_v5d_cleaned/CPDBAS_v5d_20Nov2008_mouse_TD50.csv
140
+ - test/data/CPDBAS_v5d_cleaned/CPDBAS_v5d_20Nov2008_rat_TD50.csv
141
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Hamster.csv
142
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mouse.csv
143
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_MultiCellCall.csv
144
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_MultiCellCall_no_duplicates.csv
145
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mutagenicity.csv
146
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mutagenicity_no_duplicates.csv
147
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Rat.csv
148
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_SingleCellCall.csv
149
+ - test/data/EPAFHM.csv
150
+ - test/data/EPAFHM.medi.csv
151
+ - test/data/EPAFHM.mini.csv
152
+ - test/data/EPA_v4b_Fathead_Minnow_Acute_Toxicity_LC50_mmol.csv
153
+ - test/data/FDA_v3b_Maximum_Recommended_Daily_Dose_mmol.csv
154
+ - test/data/ISSCAN-multi.csv
155
+ - test/data/LOAEL_log_mg_corrected_smiles.csv
156
+ - test/data/LOAEL_log_mmol_corrected_smiles.csv
157
+ - test/data/acetaldehyde.sdf
158
+ - test/data/boiling_points.ext.sdf
159
+ - test/data/cpdb_100.csv
160
+ - test/data/hamster_carcinogenicity.csv
161
+ - test/data/hamster_carcinogenicity.mini.bool_float.csv
162
+ - test/data/hamster_carcinogenicity.mini.bool_int.csv
163
+ - test/data/hamster_carcinogenicity.mini.bool_string.csv
164
+ - test/data/hamster_carcinogenicity.mini.csv
165
+ - test/data/hamster_carcinogenicity.ntriples
166
+ - test/data/hamster_carcinogenicity.sdf
167
+ - test/data/hamster_carcinogenicity.xls
168
+ - test/data/hamster_carcinogenicity.yaml
169
+ - test/data/hamster_carcinogenicity_with_errors.csv
170
+ - test/data/kazius.csv
171
+ - test/data/multi_cell_call.csv
172
+ - test/data/multi_cell_call_no_dup.csv
173
+ - test/data/multicolumn.csv
174
+ - test/data/rat_feature_dataset.csv
175
+ - test/data/wrong_dataset.csv
114
176
  - test/dataset-long.rb
115
177
  - test/dataset.rb
116
178
  - test/descriptor-long.rb
@@ -149,18 +211,56 @@ signing_key:
149
211
  specification_version: 4
150
212
  summary: Lazar framework
151
213
  test_files:
152
- - test/lazar-long.rb
214
+ - test/all.rb
215
+ - test/compound.rb
216
+ - test/data/CPDBAS_v5c_1547_29Apr2008part.sdf
217
+ - test/data/CPDBAS_v5d_cleaned/CPDBAS_v5d_20Nov2008_mouse_TD50.csv
218
+ - test/data/CPDBAS_v5d_cleaned/CPDBAS_v5d_20Nov2008_rat_TD50.csv
219
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Hamster.csv
220
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mouse.csv
221
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_MultiCellCall.csv
222
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_MultiCellCall_no_duplicates.csv
223
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mutagenicity.csv
224
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Mutagenicity_no_duplicates.csv
225
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_Rat.csv
226
+ - test/data/CPDBAS_v5d_cleaned/DSSTox_Carcinogenic_Potency_DBS_SingleCellCall.csv
227
+ - test/data/EPAFHM.csv
228
+ - test/data/EPAFHM.medi.csv
229
+ - test/data/EPAFHM.mini.csv
230
+ - test/data/EPA_v4b_Fathead_Minnow_Acute_Toxicity_LC50_mmol.csv
231
+ - test/data/FDA_v3b_Maximum_Recommended_Daily_Dose_mmol.csv
232
+ - test/data/ISSCAN-multi.csv
233
+ - test/data/LOAEL_log_mg_corrected_smiles.csv
234
+ - test/data/LOAEL_log_mmol_corrected_smiles.csv
235
+ - test/data/acetaldehyde.sdf
236
+ - test/data/boiling_points.ext.sdf
237
+ - test/data/cpdb_100.csv
238
+ - test/data/hamster_carcinogenicity.csv
239
+ - test/data/hamster_carcinogenicity.mini.bool_float.csv
240
+ - test/data/hamster_carcinogenicity.mini.bool_int.csv
241
+ - test/data/hamster_carcinogenicity.mini.bool_string.csv
242
+ - test/data/hamster_carcinogenicity.mini.csv
243
+ - test/data/hamster_carcinogenicity.ntriples
244
+ - test/data/hamster_carcinogenicity.sdf
245
+ - test/data/hamster_carcinogenicity.xls
246
+ - test/data/hamster_carcinogenicity.yaml
247
+ - test/data/hamster_carcinogenicity_with_errors.csv
248
+ - test/data/kazius.csv
249
+ - test/data/multi_cell_call.csv
250
+ - test/data/multi_cell_call_no_dup.csv
251
+ - test/data/multicolumn.csv
252
+ - test/data/rat_feature_dataset.csv
253
+ - test/data/wrong_dataset.csv
254
+ - test/dataset-long.rb
255
+ - test/dataset.rb
153
256
  - test/descriptor-long.rb
257
+ - test/descriptor.rb
154
258
  - test/error.rb
155
- - test/dataset-long.rb
156
259
  - test/feature.rb
157
- - test/compound.rb
260
+ - test/fminer-long.rb
158
261
  - test/fminer.rb
159
- - test/all.rb
160
- - test/dataset.rb
161
262
  - test/lazar-fminer.rb
263
+ - test/lazar-long.rb
162
264
  - test/lazar-physchem-short.rb
163
265
  - test/setup.rb
164
266
  - test/validation.rb
165
- - test/descriptor.rb
166
- - test/fminer-long.rb