dataverse-sdk 2.7.0__tar.gz → 2.8.0__tar.gz

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.
Files changed (40) hide show
  1. dataverse_sdk-2.8.0/MANIFEST.in +2 -0
  2. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/PKG-INFO +286 -8
  3. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/README.md +285 -7
  4. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/__init__.py +14 -0
  5. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/apis/backend.py +39 -0
  6. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/client.py +510 -13
  7. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/schemas/api.py +54 -1
  8. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/schemas/client.py +143 -19
  9. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/schemas/common.py +86 -0
  10. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk.egg-info/PKG-INFO +286 -8
  11. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk.egg-info/SOURCES.txt +7 -1
  12. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/setup.py +1 -1
  13. dataverse_sdk-2.8.0/tools/export_dataslice.py +118 -0
  14. dataverse_sdk-2.8.0/tools/export_dataslice_large.py +133 -0
  15. dataverse_sdk-2.8.0/tools/import_dataset_from_local.py +176 -0
  16. dataverse_sdk-2.8.0/tools/import_vqa_dataset.py +166 -0
  17. dataverse_sdk-2.8.0/tools/upload_videos_create_session.py +135 -0
  18. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/apis/__init__.py +0 -0
  19. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/apis/third_party.py +0 -0
  20. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/connections.py +0 -0
  21. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/constants.py +0 -0
  22. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/exceptions/__init__.py +0 -0
  23. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/exceptions/client.py +0 -0
  24. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/__init__.py +0 -0
  25. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/base.py +0 -0
  26. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/coco.py +0 -0
  27. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/constant.py +0 -0
  28. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/exporter.py +0 -0
  29. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/utils.py +0 -0
  30. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/visionai.py +0 -0
  31. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/vqa.py +0 -0
  32. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/export/yolo.py +0 -0
  33. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/schemas/__init__.py +0 -0
  34. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/schemas/format.py +0 -0
  35. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/utils/__init__.py +0 -0
  36. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk/utils/utils.py +0 -0
  37. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk.egg-info/dependency_links.txt +0 -0
  38. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk.egg-info/requires.txt +0 -0
  39. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/dataverse_sdk.egg-info/top_level.txt +0 -0
  40. {dataverse_sdk-2.7.0 → dataverse_sdk-2.8.0}/setup.cfg +0 -0
@@ -0,0 +1,2 @@
1
+ # tools/ is not a package, so the sdist is the only artifact that can carry the scripts.
2
+ recursive-include tools *.py
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: dataverse-sdk
3
- Version: 2.7.0
3
+ Version: 2.8.0
4
4
  Summary: Dataverse SDK For Python
5
5
  Home-page:
6
6
  Author: LinkerVision
@@ -32,8 +32,10 @@ Use Dataverse-SDK for Python to help you to interact with the Dataverse platform
32
32
  - Get Project by project-id
33
33
  - Create Dataset from your AWS storage or local
34
34
  - Get Dataset by dataset-id
35
+ - Create Dataslice from a whole dataset, or look one up by name
35
36
  - List models for your selected project-id
36
37
  - Get and download your model
38
+ - Convert a model to ONNX/TensorRT and read the convert record's metrics
37
39
 
38
40
  [Package (PyPi)](https://pypi.org/project/dataverse-sdk/) | [Source code](https://github.com/linkervision/dataverse-sdk)
39
41
 
@@ -104,9 +106,14 @@ The following sections provide examples for the most common DataVerse tasks incl
104
106
  * [Create Dataset](#create-dataset)
105
107
  * [List Dataset](#list-and-get-dataset)
106
108
  * [List Dataslices](#list-and-get-dataslices)
109
+ * [Get Dataslice by Name](#get-dataslice-by-name)
110
+ * [Create Dataslice from Dataset](#create-dataslice-from-dataset)
107
111
  * [Export Dataslice](#export-dataslice-and-download)
108
112
  * [List Models](#list-models)
109
113
  * [Get and Download Model](#get-model)
114
+ * [Get Convert Model File](#get-convert-model-file)
115
+ * [Convert Model to ONNX/TRT](#convert-model-to-onnxtrt)
116
+ * [Retrieve Convert Record](#retrieve-convert-record-and-its-metrics)
110
117
  * [Create VQA Project](#create-vqa-project)
111
118
  * [Edit VQA Ontology](#edit-vqa-ontology)
112
119
  * [Get Question List](#get-question-list)
@@ -199,9 +206,9 @@ project = client.create_project(name="Sample project", ontology=ontology, sensor
199
206
 
