upload random on init

weecology · Nov 26, 2024 · 063044a · 063044a
1 parent e1ba9a9
commit 063044a
Show file tree

Hide file tree

Showing 4 changed files with 31 additions and 3 deletions.
diff --git a/src/detection.py b/src/detection.py
@@ -209,14 +209,16 @@ def preprocess_and_train(config, validation_df=None, model_type="detection"):
                                root_dir=config.detection_model.train_image_dir,
                                save_dir=config.detection_model.crop_image_dir)
 
-    validation_df = data_processing.preprocess_images(validation_df,
+    if not validation_df.empty:
+        validation_df = data_processing.preprocess_images(validation_df,
                                     root_dir=config.detection_model.train_image_dir,
                                     save_dir=config.detection_model.crop_image_dir)
 
     # Limit empty frames
     if config.detection_model.limit_empty_frac > 0:
         train_df = limit_empty_frames(train_df, config.detection_model.limit_empty_frac)
-        validation_df = limit_empty_frames(validation_df, config.detection_model.limit_empty_frac)
+        if not validation_df.empty:
+            validation_df = limit_empty_frames(validation_df, config.detection_model.limit_empty_frac)
 
     # Train model
     # Load existing model

diff --git a/src/pipeline.py b/src/pipeline.py
@@ -1,5 +1,7 @@
 from datetime import datetime
 import os
+import glob
+import random
 
 import pandas as pd
 from omegaconf import DictConfig
@@ -37,6 +39,19 @@ def run(self):
                 **self.config.label_studio)
             if new_annotations is None:
                 print("No new annotations, exiting")
+                if self.config.force_upload:
+                    image_paths = glob.glob(os.path.join(self.config.active_learning.image_dir, "*.jpg"))
+                    image_paths = random.sample(image_paths, 10)
+                    label_studio.upload_to_label_studio(images=image_paths, 
+                                                    sftp_client=self.sftp_client, 
+                                                    label_studio_project=self.label_studio_project, 
+                                                    images_to_annotate_dir=self.config.active_learning.image_dir, 
+                                                    folder_name=self.config.label_studio.folder_name, 
+                                                        preannotations=None
+                                                           )
+                    return None
+                    # Select images to upload
+
                 return None
 
             # Given new annotations, propogate labels to nearby images

diff --git a/tests/conftest.py b/tests/conftest.py
@@ -119,5 +119,6 @@ def config(tmpdir_factory):
 
     # Reporting
     cfg.reporting.image_dir = cfg.detection_model.train_image_dir
-
+    cfg.reporting.report_dir = tmpdir_factory.mktemp("reports").strpath
+
     return cfg
diff --git a/tests/test_pipeline.py b/tests/test_pipeline.py
@@ -68,3 +68,13 @@ def test_pipeline_run(config, label_studio_client):
     """Test complete pipeline run"""
     pipeline = Pipeline(cfg=config)
     pipeline.run()
+
+
+@pytest.mark.integration
+def test_first_phase(config, label_studio_client):
+    """Test init phase with no data"""
+    # Set validation csv paths to None
+    config.detection_model.validation_csv_path = None
+    config.classification_model.validation_csv_path = None
+    pipeline = Pipeline(cfg=config)
+    pipeline.run()