nkululeko 0.86.5__py3-none-any.whl → 0.86.7__py3-none-any.whl

nkululeko/constants.py

@@ -1,2 +1,2 @@
-VERSION="0.86.5"
+VERSION="0.86.7"
 SAMPLING_RATE = 16000

nkululeko/data/dataset.py

@@ -150,6 +150,13 @@ class Dataset:
                 self.got_speaker = got_speaker2 or self.got_speaker
                 self.got_gender = got_gender2 or self.got_gender
                 self.got_age = got_age2 or self.got_age
+                if audformat.is_filewise_index(df_target.index):
+                    try:
+                        df_target = df_target.loc[df.index.get_level_values("file")]
+                        df_target = df_target.set_index(df.index)
+                    except KeyError:
+                        # just a try...
+                        pass
                 if got_target2:
                     df[self.target] = df_target[self.target]
                 if got_speaker2:
@@ -255,6 +262,7 @@ class Dataset:
         df = pd.DataFrame()
         for table in df_files:
             source_df = db.tables[table].df
+            # check if columns should be renamed
             source_df = self._check_cols(source_df)
             # create a dataframe with the index (the filenames)
             df_local = pd.DataFrame(index=source_df.index)

nkululeko/modelrunner.py

@@ -30,6 +30,8 @@ class Modelrunner:
         # intialize a new model
         model_type = glob_conf.config["MODEL"]["type"]
         self._select_model(model_type)
+        self.best_performance = 0
+        self.best_epoch = 0
 
     def do_epochs(self):
         # initialze results
@@ -51,7 +53,8 @@ class Modelrunner:
             # epochs are handled by Huggingface API
             self.model.train()
             report = self.model.predict()
-            # todo: findout the best epoch
+            # todo: findout the best epoch, no need
+            # since oad_best_model_at_end is given in training args
             epoch = epoch_num
             report.set_id(self.run, epoch)
             plot_name = self.util.get_plot_name() + f"_{self.run}_{epoch:03d}_cnf"
@@ -77,10 +80,15 @@ class Modelrunner:
                 report.set_id(self.run, epoch)
                 plot_name = self.util.get_plot_name() + f"_{self.run}_{epoch:03d}_cnf"
                 reports.append(report)
+                test_score_metric = report.get_result().get_test_result()
                 self.util.debug(
-                    f"run: {self.run} epoch: {epoch}: result: "
-                    f"{reports[-1].get_result().get_test_result()}"
+                    f"run: {self.run} epoch: {epoch}: result: {test_score_metric}"
                 )
+                # print(f"performance: {performance.split(' ')[1]}")
+                performance = float(test_score_metric.split(' ')[1])
+                if performance > self.best_performance:
+                    self.best_performance = performance
+                    self.best_epoch = epoch
                 if plot_epochs:
                     self.util.debug(f"plotting conf matrix to {plot_name}")
                     report.plot_confmatrix(plot_name, epoch)
@@ -110,11 +118,14 @@ class Modelrunner:
                             f"reached patience ({str(patience)}): early stopping"
                         )
                         break
+        # After training, report the best performance and epoch
+        best_report = reports[self.best_epoch]
+        # self.util.debug(f"Best score at epoch: {self.best_epoch}, UAR: {self.best_performance}") # move to reporter below
 
         if not plot_epochs:
             # Do at least one confusion matrix plot
             self.util.debug(f"plotting confusion matrix to {plot_name}")
-            reports[-1].plot_confmatrix(plot_name, epoch)
+            best_report.plot_confmatrix(plot_name, self.best_epoch)
         return reports, epoch
 
     def _select_model(self, model_type):

nkululeko/reporting/reporter.py

@@ -122,7 +122,7 @@ class Reporter:
         self.truths = np.digitize(self.truths, bins) - 1
         self.preds = np.digitize(self.preds, bins) - 1
 
-    def plot_confmatrix(self, plot_name, epoch):
+    def plot_confmatrix(self, plot_name, epoch=None):
         if not self.util.exp_is_classification():
             self.continuous_to_categorical()
         self._plot_confmat(self.truths, self.preds, plot_name, epoch)
@@ -156,9 +156,11 @@ class Reporter:
             pred = np.digitize(pred, bins) - 1
         self._plot_confmat(truth, pred.astype("int"), plot_name, 0)
 