200
207
  | Argument name | Type/Options | Default | Description |
201
208
  | :--- | :--- | :--- | :--- |
202
- | name | str | *-- | name of your project |
203
- | ontology | Ontology | *-- | the Ontology basemodel data of current project |
204
- | sensors | list[Sensor] | *-- | the list of Sensor basemodel data of your project |
209
+ | name | str | *-- | name of your project |
210
+ | ontology | Ontology | *-- | the Ontology basemodel data of current project |
211
+ | sensors | list[Sensor] | *-- | the list of Sensor basemodel data of your project |
205
212
  | project_tag | ProjectTag | None | your project tags |
206
213
  | description | str | None | your project description |
207
214
 
@@ -445,6 +452,87 @@ dataslice_data = client.get_dataslice(dataslice_id=504)
445
452
 
446
453
  ```
447
454
 
455
+ ### Get Dataslice by Name
456
+
457
+ Dataslice names are unique across the site, so a name identifies exactly one dataslice.
458
+
459
+ ```Python
460
+ dataslice = client.get_dataslice_by_name(project_id=101, dataslice_name="my-dataslice")
461
+ # OR from the project object
462
+ project = client.get_project(project_id=101)
463
+ dataslice = project.get_dataslice_by_name(dataslice_name="my-dataslice")
464
+
465
+ dataslice.id
466
+ dataslice.status # DataSliceStatus
467
+ dataslice.image_count
468
+ ```
469
+
470
+ `DataSliceStatus`: `CREATING`, `CREATING_FAIL`, `READY`, `ANNOTATION_UPDATING`,
471
+ `IQA_UPDATING`, `TAGGING_UPDATING`, `DELETING`.
472
+
473
+ `image_count` and `pcd_count` read the per-type datarow counts in `dataslice.metadata`.
474
+ Which of them carries a number follows the slice's own `type`:
475
+
476
+ | `dataslice.type` | `image_count` | `pcd_count` |
477
+ | -------------------- | ------------- | ----------- |
478
+ | `image` | images | `None` |
479
+ | `pcd` | `None` | pcds |
480
+ | `frame` / `sequence` | images | pcds |
481
+
482
+ The last row is the fused camera-plus-lidar case, where both are counted and both can be
483
+ non-zero. `None` means "this slice holds no datarow of that type", never zero.
484
+
485
+ `get_dataslice` and `get_dataslice_by_name` return `metadata`; `list_dataslices` answers
486
+ with plain dicts carrying only `id`, `name` and `file_count` — the slice's image and pcd
487
+ datarows counted together.
488
+
489
+ This is the only method that takes a name: everything downstream — `convert_model`
490
+ included — takes ids, so this is how a name becomes one.
491
+
492
+ Raises `ValueError` when the project holds no dataslice of that name.
493
+
494
+ ### Create Dataslice from Dataset
495
+
496
+ `create_dataslice_from_dataset` turns a whole dataset into a dataslice — the equivalent of
497
+ opening it in Data Visualization and hitting **Save Data Slice** without narrowing the
498
+ selection. It is **asynchronous**: the returned dataslice starts out `creating` with no
499
+ file count, and turns `ready` once the datarows are in.
500
+
501
+ ```Python
502
+ dataslice = client.create_dataslice_from_dataset(
503
+ project_id=101,
504
+ dataset_id=5,
505
+ dataslice_name="whole-dataset-slice",
506
+ )
507
+ # OR from the project object
508
+ dataslice = project.create_dataslice_from_dataset(
509
+ dataset_id=5, dataslice_name="whole-dataset-slice"
510
+ )
511
+
512
+ # Poll until it settles
513
+ import time
514
+
515
+ from dataverse_sdk import DataSliceStatus
516
+
517
+ SETTLED = (DataSliceStatus.READY, DataSliceStatus.CREATING_FAIL)
518
+ while dataslice.status not in SETTLED:
519
+ time.sleep(10)
520
+ dataslice = client.get_dataslice(dataslice_id=dataslice.id)
521
+ print(dataslice.status, dataslice.image_count) # "ready" 11
522
+ ```
523
+
524
+ A convert record polls the same way through `get_convert_record`, with
525
+ `ConvertRecordStatus.READY` and `ConvertRecordStatus.FAILED` as its settled states.
526
+
527
+ | Argument name | Type | Default | Description |
528
+ | -------------- | ---- | ------- | -------------------------------------------------- |
529
+ | project_id | int | *-- | Project the dataset belongs to |
530
+ | dataset_id | int | *-- | Every datarow of this dataset goes into the slice |
531
+ | dataslice_name | str | *-- | Must be unused **site-wide** |
532
+ | description | str | None | |
533
+
534
+ `*--`: required argument without default
535
+
448
536
  ### Export Dataslice and Download
449
537
  ```Python
450
538
  # Trigger export and get export record id
@@ -454,6 +542,12 @@ client.download_export_dataslice_data(dataslice_id=504, export_record_id=export_
454
542
 
