Add create_transcriptions function to call CreateTranscriptions API endpoint

Depends on https://gitlab.com/arkindex/backend/-/merge_requests/1286

Closes #27 (closed)

added 2 commits

• 1a9756d0 - 1 commit from branch master
• 41204d06 - Add create_transcriptions function to call CreateTranscriptions API endpoint

Compare with previous version

This wonderful little worker does not work with use_cache=True, and crashes on a missing ID (when using https://gitlab.com/arkindex/backend/-/merge_requests/1286 on a local backend)

# -*- coding: utf-8 -*-
from arkindex_worker.worker import ElementsWorker
class Demo(ElementsWorker):
    def process_element(self, element):
        print("Demo processing element", element)
        out = self.create_elements(
            parent=element,
            elements=[
                {
                    "name": f"demo {i}",
                    "type": "crash",
                    "polygon": [
                        [10*i, 10*i],
                        [10*i, 100*i],
                        [100*i, 100*i],
                        [100*i, 10*i],
                        [10*i, 10*i],
                    ]
                }
                for i in range(1, 3)
            ]
        )
        from pprint import pprint
        pprint(out)
        out = self.create_transcriptions([
            {
                "element_id": e["id"],
                "text": "plop",
                "score": 0.9
            }
            for e in out
        ])
        pprint(out)
def main():
    Demo(use_cache=True, description="lol").run()
if __name__ == "__main__":
    main()

Traceback:

2021-03-26 10:56:09,565 INFO/arkindex_worker: Connection to local cache /home/bastien/dev/ark/workers/base/db.sqlite established.
2021-03-26 10:56:09,565 INFO/arkindex_worker: Starting ML report for unknown worker
2021-03-26 10:56:11,059 INFO/arkindex_worker: Loaded worker Kaldi Horae revision 48ed481 from API
2021-03-26 10:56:11,268 INFO/arkindex_worker: Processing page 210545209_006_0013.jpg (1de656ab-1f01-4289-9a23-b10bffd5d4a8) (1/1)
Demo processing element page 210545209_006_0013.jpg (1de656ab-1f01-4289-9a23-b10bffd5d4a8)
[{'id': '4a7e229e-2e38-4052-9b90-63cf744172de'},
 {'id': '58d76dbf-3062-4811-a2f4-d6d77b580bf0'}]
{'confidence': 0.9,
 'element_id': '58d76dbf-3062-4811-a2f4-d6d77b580bf0',
 'text': 'plop'}
Traceback (most recent call last):
  File "test.py", line 47, in <module>
    main()
  File "test.py", line 43, in main
    Demo(use_cache=True, description="lol").run()
  File "/home/bastien/dev/ark/workers/base/arkindex_worker/worker.py", line 279, in run
    self.process_element(element)
  File "test.py", line 30, in process_element
    out = self.create_transcriptions([
  File "/home/bastien/dev/ark/workers/base/arkindex_worker/worker.py", line 577, in create_transcriptions
    to_insert = [
  File "/home/bastien/dev/ark/workers/base/arkindex_worker/worker.py", line 579, in <listcomp>
    id=convert_str_uuid_to_hex(created_tr["id"]),
KeyError: 'id'

My bad: an outdated schema.yml was inducing the error !

approved this merge request

merged