-    def _plot_confmat(self, truths, preds, plot_name, epoch):
+    def _plot_confmat(self, truths, preds, plot_name, epoch=None):
         # print(truths)
         # print(preds)
+        if epoch is None:
+            epoch = self.epoch
         fig_dir = self.util.get_path("fig_dir")
         labels = glob_conf.labels
         fig = plt.figure()  # figsize=[5, 5]
@@ -225,7 +227,7 @@ class Reporter:
 
         res_dir = self.util.get_path("res_dir")
         rpt = (
-            f"epoch: {epoch}, UAR: {uar_str}"
+            f"Best score at epoch: {epoch}, UAR: {uar_str}"
             + f", (+-{up_str}/{low_str}), ACC: {acc_str}"
         )
         # print(rpt)
@@ -237,7 +239,9 @@ class Reporter:
     def set_filename_add(self, my_string):
         self.filenameadd = f"_{my_string}"
 
-    def print_results(self, epoch):
+    def print_results(self, epoch=None):
+        if epoch is None:
+            epoch = self.epoch
         """Print all evaluation values to text file."""
         res_dir = self.util.get_path("res_dir")
         file_name = f"{res_dir}{self.util.get_exp_name()}_{epoch}{self.filenameadd}.txt"
@@ -262,12 +266,14 @@ class Reporter:
                     c_res = rpt[l]["f1-score"]
                     c_ress[i] = float(f"{c_res:.3f}")
                 self.util.debug(f"labels: {labels}")
-                f1_per_class = f"result per class (F1 score): {c_ress}"
+                f1_per_class = (
+                    f"result per class (F1 score): {c_ress} from epoch: {epoch}"
+                )
                 if len(np.unique(self.truths)) == 2:
                     fpr, tpr, _ = roc_curve(self.truths, self.preds)
                     auc_score = auc(fpr, tpr)
                     pauc_score = roc_auc_score(self.truths, self.preds, max_fpr=0.1)
-                    auc_pauc = f"auc: {auc_score:.3f}, pauc: {pauc_score:.3f}"
+                    auc_pauc = f"auc: {auc_score:.3f}, pauc: {pauc_score:.3f} from epoch: {epoch}"
                     self.util.debug(auc_pauc)
                 self.util.debug(f1_per_class)
                 rpt_str = f"{json.dumps(rpt)}\n{f1_per_class}"

nkululeko/runmanager.py

@@ -63,8 +63,7 @@ class Runmanager:
             )
             self.reports, last_epoch = self.modelrunner.do_epochs()
             # wrap up the run
-            plot_anim_progression = self.util.config_val(
-                "PLOT", "anim_progression", 0)
+            plot_anim_progression = self.util.config_val("PLOT", "anim_progression", 0)
             if plot_anim_progression:
                 plot_name_suggest = self.util.get_exp_name()
                 plot_name = (
@@ -88,8 +87,7 @@ class Runmanager:
                     + "_epoch_progression"
                 )
                 self.util.debug(f"plotting progression to {plot_name}")
-                self.reports[-1].plot_epoch_progression(
-                    self.reports, plot_name)
+                self.reports[-1].plot_epoch_progression(self.reports, plot_name)
             # remember the best run
             best_report = self.get_best_result(self.reports)
             plot_best_model = self.util.config_val("PLOT", "best_model", False)
@@ -107,9 +105,10 @@ class Runmanager:
                 )
                 self.print_model(best_report, plot_name)
             # finally, print out the numbers for this run
-            self.reports[-1].print_results(
-                int(self.util.config_val("EXP", "epochs", 1))
-            )
+            # self.reports[-1].print_results(
+            #     int(self.util.config_val("EXP", "epochs", 1))
+            # )
+            best_report.print_results(best_report.epoch)
             self.best_results.append(best_report)
             self.last_epochs.append(last_epoch)
 
@@ -145,19 +144,18 @@ class Runmanager:
         )
         self.print_model(report, plot_name)
 
-    def print_model(self, report, plot_name):
+    def print_model(self, reporter, plot_name):
         """Print a confusion matrix for a special report.
 
         Args:
             report: for which report (will be computed newly from model)
             plot_name: name of plot file
         """
-        epoch = report.epoch
         # self.load_model(report)
         # report = self.model.predict()
         self.util.debug(f"plotting conf matrix to {plot_name}")
-        report.plot_confmatrix(plot_name, epoch)
-        report.print_results(epoch)
+        reporter.plot_confmatrix(plot_name)
+        reporter.print_results()
 
     def load_model(self, report):
         """Load a model from disk for a specific run and epoch and evaluate it.