455
543
  ```
456
544
 
545
+ > **Note:** Direct export via `export_dataslice` is limited to 1,000 records. For
546
+ > dataslices exceeding this limit, use
547
+ > [`tools/export_dataslice_large.py`](tools/export_dataslice_large.py) instead
548
+ > (see [Export Large Dataslice and download files](#export-large-dataslice-and-download-files)),
549
+ > which exports the dataslice in batches.
550
+
457
551
 
458
552
  ### List Models
459
553
 
@@ -470,6 +564,16 @@ project = client.get_project(project_id=1)
470
564
  models = project.list_models()
471
565
  ```
472
566
 
567
+ Each model carries a `status`, which is what a listing is usually narrowed by:
568
+
569
+ ```Python
570
+ from dataverse_sdk import MLModelStatus
571
+
572
+ ready = [model for model in models if model.status == MLModelStatus.READY]
573
+ ```
574
+
575
+ `MLModelStatus`: `PROCESSING`, `READY`, `DELETING`.
576
+
473
577
  #### Filtering by Model Type
474
578
 
475
579
  You can filter models by type using strings or lists of strings. The SDK supports multiple model types:
@@ -501,7 +605,7 @@ models = client.list_models(
501
605
 
502
606
  | Argument name | Type/Options | Default | Description |
503
607
  | ------------- | ----------------------------------------------------------------- | ------------------- | ------------------------ |
504
- | project_id | int | \*-- | The project ID |
608
+ | project_id | int | *-- | The project ID |
505
609
  | client_alias | str | None | The client alias |
506
610
  | type | "trained", "byom", "uploaded", list["trained", "byom", "uploaded] | ["trained", "byom"] | Model types to filter by |
507
611
 
@@ -571,7 +675,7 @@ Pass an explicit `save_path` whenever you need a predictable location — it alw
571
675
 
572
676
  | Argument name | Type/Options | Default | Description |
573
677
  | ----------------- | -------------------------------------------------------------- | ----------------------------- | -------------------------------------------------------------------- |
574
- | convert_record_id | int | \*-- | The convert record to download from |
678
+ | convert_record_id | int | *-- | The convert record to download from |
575
679
  | file_type | ConvertModelFileType, "triton", "model", "raw_onnx", "calib_cache" | ConvertModelFileType.TRITON | Which artifact to download |
576
680
  | save_path | str | None | Local path to write the file; when omitted the server's filename is used (see above) |
577
681
  | timeout | int | 3000 | Maximum timeout of the request |
@@ -588,6 +692,180 @@ Pass an explicit `save_path` whenever you need a predictable location — it alw
588
692
  <br>
589
693
 
