Merge branch 'main' into add-dataset-training

Author: takahiro-tamenishi

README.md

@@ -2583,7 +2583,21 @@ dataset_object = client.find_dataset_object(
     object_name="brushwood_dog.jpg"
 )
 ```
-
+You can find a object of specified revision by version or revision_id.  
+```python
+dataset_object = client.find_dataset_object(
+    dataset_id="YOUR_DATASET_ID",
+    object_name="brushwood_dog.jpg",
+    version="YOUR_VERSION_NAME" # default is "latest"
+)   
+```
+```python
+dataset_object = client.find_dataset_object(
+    dataset_id="YOUR_DATASET_ID",
+    object_name="brushwood_dog.jpg",
+    revision_id="YOUR_REVISION_ID" # 8 characters or more
+)   
+```
 Success response is the same as when created.
 
 ### Get Dataset Object
@@ -2596,7 +2610,7 @@ dataset_objects = client.get_dataset_objects(dataset="YOUR_DATASET_NAME")
 
 The success response is the same as when created, but it is an array.
 
-You can filter by version and tags.
+You can filter by version or revision_id and tags.
 
 ```python
 dataset_objects = client.get_dataset_objects(
@@ -2605,7 +2619,12 @@ dataset_objects = client.get_dataset_objects(
     tags=["cat"],
 )
 ```
-
+```python
+dataset_objects = client.get_dataset_objects(
+    dataset="YOUR_DATASET_NAME",
+    revision_id="YOUR_REVISION_ID" # 8 characters or more
+)
+```
 ### Download Dataset Objects
 
 Download dataset objects in the dataset to specific directories.

examples/create_dataset.py

@@ -0,0 +1,9 @@
+from pprint import pprint
+
+import fastlabel
+
+client = fastlabel.Client()
+
+dataset = client.create_dataset(name="object-detection", license="The MIT License")
+
+pprint(dataset)

examples/create_dataset_object.py

@@ -0,0 +1,12 @@
+from pprint import pprint
+
+import fastlabel
+
+client = fastlabel.Client()
+
+dataset_object = client.create_dataset_object(
+    dataset="YOUR_DATASET_NAME",
+    name="NAME",
+    file_path="FILE_PATH",
+)
+pprint(dataset_object)

examples/create_image_task.py

@@ -0,0 +1,28 @@
+from pprint import pprint
+
+import fastlabel
+
+# Initialize client
+client = fastlabel.Client()
+
+project = "YOUR_PROJECT_SLUG"
+name = "YOUR_DATA_NAME"
+file_path = "YOUR_DATA_FILE_PATH"  # e.g.) ./cat.jpg
+annotations = [
+    {
+        "type": "bbox",
+        "value": "cat",
+        "attributes": [{"key": "kind", "value": "Scottish field"}],
+        "points": [
+            100,  # top-left x
+            100,  # top-left y
+            200,  # bottom-right x
+            200,  # bottom-right y
+        ],
+    }
+]
+
+task_id = client.create_image_task(
+    project=project, name=name, file_path=file_path, annotations=annotations
+)
+pprint(task_id)

examples/delete_dataset.py

@@ -0,0 +1,5 @@
+import fastlabel
+
+client = fastlabel.Client()
+
+client.delete_dataset(dataset_id="YOUR_DATASET_ID")

examples/delete_dataset_object.py

@@ -0,0 +1,7 @@
+import fastlabel
+
+client = fastlabel.Client()
+
+client.delete_dataset_object(
+    dataset_id="YOUR_DATASET_OBJECT_ID", object_name="YOUR_OBJECT_NAME"
+)

examples/download_dataset_objects.py

@@ -0,0 +1,7 @@
+import fastlabel
+
+client = fastlabel.Client()
+
+client.download_dataset_objects(
+    dataset="object-detection", path="./downloads", types=["train", "valid"]
+)

examples/download_images.py

@@ -0,0 +1,42 @@
+import os
+import time
+import urllib.request
+from multiprocessing import Pool, cpu_count
+
+import fastlabel
+
+client = fastlabel.Client()
+IMAGE_DIR = "images"
+PROJECT_SLUG = "YOUR_PROJECT_SLUG"
+
+
+def get_all_tasks() -> list:
+    # Iterate pages until new tasks are empty.
+    all_tasks = []
+    offset = None
+    while True:
+        time.sleep(1)
+
+        tasks = client.get_image_classification_tasks(
+            project=PROJECT_SLUG, limit=1000, offset=offset
+        )
+        all_tasks.extend(tasks)
+
+        if len(tasks) > 0:
+            offset = len(all_tasks)  # Set the offset
+        else:
+            break
+
+    return all_tasks
+
+
+def download_image(task: dict):
+    urllib.request.urlretrieve(task["url"], os.path.join(IMAGE_DIR, task["name"]))
+
+
+if __name__ == "__main__":
+    os.makedirs(IMAGE_DIR, exist_ok=True)
+
+    tasks = get_all_tasks()
+    with Pool(cpu_count()) as p:
+        p.map(download_image, tasks)

examples/find_dataset.py

@@ -0,0 +1,8 @@
+from pprint import pprint
+
+import fastlabel
+
+client = fastlabel.Client()
+
+dataset = client.find_dataset(dataset_id="YOUR_DATASET_ID")
+pprint(dataset)

examples/find_dataset_object.py

@@ -0,0 +1,10 @@
+from pprint import pprint
+
+import fastlabel
+
+client = fastlabel.Client()
+
+dataset_object = client.find_dataset_object(
+    dataset_id="YOUR_DATASET_OBJECT_ID", object_name="YOUR_OBJECT_NAME"
+)
+pprint(dataset_object)