Tomatillo commited on
Commit
774092c
·
verified ·
1 Parent(s): c3a4698

updated get_samples by Tom S

Browse files
Files changed (1) hide show
  1. src/get_labels_from_samples.py +6 -9
src/get_labels_from_samples.py CHANGED
@@ -1,19 +1,16 @@
1
  from segments import SegmentsClient
 
2
 
3
  def get_samples(client, dataset_identifier):
 
 
 
4
  page = 1
5
  per_page = 1000
6
  samples = []
7
- while True:
8
  response = client.get_samples(dataset_identifier, per_page=per_page, page=page)
9
- # Handle both paginated and direct list responses
10
- if isinstance(response, list):
11
- samples.extend(response)
12
- break # No pagination in this case
13
- else:
14
- samples.extend(response.results)
15
- if not response.has_next:
16
- break
17
  page += 1
18
  return samples
19
 
 
1
  from segments import SegmentsClient
2
+ import math
3
 
4
  def get_samples(client, dataset_identifier):
5
+ dataset = client.get_dataset(dataset_identifier)
6
+ samples_count = getattr(dataset, "samples_count", 0)
7
+ total_pages = math.ceil(samples_count / 1000)
8
  page = 1
9
  per_page = 1000
10
  samples = []
11
+ while page <= total_pages:
12
  response = client.get_samples(dataset_identifier, per_page=per_page, page=page)
13
+ samples.extend(response)
 
 
 
 
 
 
 
14
  page += 1
15
  return samples
16