590
694
 
695
+ ### Convert Model to ONNX/TRT
696
+
697
+ `convert_model` starts a model conversion. It is **asynchronous**: the call returns the
698
+ ids of the created convert records, and the conversion itself finishes later.
699
+
700
+ ```Python
701
+ from dataverse_sdk import ConvertFormat, ConvertPrecision, QuantizationMethod
702
+
703
+ # Method 1: Using client -- model 6418 is a YOLOv9, which is NMS-based
704
+ record_ids = client.convert_model(
705
+ model_id=6418,
706
+ name="onnx-fp16",
707
+ target_dataslice_id=504, # ids only; see Get Dataslice by Name
708
+ model_type=ConvertFormat.ONNX,
709
+ data_type=ConvertPrecision.FP16,
710
+ confidence_score=10,
711
+ nms_threshold=50,
712
+ client_alias=client.alias,
713
+ )
714
+
715
+ # Method 2: Using the model object -- model 7201 is a D-FINE, which is NMS-free
716
+ model = client.get_model(model_id=7201, client_alias=client.alias)
717
+ record_ids = model.convert(
718
+ name="trt-int8-ptq",
719
+ target_dataslice_id=504,
720
+ model_type=ConvertFormat.TRT,
721
+ data_type=ConvertPrecision.INT8,
722
+ quantizations=[QuantizationMethod.PTQ],
723
+ quantize_dataslice_id=508,
724
+ )
725
+ ```
726
+
727
+ #### Input Arguments
728
+
729
+ | Argument name | Type/Options | Default | Description |
730
+ | ------------------ | ------------------------------------- | --------- | ------------------------------------------------------------------------------- |
731
+ | model_id | int | *-- | The source model to convert |
732
+ | name | str | *-- | Convert record name; must be unused **under this model** |
733
+ | target_dataslice_id| int | *-- | Dataslice the converted model is evaluated on |
734
+ | model_type | ConvertFormat \| str | *-- | Target format |
735
+ | data_type | ConvertPrecision \| str | *-- | Numeric precision; `fp32` is D-FINE only |
736
+ | confidence_score | int | 10 | Confidence threshold, 10-90 |
737
+ | iou | int | 50 | IoU threshold, 1-99 |
738
+ | topk | int | None | 50-300; 300 for D-FINE, 100 for every other structure |
739
+ | main_obj_low | int | 1024 | Main object size lower bound |
740
+ | main_obj_high | int | 9216 | Main object size upper bound |
741
+ | resolution_width | int | None | Defaults to the source model's own resolution |
742
+ | resolution_height | int | None | Defaults to the source model's own resolution |
743
+ | nms_threshold | int | None | 10-90; required for every structure except D-FINE |
744
+ | nms_class_agnostic | bool | None | **Rejected for D-FINE** |
745
+ | machine_type | str | None | |
746
+ | quantize_dataslice_id | int | None | Calibration dataslice; has to be given together with `quantizations` |
747
+ | quantizations | list[QuantizationMethod \| str] | None | Exactly one method; **required for `int8`, rejected for every other `data_type`**; only `PTQ` is supported for D-FINE |
748
+ | model_structure | ModelStructure \| str | None | The source model's architecture; sets the topk default and the D-FINE rules. Read back from the model when omitted |
749
+
750
+ `*--`: required argument without default
751
+
752
+ Every enum below is a `str` enum, so the member and its plain string are interchangeable:
753
+
754
+ | Enum | Members | Strings |
755
+ | -------------------- | ------------------------------------------------ | ------------------------------------------------------------------------------------------- |
756
+ | `ConvertFormat` | `ONNX`, `TRT` | `"onnx"`, `"trt"` |
757
+ | `ConvertPrecision` | `FP32`, `FP16`, `INT8` | `"fp32"`, `"fp16"`, `"int8"` |
758
+ | `QuantizationMethod` | `PTQ`, `QAT_TRAIN`, `QAT_DISTILL` | `"ptq"`, `"qat_train"`, `"qat_distill"` |
759
+ | `ModelStructure` | `YOLOV9_C`, `YOLOV9_E`, `YOLOV9_S` | `"yolov9-c"`, `"yolov9-e"`, `"yolov9-s"` |
760
+ | | `DFINE_N`, `DFINE_S`, `DFINE_M`, `DFINE_L`, `DFINE_X` | `"dfine-n"`, `"dfine-s"`, `"dfine-m"`, `"dfine-l"`, `"dfine-x"` |
761
+
762
+ `convert_model`, `get_convert_record`, `get_convert_record_by_name` and
763
+ `list_convert_records` are static methods, so they need `client_alias` (or `client`)
764
+ passed in; without either they raise `ValueError`. The dataslice methods above are
765
+ instance methods and fall back to the client's own alias.
766
+
767
+ Supported resolutions are `640x480`, `640x640`, `1024x576`, `1024x768` and `1024x1024`; an unsupported one raises `APIValidationError` without sending the request.
768
+
769
+ The server decides the rest, so these can change; at the time of writing:
770
+
771
+ | `model_structure` | `data_type` | `model_type` | `quantizations` |
772
+ | ----------------- | ----------- | --------------- | --------------------------------- |
773
+ | yolov9 | `fp16` | `onnx` or `trt` | none |
774
+ | yolov9 | `int8` | `trt` | `ptq`, `qat_train`, `qat_distill` |