{nkululeko-0.86.5.dist-info → nkululeko-0.86.7.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.86.5
+Version: 0.86.7
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -343,6 +343,14 @@ F. Burkhardt, Johannes Wagner, Hagen Wierstorf, Florian Eyben and Björn Schulle
 Changelog
 =========
 
+Version 0.86.7
+--------------
+* handles now audformat tables where the target is in a file index
+  
+Version 0.86.6
+--------------
+* now best (not last) result is shown at end
+
 Version 0.86.5
 --------------
 * fix audio path detection in data csv import

{nkululeko-0.86.5.dist-info → nkululeko-0.86.7.dist-info}/RECORD

@@ -2,7 +2,7 @@ nkululeko/__init__.py,sha256=62f8HiEzJ8rG2QlTFJXUCMpvuH3fKI33DoJSj33mscc,63
 nkululeko/aug_train.py,sha256=YhuZnS_WVWnun9G-M6g5n6rbRxoVREz6Zh7k6qprFNQ,3194
 nkululeko/augment.py,sha256=4MG0apTAG5RgkuJrYEjGgDdbodZWi_HweSPNI1JJ5QA,3051
 nkululeko/cacheddataset.py,sha256=lIJ6hUo5LoxSrzXtWV8mzwO7wRtUETWnOQ4ws2XfL1E,969
-nkululeko/constants.py,sha256=ctptCGup_HGCOxioUojLqMivtVfYq8CZDLHJprDr9aE,39
+nkululeko/constants.py,sha256=CscqJhC7nceHk2wmZd2bBFSeFExtr0HkXt99qpAZU4E,39
 nkululeko/demo.py,sha256=WSKr-W5uJ9DQfemK923g7Hd5V3kgAn03Er0JX1Pa45I,5142
 nkululeko/demo_feats.py,sha256=sAeGFojhEj9WEDFtG3SzPBmyYJWLF2rkbpp65m8Ujo4,2025
 nkululeko/demo_predictor.py,sha256=es56xbT8ifkS_vnrlb5NTZT54gNmeUtNlA4zVA_gnN8,4757
@@ -13,14 +13,14 @@ nkululeko/feature_extractor.py,sha256=8mssYKmo4LclVI-hiLmJEDZ0ZPyDavFG2YwtXcrGzw
 nkululeko/file_checker.py,sha256=LoLnL8aHpW-axMQ46qbqrManTs5otG9ShpEZuz9iRSk,3474
 nkululeko/filter_data.py,sha256=w-X2mhKdYr5DxDIz50E5yzO6Jmzk4jjDBoXsgOOVtcA,7222
 nkululeko/glob_conf.py,sha256=KL9YJQTHvTztxo1vr25qRRgaPnx4NTg0XrdbovKGMmw,525
-nkululeko/modelrunner.py,sha256=iCmfJxsS2UafcikjRdUqPQuqQMOYA-Ctr3et3HeNR3c,10452
+nkululeko/modelrunner.py,sha256=OU35qwP94GxW_EtL4I2-RhqB-wxbjNvp8CIHNbtnt7Q,11155
 nkululeko/multidb.py,sha256=fG3VukEWP1vreVN4gB1IRXxwwg4jLftsSEYtu0o1f78,5634
 nkululeko/nkuluflag.py,sha256=PGWSmZz-PiiHLgcZJAoGOI_Y-sZDVI1ksB8p5r7riWM,3725
 nkululeko/nkululeko.py,sha256=Kn3s2E3yyH8cJ7z6lkMxrnqtCxTu7-qfe9Zr_ONTD5g,1968
 nkululeko/plots.py,sha256=C2mwQFK0Vxfl5ZM7CO87tULDoEf7G16ek0nU77bhOc4,23070
 nkululeko/predict.py,sha256=sF091sSSLnEWcISx9ZcULLie3tY5XeFsQJd6b3vrxFg,2409
 nkululeko/resample.py,sha256=2d9eao_0sLrGZ_KSl8OVKsPor3BkFrlmMhrpB9WelIs,4267
-nkululeko/runmanager.py,sha256=eTM1DNQKt1lxYhzt4vZyZluPXW9sWlIJHNQzex4lkJU,7624
+nkululeko/runmanager.py,sha256=Na8oPn59lRFiNMsYChRHBRgw40mBcw0Rwl2Kz1RUsA0,7614
 nkululeko/scaler.py,sha256=4nkIqoajkIkuTPK0Z02ifMN_awl6fP_i-GBYdoGYgGM,4101
 nkululeko/segment.py,sha256=YLKckX44tbvTb3LrdgYw9X4guzuF27sutl92z9DkpZU,4835
 nkululeko/syllable_nuclei.py,sha256=Sky-C__MeUDaxqHnDl2TGLLYOYvsahD35TUjWGeG31k,10047
