rapidata 0.1.25__py3-none-any.whl → 0.2.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- rapidata/rapidata_client/dataset/rapidata_dataset.py +20 -4
- rapidata/rapidata_client/order/rapidata_order_builder.py +2 -2
- {rapidata-0.1.25.dist-info → rapidata-0.2.0.dist-info}/METADATA +1 -1
- {rapidata-0.1.25.dist-info → rapidata-0.2.0.dist-info}/RECORD +6 -6
- {rapidata-0.1.25.dist-info → rapidata-0.2.0.dist-info}/LICENSE +0 -0
- {rapidata-0.1.25.dist-info → rapidata-0.2.0.dist-info}/WHEEL +0 -0
|
@@ -11,6 +11,8 @@ from rapidata.api_client.models.upload_text_sources_to_dataset_model import (
|
|
|
11
11
|
from rapidata.rapidata_client.metadata.base_metadata import Metadata
|
|
12
12
|
from rapidata.service import LocalFileService
|
|
13
13
|
from rapidata.service.openapi_service import OpenAPIService
|
|
14
|
+
from concurrent.futures import ThreadPoolExecutor, as_completed
|
|
15
|
+
from tqdm import tqdm
|
|
14
16
|
|
|
15
17
|
|
|
16
18
|
class RapidataDataset:
|
|
@@ -32,19 +34,18 @@ class RapidataDataset:
|
|
|
32
34
|
self,
|
|
33
35
|
image_paths: list[str | list[str]],
|
|
34
36
|
metadata: list[Metadata] | None = None,
|
|
37
|
+
max_workers: int = 10,
|
|
35
38
|
):
|
|
36
39
|
if metadata is not None and len(metadata) != len(image_paths):
|
|
37
40
|
raise ValueError(
|
|
38
41
|
"metadata must be None or have the same length as image_paths"
|
|
39
42
|
)
|
|
40
43
|
|
|
41
|
-
|
|
42
|
-
|
|
44
|
+
def upload_datapoint(media_paths_rapid: str | list[str], meta: Metadata | None) -> None:
|
|
43
45
|
if isinstance(media_paths_rapid, list) and not all(
|
|
44
46
|
os.path.exists(media_path) for media_path in media_paths_rapid
|
|
45
47
|
):
|
|
46
48
|
raise FileNotFoundError(f"File not found: {media_paths_rapid}")
|
|
47
|
-
|
|
48
49
|
elif isinstance(media_paths_rapid, str) and not os.path.exists(
|
|
49
50
|
media_paths_rapid
|
|
50
51
|
):
|
|
@@ -60,4 +61,19 @@ class RapidataDataset:
|
|
|
60
61
|
),
|
|
61
62
|
)
|
|
62
63
|
|
|
63
|
-
self.openapi_service.dataset_api.dataset_create_datapoint_post(
|
|
64
|
+
self.openapi_service.dataset_api.dataset_create_datapoint_post(
|
|
65
|
+
model=model,
|
|
66
|
+
files=media_paths_rapid if isinstance(media_paths_rapid, list) else [media_paths_rapid] # type: ignore
|
|
67
|
+
)
|
|
68
|
+
|
|
69
|
+
total_uploads = len(image_paths)
|
|
70
|
+
with ThreadPoolExecutor(max_workers=max_workers) as executor:
|
|
71
|
+
futures = [
|
|
72
|
+
executor.submit(upload_datapoint, media_paths, meta)
|
|
73
|
+
for media_paths, meta in zip_longest(image_paths, metadata or [])
|
|
74
|
+
]
|
|
75
|
+
|
|
76
|
+
with tqdm(total=total_uploads, desc="Uploading datapoints") as pbar:
|
|
77
|
+
for future in as_completed(futures):
|
|
78
|
+
future.result() # This will raise any exceptions that occurred during execution
|
|
79
|
+
pbar.update(1)
|
|
@@ -87,7 +87,7 @@ class RapidataOrderBuilder:
|
|
|
87
87
|
],
|
|
88
88
|
)
|
|
89
89
|
|
|
90
|
-
def create(self, submit=True) -> RapidataOrder:
|
|
90
|
+
def create(self, submit=True, max_workers=10) -> RapidataOrder:
|
|
91
91
|
"""Actually makes the API calls to create the order based on how the order builder was configured.
