Merge 59ace0a into ccd7121

conda_requirements.txt

@@ -1,14 +1,14 @@
 python>=3.6.3,<3.7
-ipython=6.3.1
+ipython=6.5.0
 jupyter=1.0.0
 joblib=0.11
 matplotlib=2.1.2
 nose=1.3.7
-notebook=5.4.1
+notebook=5.7.2
 numpy=1.13.3
-pandas=0.22.0
+pandas=0.23.4
 scikit-learn=0.19.1
-scipy=1.0.0
+scipy=1.1.0
 seaborn=0.9.0
 skll=1.5.2
 statsmodels=0.9.0
@@ -22,4 +22,4 @@ sphinx
 sphinx_rtd_theme
 coverage
 parameterized
-setuptools
+setuptools

requirements.txt

@@ -1,13 +1,13 @@
-ipython==6.3.1
+ipython==6.5.0
 jupyter==1.0.0
 joblib==0.11
 matplotlib==2.1.2
 nose==1.3.7
-notebook==5.4.1
+notebook==5.7.2
 numpy==1.13.3
-pandas==0.22.0
+pandas==0.23.4
 scikit-learn==0.19.1
-scipy==1.0.0
+scipy==1.1.0
 seaborn==0.9.0
 skll==1.5.2
 statsmodels==0.9.0
@@ -18,4 +18,4 @@ sphinx
 sphinx_rtd_theme
 coverage
 parameterized
-setuptools
+setuptools

rsmtool/analyzer.py

@@ -310,9 +310,8 @@ def compute_basic_descriptives(df, selected_features):
                                   'N': len(df_desc)})
 
         # reorder the columns to make it look better
-        df_output = df_output[['mean', 'std. dev.', 'min', 'max',
-                               'skewness', 'kurtosis', 'Correlation',
-                               'p', 'N']]
+        df_output = df_output[['mean', 'std. dev.', 'min', 'max', 'skewness',
+                               'kurtosis', 'Correlation', 'p', 'N']]
 
         return df_output
 
@@ -599,7 +598,7 @@ def metrics_helper(human_scores,
             Reference standard deviation for system scores.  This is used to compute SMD and
             should be the standard deviation for the whole population when SMD are computed
             for individual subgroups.
-            When None, this will be computed as the standard devaiation of `system_scores`.
+            When None, this will be computed as the standard deviation of `system_scores`.
 
         Returns
         -------
@@ -679,7 +678,6 @@ def metrics_helper(human_scores,
         rmse = np.sqrt(mse)
 
         # return everything as a series
-
         metrics = pd.Series({'kappa': unweighted_kappa,
                              'wtkappa': quadratic_weighted_kappa,
                              'exact_agr': human_system_agreement,
@@ -715,7 +713,8 @@ def compute_disattenuated_correlations(human_machine_corr,
             Series containing of pearson's correlation coefficients human-machine correlations
         human_human_corr : pandas Series
             Series containing of pearson's correlation coefficients for human-human correlations.
-            This can contain a single value or have the index matching that of human-machine correlations
+            This can contain a single value or have the index matching that of human-machine
+            correlations
 
         Returns
         -------
@@ -733,14 +732,16 @@ def compute_disattenuated_correlations(human_machine_corr,
         # we now concatenate the two series on index
         df_correlations = pd.concat([human_machine_corr, human_human_corr],
                                     axis=1,
+                                    sort=True,
                                     keys=['corr_HM', 'corr_HH'])
 
         # if any of the HH correlations are negative, we will ignore these
         # and treat them as Nones
         with np.errstate(invalid='ignore'):
             df_correlations['sqrt_HH'] = np.sqrt(df_correlations['corr_HH'])
 
-        df_correlations['corr_disattenuated'] = df_correlations['corr_HM'] / df_correlations['sqrt_HH']
+        df_correlations['corr_disattenuated'] = (df_correlations['corr_HM'] /
+                                                 df_correlations['sqrt_HH'])
 
         return df_correlations
 
@@ -791,7 +792,7 @@ def compute_correlations_by_group(self,
         df_desc_all[grouping_variable] = 'All data'
 
         # combine the two data frames
-        df_desc_combined = pd.concat([df_desc, df_desc_all])
+        df_desc_combined = pd.concat([df_desc, df_desc_all], sort=True)
         df_desc_combined.reset_index(drop=True, inplace=True)
 
         # compute the various (marginal and partial) correlations with score
@@ -1085,7 +1086,7 @@ def compute_metrics_by_group(self,
         df_preds_all[grouping_variable] = 'All data'
 
         # combine the two data frames
-        df_preds_combined = pd.concat([df_test, df_preds_all])
+        df_preds_combined = pd.concat([df_test, df_preds_all], sort=True)
         df_preds_combined.reset_index(drop=True, inplace=True)
 
