Awesome

STAC Task (stac-task)

• Quickstart for Creating New Tasks
• Task Input
  • ProcessDefinition Object
    • UploadOptions Object
      • path_template
      • collections
    • tasks
    • TaskConfig Object
• Full Process Definition Example
• Migration
  • 0.4.x -> 0.5.x
• Development
• Contributing

This Python library consists of the Task class, which is used to create custom tasks based on a "STAC In, STAC Out" approach. The Task class acts as wrapper around custom code and provides several convenience methods for modifying STAC Items, creating derived Items, and providing a CLI.

This library is based on a branch of cirrus-lib except aims to be more generic.

Quickstart for Creating New Tasks

from typing import Any

from stactask import Task, DownloadConfig

class MyTask(Task):
    name = "my-task"
    description = "this task does it all"

    def validate(self, payload: dict[str, Any]) -> bool:
        return len(self.items) == 1

    def process(self, **kwargs: Any) -> list[dict[str, Any]]:
        item = self.items[0]

        # download a datafile
        item = self.download_item_assets(
            item,
            config=DownloadConfig(include=['data'])
        )

        # operate on the local file to create a new asset
        item = self.upload_item_assets_to_s3(item)

        # this task returns a single item
        return [item.to_dict(include_self_link=True, transform_hrefs=False)]

Task Input

ProcessDefinition Object

A STAC task can be provided additional configuration via the 'process' field in the input ItemCollection.

UploadOptions Object

path_template

The path_template string is a way to control the output location of uploaded assets from a STAC Item using metadata from the Item itself. The template can contain fixed strings along with variables used for substitution. See the PySTAC documentation for LayoutTemplate for a list of supported template variables and their meaning.

collections

The collections dictionary provides a collection ID and JSONPath pattern for matching against STAC Items. At the end of processing, before the final STAC Items are returned, the Task class can be used to assign all of the Items to specific collection IDs. For each Item the JSONPath pattern for all collections will be compared. The first match will cause the Item's Collection ID to be set to the provided value.

For example:

"collections": {
    "landsat-c2l2": "$[?(@.id =~ 'LC08.*')]"
}

In this example, the task will set any STAC Items that have an ID beginning with "LC08" to the landsat-c2l2 collection.

See JSONPath Online Evaluator to experiment with JSONPath and regex101 to experiment with regex.

tasks

The tasks field is a dictionary with an optional key for each task. If present, it contains a dictionary that is converted to a set of keywords and passed to the Task's process function. The documentation for each task will provide the list of available parameters.

{
    "tasks": {
        "task-a": {
            "param1": "value1"
        },
        "task-c": {
            "param2": "value2"
        }
    }
}

In the example above a task named task-a would have the param1=value1 passed as a keyword, while task-c would have param2=value2 passed. If there were a task-b to be run it would not be passed any keywords.

TaskConfig Object

DEPRECATED: tasks should be a dictionary of parameters, with task names as keys. See tasks for more information.

A Task Configuration contains information for running a specific task.

Full Process Definition Example

Process definitions are sometimes called "Payloads":

{
    "description": "My process configuration",
    "collections": {
        "landsat-c2l2": "$[?(@.id =~ 'LC08.*')]"
    },
    "upload_options": {
        "path_template": "s3://my-bucket/${collection}/${year}/${month}/${day}/${id}"
    },
    "tasks": {
        "task-name": {
            "param": "value"
        }
    }
}

Migration

0.4.x -> 0.5.x

In 0.5.0, the previous use of fsspec to download Item Assets has been replaced with the stac-asset library. This has necessitated a change in the parameters that the download methods accept.

The primary change is that the Task methods download_item_assets and download_items_assets (items plural) now accept fewer explicit and implicit (kwargs) parameters.

Previously, the methods looked like:

  def download_item_assets(
        self,
        item: Item,
        path_template: str = "${collection}/${id}",
        keep_original_filenames: bool = False,
        **kwargs: Any,
    ) -> Item:

but now look like:

    def download_item_assets(
        self,
        item: Item,
        path_template: str = "${collection}/${id}",
        config: Optional[DownloadConfig] = None,
    ) -> Item:

Similarly, the asset_io package methods were previously:

async def download_item_assets(
    item: Item,
    assets: Optional[list[str]] = None,
    save_item: bool = True,
    overwrite: bool = False,
    path_template: str = "${collection}/${id}",
    absolute_path: bool = False,
    keep_original_filenames: bool = False,
    **kwargs: Any,
) -> Item:

and are now:

async def download_item_assets(
    item: Item,
    path_template: str = "${collection}/${id}",
    config: Optional[DownloadConfig] = None,
) -> Item:

Additionally, kwargs keys were set to pass configuration through to fsspec. The most common parameter was requester_pays, to set the Requester Pays flag in AWS S3 requests.

Many of these parameters can be directly translated into configuration passed in a DownloadConfig object, which is just a wrapper over the stac_asset.Config object.

Migration of these various parameters to DownloadConfig are as follows:

• assets: set include
• requester_pays: set s3_requester_pays = True
• keep_original_filenames: set file_name_strategy to FileNameStrategy.FILE_NAME if True or FileNameStrategy.KEY if False
• overwrite: set overwrite
• save_item: none, Item is always saved
• absolute_path: none. To create or retrieve the Asset hrefs as absolute paths, use either Item#make_all_asset_hrefs_absolute() or Asset#get_absolute_href()

0.5.x -> 0.6.0

Previously, the validate method was a classmethod, validating the payload argument passed. This has now been made an instance method, which validates the self._payload copy of the payload, from which the Task instance is constructed. This is behaviorally the same, in that construction will fail if validation fails, but allows implementers to utilize the instance method's convenience functions.

Previous implementations of validate would have been similar to this:

    @classmethod
    def validate(payload: dict[str, Any]) -> bool:
        # Check The Things™
        return isinstance(payload, dict)

And will now need to be updated to this form:

    def validate(self) -> bool:
        # Check The Things™
        return isinstance(self._payload, dict)

Development

Clone, install in editable mode with development requirements, and install the pre-commit hooks:

git clone https://github.com/stac-utils/stac-task
cd stac-task
pip install -e '.[dev]'
pre-commit install

To run the tests:

pytest

To lint all the files:

pre-commit run --all-files

Contributing

Use Github issues and pull requests.