775
+ | D-FINE | `fp32` | `onnx` | none |
776
+ | D-FINE | `fp16` | `trt` | none |
777
+ | D-FINE | `int8` | `trt` | `ptq` |
778
+
779
+ `int8` must carry exactly one quantization method and every other `data_type` must
780
+ carry none -- that holds for both families, so a yolov9 `int8` convert without
781
+ `quantizations` raises `APIValidationError` instead of reaching the backend. The
782
+ method itself is what differs: D-FINE's `int8` only takes `ptq`, while yolov9 takes
783
+ any of the three. A structure the SDK does not know is left to the server.
784
+
785
+ #### Return
786
+
787
+ `list[int]` — ids of the created convert records, one per quantization method. The
788
+ backend takes only one quantization method for now, so today the list always holds
789
+ exactly one id.
790
+
791
+ #### Keeping it to one request
792
+
793
+ Both dataslices are given by id, so the only extra request left is the read-back of the
794
+ source model — and passing `model_structure` **together with both resolutions** skips
795
+ that too, which is what `model.convert()` does with the three it already holds. Chaining
796
+ from objects you already have is therefore the cheapest form:
797
+
798
+ ```Python
799
+ model = client.get_model(model_id=6418, client_alias=client.alias)
800
+ dataslice = project.get_dataslice_by_name("my-eval-dataslice")
801
+
802
+ record_ids = model.convert(
803
+ name="onnx-fp16",
804
+ target_dataslice_id=dataslice.id,
805
+ model_type=ConvertFormat.ONNX,
806
+ data_type=ConvertPrecision.FP16,
807
+ nms_threshold=50,
808
+ )
809
+
810
+ record = client.get_convert_record(convert_record_id=record_ids[0], client_alias=client.alias)
811
+ record.status
812
+ ```
813
+
814
+ <br>
815
+
816
+
817
+ ### Retrieve Convert Record and its Metrics
818
+
819
+ A convert name is only unique **per source model** — two models may each own a record
820
+ called `onnx-fp16` — so a lookup by name takes the model alongside it.
821
+
822
+ ```Python
823
+ # By id
824
+ record = client.get_convert_record(convert_record_id=2212, client_alias=client.alias)
825
+
826
+ # By name, under a given model
827
+ record = client.get_convert_record_by_name(
828
+ model_id=6418, convert_name="onnx-fp16", client_alias=client.alias
829
+ )
830
+ # OR from the model object
831
+ model = client.get_model(model_id=6418, client_alias=client.alias)
832
+ record = model.get_convert_record_by_name(convert_name="onnx-fp16")
833
+
834
+ record.status # ConvertRecordStatus
835
+ record.id
836
+ record.model_type # "onnx" / "trt" -- reads configuration["format"]
837
+ record.data_type # "fp32"/"fp16"/"int8" -- reads configuration["precision"]
838
+ record.f1_score # and .precision, .recall, .map_5_95
839
+ ```
840
+
841
+ `record.configuration` carries every other convert setting. The four metrics read as
842
+ `0.0` until the conversion finishes, and `None` when the server did not return them at all
843
+ — mind that `record.precision` is one of them, and the precision the model was converted
844
+ at is `record.data_type`.
845
+
846
+ Looking a record up by name, filtering a listing, and reading the ids back from
847
+ `convert_model` all need a Dataverse recent enough to serialise them; against an older
848
+ site the SDK falls back to filtering the full listing itself.
849
+
850
+ #### List Convert Records
851
+
852
+ ```Python
853
+ # Every convert record in a project
854
+ records = client.list_convert_records(project_id=1992, client_alias=client.alias)
855
+ # OR
856
+ records = project.list_convert_records()
857
+
858
+ # Narrow by model / name / status
859
+ records = client.list_convert_records(
860
+ model_id=6418, status="ready", client_alias=client.alias
861
+ )
862
+ # OR
863
+ records = model.list_convert_records(status="ready")
864
+ ```
865
+
866
+ <br>
867
+
868
+
591
869
  ### Create VQA Project
592
870
 
593
871
  The `create_vqa_project` method will create project on the connected site with the defined questions/answer_type.
@@ -610,8 +888,8 @@ project = client.create_vqa_project(name="vqa-project", sensor_name="camera1", o
610
888
 
611
889
  | Argument name | Type/Options | Default | Description |
612
890
  | :--- | :--- | :--- | :--- |
613
- | name | str | *-- | name of your project |
614
- | sensor_name | str | *-- | the camera sensor name |
891
+ | name | str | *-- | name of your project |
892
+ | sensor_name | str | *-- | the camera sensor name |
615
893
  | ontology_name | str | *-- | the ontology name |
616
894
  | question_answer| list[QuestionClass] | *-- | your question/answer_type. `QuestionClass.answer_type` valid values: `boolean`, `option`, `number`, `text` |
617
895
  | description | str | None | your project description |