         # group by the grouping_variable columns

rsmtool/comparer.py

@@ -103,7 +103,7 @@ def make_summary_stat_df(df):
                 warnings.simplefilter('ignore')
                 series.append(df.apply(summary_func))
 
-        res = pd.concat(series, axis=1)
+        res = pd.concat(series, axis=1, sort=True)
         res.columns = ['MEAN', 'SD', 'MEDIAN', 'MIN', 'MAX']
         return res
 
@@ -197,7 +197,7 @@ def compute_correlations_between_versions(df_old,
                                                        df_merged['{}%%%old'.format(feature)])[0]})
             correlation_list.append(df_cor)
 
-        df_correlations = pd.concat(correlation_list)
+        df_correlations = pd.concat(correlation_list, sort=True)
         df_correlations.index = df_correlations['Feature']
         df_correlations.index.name = None
 

rsmtool/modeler.py

@@ -169,7 +169,7 @@ def ols_coefficients_to_dataframe(coefs):
                                       'coefficient': coefs['const']}])
 
         # append the non-intercept frame to the intercept one
-        df_coef = df_intercept.append(df_non_intercept, ignore_index=True)
+        df_coef = df_intercept.append(df_non_intercept, sort=True, ignore_index=True)
 
         # we always want to have the feature column first
         df_coef = df_coef[['feature', 'coefficient']]
@@ -216,7 +216,7 @@ def skll_learner_params_to_dataframe(learner):
                                       'coefficient': intercept}])
 
         # append the non-intercept frame to the intercept one
-        df_coef = df_intercept.append(df_non_intercept, ignore_index=True)
+        df_coef = df_intercept.append(df_non_intercept, sort=True, ignore_index=True)
 
         # we always want to have the feature column first
         df_coef = df_coef[['feature', 'coefficient']]

rsmtool/notebooks/comparison/consistency.ipynb

@@ -10,7 +10,7 @@
    "source": [
     "def create_degradation_df(df_degradation, df_eval, df_consistency):\n",
     "    df_eval_for_degradation = df_eval[df_degradation.columns].copy()\n",
-    "    df_consistency_for_degradation = pd.concat([df_consistency]*len(df_eval))\n",
+    "    df_consistency_for_degradation = pd.concat([df_consistency]*len(df_eval), sort=True)\n",
     "    df_consistency_for_degradation = df_consistency_for_degradation[df_degradation.columns].copy()\n",
     "    df_consistency_for_degradation.index = df_eval_for_degradation.index\n",
     "\n",
@@ -21,7 +21,7 @@
     "    df = pd.concat(\n",
     "        [df_consistency_for_degradation,\n",
     "         df_eval_for_degradation,\n",
-    "         df_degradation])\n",
+    "         df_degradation], sort=True)\n",
     "    return df\n",
     "\n",
     "\n",
@@ -84,7 +84,7 @@
     "    df_degradation_new['version'] = 'new'\n",
     "    \n",
     "    \n",
-    "    df = pd.concat([df_degradation_old, df_degradation_new])\n",
+    "    df = pd.concat([df_degradation_old, df_degradation_new], sort=True)\n",
     "    df = df[['type', 'version', 'corr', 'kappa',\n",
     "             'wtkappa', 'exact_agr', 'adj_agr', 'SMD']]\n",
     "\n",
@@ -137,7 +137,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.4"
+   "version": "3.6.7"
   }
  },
  "nbformat": 4,

rsmtool/notebooks/comparison/feature_descriptives.ipynb

@@ -125,7 +125,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.3"
+   "version": "3.6.7"
   }
  },
  "nbformat": 4,

rsmtool/notebooks/comparison/header.ipynb

@@ -334,8 +334,8 @@
     "    df_diff['version'] = 'change'\n",
     "\n",
     "    tmp_df = pd.DataFrame(df_old, copy=True)\n",
-    "    tmp_df = tmp_df.append(df_new)\n",
-    "    tmp_df = tmp_df.append(df_diff)\n",
+    "    tmp_df = tmp_df.append(df_new, sort=True)\n",
+    "    tmp_df = tmp_df.append(df_diff, sort=True)\n",
     "    tmp_df.index.name = 'for'  \n",
     "    tmp_df = tmp_df.reset_index().sort_values(by=['for', 'version'], ascending=[True, False]).set_index(tmp_df.index.names)\n",
     "    tmp_df.index.name = None\n",