|
|
92
92
|
|
|
93
93
|
Args:
|
|
@@ -113,7 +113,7 @@ class RapidataOrderBuilder:
|
|
|
113
113
|
openapi_service=self._openapi_service,
|
|
114
114
|
)
|
|
115
115
|
|
|
116
|
-
order.dataset.add_media_from_paths(self._media_paths, self._metadata)
|
|
116
|
+
order.dataset.add_media_from_paths(self._media_paths, self._metadata, max_workers)
|
|
117
117
|
|
|
118
118
|
if submit:
|
|
119
119
|
order.submit()
|
|
@@ -222,7 +222,7 @@ rapidata/rapidata_client/config.py,sha256=tQLgN6k_ATOX1GzZh38At2rgBDLStV6rJ6z0vs
|
|
|
222
222
|
rapidata/rapidata_client/country_codes/__init__.py,sha256=Y8qeG2IMjvMGvhaPydq0nhwRQHb6dQqilctlEXu0_PE,55
|
|
223
223
|
rapidata/rapidata_client/country_codes/country_codes.py,sha256=Q0HMX7uHJQDeLCFPP5bq4iYi6pgcDWEcl2ONGhjgoeU,286
|
|
224
224
|
rapidata/rapidata_client/dataset/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
225
|
-
rapidata/rapidata_client/dataset/rapidata_dataset.py,sha256=
|
|
225
|
+
rapidata/rapidata_client/dataset/rapidata_dataset.py,sha256=QDsl7ZCZxuG02yfEpBSphfSDZh_qHz6m5HUCGwZflfw,3205
|
|
226
226
|
rapidata/rapidata_client/dataset/rapidata_validation_set.py,sha256=YrnUzia9AXgq2z917FtztFxj4fD5EgTWXVPBzLVIujY,9374
|
|
227
227
|
rapidata/rapidata_client/dataset/validation_rapid_parts.py,sha256=SIeQesEXPPOW5kclxYLNWaKllBXHm7DQKBdMU-GXnfc,2104
|
|
228
228
|
rapidata/rapidata_client/dataset/validation_set_builder.py,sha256=B9D-uNCo_PO0NCUHju_7dsWtz_KcOmvFIsxUgQ67Q2M,7471
|
|
@@ -235,7 +235,7 @@ rapidata/rapidata_client/metadata/prompt_metadata.py,sha256=_FypjKWrC3iKUO_G2CVw
|
|
|
235
235
|
rapidata/rapidata_client/metadata/transcription_metadata.py,sha256=THtDEVCON4UlcXHmXrjilaOLHys4TrktUOPGWnXaCcc,631
|
|
236
236
|
rapidata/rapidata_client/order/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
237
237
|
rapidata/rapidata_client/order/rapidata_order.py,sha256=VRDLTPBf2k6UihF0DnWq3nfBLWExfWHzh3T-cJgFF1w,2437
|
|
238
|
-
rapidata/rapidata_client/order/rapidata_order_builder.py,sha256=
|
|
238
|
+
rapidata/rapidata_client/order/rapidata_order_builder.py,sha256=Wx7lhThTd6SwJNSbXzhGCsEgMhizjsriGz2zMjbQyEI,8134
|
|
239
239
|
rapidata/rapidata_client/rapidata_client.py,sha256=z3vz5_GNivnShj7kqii-eUff16rvwSy62zwi8WZqAWo,2776
|
|
240
240
|
rapidata/rapidata_client/referee/__init__.py,sha256=x0AxGCsR6TlDjfqQ00lB9V7QVS9EZCJzweNEIzx42PI,207
|
|
241
241
|
rapidata/rapidata_client/referee/base_referee.py,sha256=bMy7cw0a-pGNbFu6u_1_Jplu0A483Ubj4oDQzh8vu8k,493
|
|
@@ -259,7 +259,7 @@ rapidata/service/local_file_service.py,sha256=pgorvlWcx52Uh3cEG6VrdMK_t__7dacQ_5
|
|
|
259
259
|
rapidata/service/openapi_service.py,sha256=-vrM2jEzQxr9KAerOYkVhpvMEeHwjzRwm9L_VFyzOT0,1537
|
|
260
260
|
rapidata/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
261
261
|
rapidata/utils/image_utils.py,sha256=TldO3eJWG8IhfJjm5MfNGO0mEDm1mQTsRoA0HLU1Uxs,404
|
|
262
|
-
rapidata-0.
|
|
263
|
-
rapidata-0.
|
|
264
|
-
rapidata-0.
|
|
265
|
-
rapidata-0.
|
|
262
|
+
rapidata-0.2.0.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
|
|
263
|
+
rapidata-0.2.0.dist-info/METADATA,sha256=Oymi81UjDR9fM9f-axaTKUWOCnlExpeA1i1voXC7ruU,923
|
|
264
|
+
rapidata-0.2.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
|
|
265
|
+
rapidata-0.2.0.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|