Export Project
Last updated
Last updated
Two types: export a single file in a specific project or export all files in a project.
Use the below API docs reference and follow this .
Please note that the project ID refers to the resource ID that can be accessed from the URL when opening a project.
Please note that this API will only return the latest state of the project.
Query:
Specifically for this, there is a that you can refer to or use. This API will return a zip which consist of the latest state of the project, as well as each labeler's work.
Query:
All the processes above are done asynchronously. To check the export job, you need to request query.
Specific for FILE_STORAGE
method, the fileUrl
response could return 404. If this happens, it means the export result hasn't been uploaded yet. Please wait or polling the query above to know when exactly the fileUrl
is ready and can be downloaded.
The method will be the same and applicable for both types of export. Below is the additional explanation for each method. Don't forget to fill all the required attributes, then follow these hints.
Download
- For the method
attribute, fill it with FILE_STORAGE
.
Email
- For the method
attribute, fill it with EMAIL
.
External Object Storage
- For the method attribute, fill with EXTERNAL_OBJECT_STORAGE
.
- You also need to fill externalObjectStorageParameter
.
DATASAUR_SCHEMA
can be used for any kinds of project.
XLSX
, CSV
JSON_TABULAR
and TSV
for Row and Doc Labeling.
CSV will also work and is compatible for Hugging Face.
TSV_IOB
, TSV_NON_IOB
, JSON_ADVANCED
for Token Labeling.
JSON
for JSON Simplified format.
PLAIN
for exporting only the text (without any labels) of a Token Labeling project.
There is a specific scenario to consider with Token Labeling. Since Comprehend can only reference to a file on S3, it is important to export the text after making any edits by doing another export with PLAIN
extension as explained above. This ensures that you can correctly reference the annotation data.
See how to get the value .
There are four different methods to obtain the export result, i.e. download, email, webhook, and external object storage. These methods are fully .
For the API payload reference, it can be accessed .
Webhook
- For the method
attribute, fill it with CUSTOM_WEBHOOK
.
- You also need to fill url
and secret
attribute.
- The detailed explanation can be seen .
CUSTOM
for export using .
AMAZON_COMPREHEND_CSV
can be used for both and .
AZURE_AUTOML_CSV
can be used for both .
GCP_VERTEX_AI_CSV
can be used for both and .