Open mfisher87 opened 8 months ago
I am currently working on updating the earthaccess
docs. My focus is on the "Getting Started", "User Guide" and "How Tos". I can think about how we can do this as part of my doc work, so I'll assign it to myself. But it seems like this API documentation might be automated, I will need to learn about this.
Hi, i have the same difficulties. I wanted to download AST L1T data and download it with a filter based on cloud coverage. Is it possible to print a dataframe with the dates and specifications such as Day/ night and cloud coverage? Thanks in advance
Hey @paolodep36, thanks for posting your question! Is there a specific parameter you're expecting to see in the docs for search_data()
? If not, let's move your question to another place so we can focus on it better:
If you think you have a bug on your hands (or would like to request a new feature), can you please open a new issue?
If you're looking for more general support (i.e. you aren't sure whether you found a bug or need help), can you please post on our Q&A discussion board?
More confusion that I think comes from this docs problem: https://github.com/nsidc/earthaccess/discussions/478
In this case the undocumented parameter was point
.
More pain for our users: https://github.com/nsidc/earthaccess/issues/504#issuecomment-2026873110
I'm going to pin this.
search_data
includes undocumented parametercloud_hosted
in its example.downloadable
parameter is not documented.Currently,
kwargs
are passed fromsearch_data
to aDataGranules
object'sparameters
method, which also acceptskwargs
. The possible keys are fixed by the methods onDataGranules
, and populated dynamically by matching keys to method names.https://github.com/nsidc/earthaccess/blob/ff0c59a79003df67265dfbfec8c51e1921739a18/earthaccess/api.py#L116-L119
The same technique is used for both data & dataset search functions.
However, the interface is untyped and the documentation is generated from a hand-maintained docstring. I feel we should move towards fully typing the interface and generating the docs from the annotations. Python's type system feels like a toy at times like this. How do we ensure that a
DataGranules
method exists for each parameter on thesearch_data
interface? Maybe we should instead collect the parameter methods in a dict? Just thinking out loud.