Datasets:
Update durhamtrees.py
Browse files- durhamtrees.py +1 -43
durhamtrees.py
CHANGED
@@ -34,7 +34,7 @@ _URLS = {
|
|
34 |
"first_domain1": {
|
35 |
"csv_file": "https://drive.google.com/uc?export=download&id=18HmgMbtbntWsvAySoZr4nV1KNu-i7GCy",
|
36 |
"geojson_file": "https://drive.google.com/uc?export=download&id=1cbn7EY7RofXN7c6Ph0eIGFIZowPZ5lKE",
|
37 |
-
|
38 |
},
|
39 |
"first_domain2": {
|
40 |
"csv_file2": "https://drive.google.com/uc?export=download&id=1RVdaI5dSTPStjhOHO40ypDv2cAQZpi_Y",
|
@@ -63,8 +63,6 @@ class DurhamTrees(GeneratorBasedBuilder):
|
|
63 |
"OBJECTID": Value("int64"),
|
64 |
"X": Value("float64"),
|
65 |
"Y": Value("float64"),
|
66 |
-
"image": Value("binary"),
|
67 |
-
"label": Value("int64"),
|
68 |
"feature1_from_class2": Value("string"),
|
69 |
"streetaddress": Value("string"),
|
70 |
"city": Value("string"),
|
@@ -139,7 +137,6 @@ class DurhamTrees(GeneratorBasedBuilder):
|
|
139 |
"class1_data_file": downloaded_files["first_domain1"]["csv_file"],
|
140 |
"class1_geojson_file": downloaded_files["first_domain1"]["geojson_file"],
|
141 |
"class2_data_file": downloaded_files["first_domain2"]["csv_file2"],
|
142 |
-
"parquet_file": downloaded_files["first_domain1"]["parquet_file"],
|
143 |
"split": Split.TRAIN,
|
144 |
},
|
145 |
),
|
@@ -152,9 +149,6 @@ class DurhamTrees(GeneratorBasedBuilder):
|
|
152 |
class1_examples = list(self._generate_examples_from_class1(class1_data_file, class1_geojson_file))
|
153 |
class2_examples = list(self._generate_examples_from_class2(class2_data_file))
|
154 |
|
155 |
-
# Load Parquet file
|
156 |
-
parquet_data = pq.read_table(parquet_file).to_pandas()
|
157 |
-
class1_examples += list(self._generate_examples_from_parquet(parquet_data))
|
158 |
|
159 |
examples = class1_examples + class2_examples
|
160 |
df = pd.DataFrame(examples)
|
@@ -209,25 +203,6 @@ class DurhamTrees(GeneratorBasedBuilder):
|
|
209 |
yield id_, example
|
210 |
|
211 |
|
212 |
-
def _generate_examples_from_parquet(self, parquet_data):
|
213 |
-
for id_, row in parquet_data.iterrows():
|
214 |
-
# Check if the "image" column is present and not empty
|
215 |
-
if "image" in row and "bytes" in row["image"]:
|
216 |
-
# Decode the base64-encoded image bytes
|
217 |
-
image_data = base64.b64decode(row["image"]["bytes"])
|
218 |
-
example = {"image": image_data, "label": row["label"]}
|
219 |
-
|
220 |
-
# Display the image
|
221 |
-
image_bytes = example.get('image', None)
|
222 |
-
if image_bytes:
|
223 |
-
img = mpimg.imread(io.BytesIO(image_bytes), format='PNG') # Use 'PNG' instead of 'JPG'
|
224 |
-
plt.imshow(img)
|
225 |
-
plt.show()
|
226 |
-
|
227 |
-
yield id_, example
|
228 |
-
else:
|
229 |
-
print(f"Skipping example {id_} as it has missing or invalid image data")
|
230 |
-
|
231 |
|
232 |
|
233 |
def _correlation_analysis(self, df):
|
@@ -250,14 +225,6 @@ durham_trees_dataset.download_and_prepare()
|
|
250 |
# Access the dataset
|
251 |
dataset = durham_trees_dataset.as_dataset()
|
252 |
|
253 |
-
# Iterate through the dataset and display images
|
