Add support for async DocParse calls in aryn-sdk
#1116
Conversation
|
Some of the aryn-sdk tests currently check for exact matches of DocParse partitioned output. Since floating point calculations are nondeterministic, we will want to fix this in a future pr so we don't falsely think a regression occurred. |
…loop in multi-async call example
…c_with_unsupported_file_format`
| A dictionary containing "job_id" which can be used with the `partition_file_result_async` | ||
| function to get the results. | ||
|
|
||
| Single Job Example: |
There was a problem hiding this comment.
There's still too much and it's too duplicated. It's not standard to have multiple examples in a function's docstring. If an example is needed at all, make it the simple one, and leave out the obvious imports. We don't need both use_ocr and extract_table_structure demonstrated.
The multi-job example is more complicated and such things are usually included only in a user guide or README.
The problem here is that when people encounter a wall of text, they don't read it. They either skip it, or assume the thing is too complicated and run away. The secondary problem is maintenance. Engineers modifying code often neglect to update the docstrings, and the bigger the job, the more they neglect. Updating the README will be even less likely, but there's a chance if the docstring says "For examples of usage see README.md". When documentation (or code) is duplicated, there's a high chance it will diverge.
|
Further work on async DocParse via the |
No description provided.