Update new_dataset_script.py
Browse files- new_dataset_script.py +48 -83
new_dataset_script.py
CHANGED
@@ -106,95 +106,58 @@ class NewDataset(datasets.GeneratorBasedBuilder):
|
|
106 |
citation=_CITATION,
|
107 |
)
|
108 |
|
109 |
-
|
110 |
def _split_generators(self, dl_manager):
|
111 |
-
|
112 |
-
|
113 |
-
|
114 |
-
|
115 |
-
|
116 |
-
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
122 |
-
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
# Shuffle the list for random split
|
127 |
-
random.seed(42) # Set a random seed for reproducibility
|
128 |
-
random.shuffle(all_image_filenames)
|
129 |
-
|
130 |
-
# Split the files into train/validation/test
|
131 |
-
num_files = len(all_image_filenames)
|
132 |
-
train_split_end = int(num_files * 0.7)
|
133 |
-
val_split_end = train_split_end + int(num_files * 0.15)
|
134 |
-
|
135 |
-
train_files = all_image_filenames[:train_split_end]
|
136 |
-
val_files = all_image_filenames[train_split_end:val_split_end]
|
137 |
-
test_files = all_image_filenames[val_split_end:]
|
138 |
-
|
139 |
-
return [
|
140 |
-
datasets.SplitGenerator(
|
141 |
name=datasets.Split.TRAIN,
|
142 |
gen_kwargs={
|
143 |
-
"filepaths":
|
144 |
-
"species_info": species_info,
|
145 |
-
"data_dir": extracted_images_path,
|
146 |
-
"split": "train",
|
147 |
-
},
|
148 |
-
),
|
149 |
-
datasets.SplitGenerator(
|
150 |
-
name=datasets.Split.VALIDATION,
|
151 |
-
gen_kwargs={
|
152 |
-
"filepaths": val_files,
|
153 |
"species_info": species_info,
|
154 |
"data_dir": extracted_images_path,
|
155 |
-
"split": "validation",
|
156 |
},
|
157 |
-
)
|
158 |
-
datasets.SplitGenerator(
|
159 |
-
name=datasets.Split.TEST,
|
160 |
-
gen_kwargs={
|
161 |
-
"filepaths": test_files,
|
162 |
-
"species_info": species_info,
|
163 |
-
"data_dir": extracted_images_path,
|
164 |
-
"split": "test",
|
165 |
-
},
|
166 |
-
),
|
167 |
-
]
|
168 |
|
169 |
-
|
170 |
|
171 |
# ... other necessary imports and class definitions
|
172 |
def _parse_yolo_labels(self, label_path, width, height):
|
173 |
-
|
174 |
-
|
175 |
-
|
176 |
-
|
177 |
-
|
178 |
-
|
179 |
-
|
180 |
-
|
181 |
-
|
182 |
-
|
183 |
-
|
184 |
-
|
185 |
-
|
186 |
-
|
187 |
-
|
188 |
-
|
189 |
-
|
190 |
-
|
191 |
-
|
192 |
-
|
193 |
-
|
194 |
-
def _generate_examples(self, filepaths, species_info, data_dir
|
195 |
-
|
196 |
for file_name in filepaths:
|
197 |
-
image_id = os.path.splitext(file_name)[0]
|
198 |
image_path = os.path.join(data_dir, f"{image_id}.jpg")
|
199 |
label_path = os.path.join(data_dir, f"{image_id}.txt")
|
200 |
|
@@ -207,12 +170,14 @@ class NewDataset(datasets.GeneratorBasedBuilder):
|
|
207 |
scientific_name = species_row['ScientificName'].values[0] if not species_row.empty else None
|
208 |
|
209 |
annotations = self._parse_yolo_labels(label_path, width, height)
|
210 |
-
|
211 |
-
|
|
|
212 |
"image_id": image_id,
|
213 |
"species": species,
|
214 |
"scientific_name": scientific_name,
|
215 |
-
"
|
216 |
-
"image_resolution": {"width": width, "height": height},
|
217 |
"annotations": annotations
|
218 |
-
}
|
|
|
|
|
|
106 |
citation=_CITATION,
|
107 |
)
|
108 |
|
|
|
109 |