254 |
-
for example in dataset['train']:
|
255 |
-
if "image" in example and example["image"] is not None and "bytes" in example["image"]:
|
256 |
-
# Display the image
|
257 |
-
image_data = base64.b64decode(example["image"]["bytes"])
|
258 |
-
img = mpimg.imread(io.BytesIO(image_data), format='PNG')
|
259 |
-
plt.imshow(img)
|
260 |
-
plt.show()
|
261 |
|
262 |
# Create an instance of the DurhamTrees class for another configuration
|
263 |
durham_trees_dataset_another = DurhamTrees(name='class2_domain1')
|
@@ -267,12 +234,3 @@ durham_trees_dataset_another.download_and_prepare()
|
|
267 |
|
268 |
# Access the dataset for the new instance
|
269 |
dataset_another = durham_trees_dataset_another.as_dataset()
|
270 |
-
|
271 |
-
# Iterate through the dataset for the new instance and display images
|
272 |
-
for example in dataset_another['train']:
|
273 |
-
if "image" in example and example["image"] is not None and "bytes" in example["image"]:
|
274 |
-
# Display the image
|
275 |
-
image_data = base64.b64decode(example["image"]["bytes"])
|
276 |
-
img = mpimg.imread(io.BytesIO(image_data), format='PNG')
|
277 |
-
plt.imshow(img)
|
278 |
-
plt.show()
|
|
|
34 |
"first_domain1": {
|
35 |
"csv_file": "https://drive.google.com/uc?export=download&id=18HmgMbtbntWsvAySoZr4nV1KNu-i7GCy",
|
36 |
"geojson_file": "https://drive.google.com/uc?export=download&id=1cbn7EY7RofXN7c6Ph0eIGFIZowPZ5lKE",
|
37 |
+
|
38 |
},
|
39 |
"first_domain2": {
|
40 |
"csv_file2": "https://drive.google.com/uc?export=download&id=1RVdaI5dSTPStjhOHO40ypDv2cAQZpi_Y",
|
|
|
63 |
"OBJECTID": Value("int64"),
|
64 |
"X": Value("float64"),
|
65 |
"Y": Value("float64"),
|
|
|
|
|
66 |
"feature1_from_class2": Value("string"),
|
67 |
"streetaddress": Value("string"),
|
68 |
"city": Value("string"),
|
|
|
137 |
"class1_data_file": downloaded_files["first_domain1"]["csv_file"],
|
138 |
"class1_geojson_file": downloaded_files["first_domain1"]["geojson_file"],
|
139 |
"class2_data_file": downloaded_files["first_domain2"]["csv_file2"],
|
|
|
140 |
"split": Split.TRAIN,
|
141 |
},
|
142 |
),
|
|
|
149 |
class1_examples = list(self._generate_examples_from_class1(class1_data_file, class1_geojson_file))
|
150 |
class2_examples = list(self._generate_examples_from_class2(class2_data_file))
|
151 |
|
|
|
|
|
|
|
152 |
|
153 |
examples = class1_examples + class2_examples
|
154 |
df = pd.DataFrame(examples)
|
|
|
203 |
yield id_, example
|
204 |
|
205 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
206 |
|
207 |
|
208 |
def _correlation_analysis(self, df):
|
|
|
225 |
# Access the dataset
|
226 |
dataset = durham_trees_dataset.as_dataset()
|
227 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
228 |
|
229 |
# Create an instance of the DurhamTrees class for another configuration
|
230 |
durham_trees_dataset_another = DurhamTrees(name='class2_domain1')
|
|
|
234 |
|
235 |
# Access the dataset for the new instance
|
236 |
dataset_another = durham_trees_dataset_another.as_dataset()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|