@@ -45,7 +45,7 @@ nkululeko/autopredict/ap_stoi.py,sha256=It0Lk-ki-gohA2AzD8nkLAN2WahYvD9rPDGTQuvd
 nkululeko/autopredict/ap_valence.py,sha256=n-hctRKySzhmJtowuMOTUu0T_ld3uK5pnfOzWeWW4VM,1024
 nkululeko/autopredict/estimate_snr.py,sha256=S-bpS0xFkwWc4Ch75UrjbS8y538lQ0U3g_iLRFXureY,5048
 nkululeko/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-nkululeko/data/dataset.py,sha256=JGzMD6HIvkFkYBekmbmslIKc5ADaCj06T-8gpqH_kFo,27650
+nkululeko/data/dataset.py,sha256=hUD0NqWCfRaSHG8JNs1MsPb0zjUZAf8FJkg_c0ebq0Q,28046
 nkululeko/data/dataset_csv.py,sha256=dzOrbKB8t0UATAIYaKAOqHTogmYPBqskt6Hak7VjbSM,4537
 nkululeko/feat_extract/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nkululeko/feat_extract/feats_agender.py,sha256=Qm69G4kqAyTVVk7wwRgrXlNwGaDMGRYyKGpuf0vOEgM,3113
@@ -96,7 +96,7 @@ nkululeko/reporting/defines.py,sha256=IsY1YgKRMaABpylVKjBJgJ5bNCEbGCVA_E6pivraqS
 nkululeko/reporting/latex_writer.py,sha256=qiCRSmB4KOD_za4oHu5x-PhwjZohzfo8wecMOwlXZwc,1886
 nkululeko/reporting/report.py,sha256=W0rcigDdjBvxZQ3pZja_gvToILYvaZ1BFtnN2qFRfYI,1060
 nkululeko/reporting/report_item.py,sha256=siWeGNgo4bAE46YBMNcsdf3jTMTy76BO9Fi6DTvDig4,533
-nkululeko/reporting/reporter.py,sha256=II3QyeneAv8xQDBZ-qE_GJL8_WV_yXqLwBUYqrjqwPo,13938
+nkululeko/reporting/reporter.py,sha256=S9A62AxdMTEV-9XDUQNxdoevGLXBP52WiDmZ694QMV4,14161
 nkululeko/reporting/result.py,sha256=nSN5or-Py2GPRWHkWpGRh7UCi1W0er7WLEHz8fYLk-A,742
 nkululeko/segmenting/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nkululeko/segmenting/seg_inaspeechsegmenter.py,sha256=pmLHuXsaqvcdYxB4PSW9l1mbQWZZBJFhi_CGabqydas,1947
@@ -105,8 +105,8 @@ nkululeko/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nkululeko/utils/files.py,sha256=UiGAtZRWYjHSvlmPaTMtzyNNGE6qaLaxQkybctS7iRM,4021
 nkululeko/utils/stats.py,sha256=1yUq0FTOyqkU8TwUocJRYdJaqMU5SlOBBRUun9STo2M,2829
 nkululeko/utils/util.py,sha256=ILpfNuaeq-hy1bUkRhVrzO2wG9z9Upaozs9EBoIaMG0,14123
-nkululeko-0.86.5.dist-info/LICENSE,sha256=0zGP5B_W35yAcGfHPS18Q2B8UhvLRY3dQq1MhpsJU_U,1076
-nkululeko-0.86.5.dist-info/METADATA,sha256=HrTVTfGh3KDsmyBFijAp5tMINdiBvHhsC8E0_YwBjwE,37848
-nkululeko-0.86.5.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-nkululeko-0.86.5.dist-info/top_level.txt,sha256=DPFNNSHPjUeVKj44dVANAjuVGRCC3MusJ08lc2a8xFA,10
-nkululeko-0.86.5.dist-info/RECORD,,
+nkululeko-0.86.7.dist-info/LICENSE,sha256=0zGP5B_W35yAcGfHPS18Q2B8UhvLRY3dQq1MhpsJU_U,1076
+nkululeko-0.86.7.dist-info/METADATA,sha256=t5cI43YRp3qmyJj03ACfgCbKoAuLYImDCLS1QkYbMQM,38024
+nkululeko-0.86.7.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+nkululeko-0.86.7.dist-info/top_level.txt,sha256=DPFNNSHPjUeVKj44dVANAjuVGRCC3MusJ08lc2a8xFA,10
+nkululeko-0.86.7.dist-info/RECORD,,