rsmtool/notebooks/consistency.ipynb

@@ -62,15 +62,17 @@
     "    markdown_strs.append(' - `wtkappa` < -0.1')\n",
     "    display(Markdown('\\n'.join(markdown_strs)))\n",
     "    df_eval_for_degradation = df_eval[df_degradation.columns].copy()\n",
-    "    df_consistency_for_degradation = pd.concat([df_consistency]*len(df_eval))\n",
+    "    df_consistency_for_degradation = pd.concat([df_consistency]*len(df_eval), sort=True)\n",
     "    df_consistency_for_degradation = df_consistency_for_degradation[df_degradation.columns].copy()\n",
     "    df_consistency_for_degradation.index = df_eval_for_degradation.index\n",
     "\n",
     "    df_consistency_for_degradation['type'] = 'H-H'\n",
     "    df_eval_for_degradation['type'] = 'H-M'\n",
     "    df_degradation['type'] = 'diff'\n",
     "\n",
-    "    df = pd.concat([df_consistency_for_degradation, df_eval_for_degradation, df_degradation])\n",
+    "    df = pd.concat([df_consistency_for_degradation,\n",
+    "                    df_eval_for_degradation,\n",
+    "                    df_degradation], sort=True)\n",
     "    df = df[['type','corr', 'kappa', 'wtkappa', 'exact_agr', 'adj_agr', 'SMD']]\n",
     "    df = df.reset_index()\n",
     "    df = df.set_index(['index', 'type']).sort_index(level='index')\n",
@@ -127,7 +129,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.4"
+   "version": "3.6.7"
   }
  },
  "nbformat": 4,

rsmtool/notebooks/features_by_group.ipynb

@@ -35,7 +35,7 @@
     "        df_train_feats_all = df_train_merged.copy()\n",
     "        df_train_feats_all[group] = 'All data'\n",
     "\n",
-    "        df_train_combined = pd.concat([df_train_feats, df_train_feats_all])\n",
+    "        df_train_combined = pd.concat([df_train_feats, df_train_feats_all], sort=True)\n",
     "        df_train_combined.reset_index(drop=True, inplace=True)\n",
     "\n",
     "        # decide on the the height per plot\n",

rsmtool/notebooks/summary/evaluation.ipynb

@@ -44,7 +44,7 @@
     "            df_eval.columns = new_column_names\n",
     "            evals.append(df_eval)          \n",
     "    if len(evals) > 0:\n",
-    "        df_evals = pd.concat(evals)\n",
+    "        df_evals = pd.concat(evals, sort=True)\n",
     "    else:\n",
     "        df_evals = pd.DataFrame()\n",
     "    return(df_evals)\n",
@@ -133,7 +133,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.3"
+   "version": "3.6.7"
   }
  },
  "nbformat": 4,

rsmtool/notebooks/summary/model.ipynb

@@ -49,7 +49,7 @@
     "                summs.append(model_summary)\n",
     "   \n",
     "    if not len(summs) == 0:\n",
-    "        df_summ = pd.concat(summs)\n",
+    "        df_summ = pd.concat(summs, sort=True)\n",
     "        display(Markdown(\"## Model summary\"))\n",
     "        display(HTML(df_summ[['N features', 'N negative',\n",
     "                              'learner', 'train_label']].to_html(index=True, \n",
@@ -63,7 +63,7 @@
     "                                       file_format=file_format_summarize)\n",
     "        \n",
     "    if not len(betas) == 0:\n",
-    "        df_betas_all = pd.concat(betas, axis=1)\n",
+    "        df_betas_all = pd.concat(betas, axis=1, sort=True)\n",
     "        df_betas_all.fillna('-', inplace=True)\n",
     "        display(Markdown(\"## Standardized coefficients\"))\n",
     "        display(HTML(df_betas_all.to_html(index=True, \n",
@@ -96,7 +96,7 @@
     "            fit.index = [model_id]\n",
     "            fits.append(fit)\n",
     "    if len(fits)>0:\n",
-    "        df_fit = pd.concat(fits)\n",
+    "        df_fit = pd.concat(fits, sort=True)\n",
     "        display(Markdown(\"## Model fit\"))\n",
     "        display(HTML(df_fit[['N responses', 'N features',\n",
     "                             'R2','R2_adjusted']].to_html(index=True,\n",
@@ -132,7 +132,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.3"
+   "version": "3.6.7"
   }
  },
  "nbformat": 4,

rsmtool/notebooks/summary/preprocessed_features.ipynb

@@ -24,7 +24,7 @@
     "            model_corrs.index = [model_id]\n",
     "            corrs.append(model_corrs)\n",
     "    if not len(corrs) == 0:\n",
-    "        df_summ = pd.concat(corrs)\n",
+    "        df_summ = pd.concat(corrs, sort=True)\n",
     "        display(header)\n",
     "        display(HTML(df_summ.to_html(index=True, classes = ['sortable'],\n",
     "                                     escape=False,\n",
@@ -109,7 +109,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.3"
+   "version": "3.6.7"
   }
  },
  "nbformat": 4,