def _split_generators(self, dl_manager):
|
110 |
+
# Only download data, no need to split
|
111 |
+
data_files = dl_manager.download_and_extract({
|
112 |
+
"csv": "https://huggingface.co/datasets/XintongHe/Populus_Stomatal_Images_Datasets/resolve/main/Labeled Stomatal Images.csv",
|
113 |
+
"zip": "https://huggingface.co/datasets/XintongHe/Populus_Stomatal_Images_Datasets/resolve/main/Labeled Stomatal Images.zip"
|
114 |
+
})
|
115 |
+
|
116 |
+
species_info = pd.read_csv(data_files["csv"])
|
117 |
+
extracted_images_path = os.path.join(data_files["zip"], "Labeled Stomatal Images")
|
118 |
+
|
119 |
+
# Get all image filenames
|
120 |
+
all_image_filenames = species_info['FileName'].apply(lambda x: x + '.jpg').tolist()
|
121 |
+
|
122 |
+
# No longer need to randomize and split the dataset
|
123 |
+
return [datasets.SplitGenerator(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
124 |
name=datasets.Split.TRAIN,
|
125 |
gen_kwargs={
|
126 |
+
"filepaths": all_image_filenames,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
127 |
"species_info": species_info,
|
128 |
"data_dir": extracted_images_path,
|
|
|
129 |
},
|
130 |
+
)]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
131 |
|
132 |
+
|
133 |
|
134 |
# ... other necessary imports and class definitions
|
135 |
def _parse_yolo_labels(self, label_path, width, height):
|
136 |
+
annotations = []
|
137 |
+
with open(label_path, 'r') as file:
|
138 |
+
yolo_data = file.readlines()
|
139 |
+
|
140 |
+
for line in yolo_data:
|
141 |
+
class_id, x_center_rel, y_center_rel, width_rel, height_rel = map(float, line.split())
|
142 |
+
x_min = (x_center_rel - width_rel / 2) * width
|
143 |
+
y_min = (y_center_rel - height_rel / 2) * height
|
144 |
+
x_max = (x_center_rel + width_rel / 2) * width
|
145 |
+
y_max = (y_center_rel + height_rel / 2) * height
|
146 |
+
annotations.append({
|
147 |
+
"category_id": int(class_id),
|
148 |
+
"bounding_box": {
|
149 |
+
"x_min": x_min,
|
150 |
+
"y_min": y_min,
|
151 |
+
"x_max": x_max,
|
152 |
+
"y_max": y_max
|
153 |
+
}
|
154 |
+
})
|
155 |
+
return annotations
|
156 |
+
|
157 |
+
def _generate_examples(self, filepaths, species_info, data_dir):
|
158 |
+
"""Yields examples as (key, example) tuples."""
|
159 |
for file_name in filepaths:
|
160 |
+
image_id = os.path.splitext(file_name)[0]
|
161 |
image_path = os.path.join(data_dir, f"{image_id}.jpg")
|
162 |
label_path = os.path.join(data_dir, f"{image_id}.txt")
|
163 |
|
|
|
170 |
scientific_name = species_row['ScientificName'].values[0] if not species_row.empty else None
|
171 |
|
172 |
annotations = self._parse_yolo_labels(label_path, width, height)
|
173 |
+
|
174 |
+
# Create a structured object that includes the image and its metadata
|
175 |
+
img_with_metadata_and_annotations = {
|
176 |
"image_id": image_id,
|
177 |
"species": species,
|
178 |
"scientific_name": scientific_name,
|
179 |
+
"image": img, # Assuming you want to keep the PIL Image object; otherwise use pics_array
|
|
|
180 |
"annotations": annotations
|
181 |
+
}
|
182 |
+
|
183 |
+
yield image_id, img_with_metadata_and_annotations
|