feat: Add stactool DAGs

[ividito]

Summary:

Addresses #277

stactools-pipeline DAGs

• Adds generic/dynamic pipeline, as well as a noaa-hrrr pipeline to illustrate how more customized stactools packages can be implemented
• The generic pipeline has been tested on sentinel2, cop-dem, and landsat datasets with varied results.
• sentinel2 doesn't implement create_collection(), and so it fails with an AttributeError
• landsat requires a metadata file href as a granule, and works without additional parameters (although I have questions about how it will handle s3-hosted data)
• cop_dem works with additional parameters (this test case is included as the default input for the DAG)
• The noaa-hrrr package requires unique inputs, and would not work when used in the generic pipeline.

How to test

• Both DAGs have been pre-loaded with working, valid inputs. Run sm2a locally, point it at a working ingest API (either local or otherwise - I used local to avoid setting up auth), and trigger the DAGs with the default config.
• To try other stactools packages, add them to the worker requirements. If creating a new custom DAG (like noaa-hrrr, either add the package to the services requirements, or avoid importing the package in the same file that defines a DAG.

[amarouane-ABDELHAK]

@ividito will you be able to demo this for the team?

[amarouane-ABDELHAK]

Why are we installing libraries without constraints?

[ividito]

I hit a barrier installing stactools which requires a more secure version of httpx. According to the Airflow docs, once the Airflow installation is complete, it's recommended that additional dependencies are added as a second step without constraints.

[amarouane-ABDELHAK]

What do you think about us creating our own constraints.txt? that way we will not have to track two requirements.txt

[ividito]

I didn't want to spend time untangling it in this PR, but I think if we put the time in, we could end up with something like this:

pip install "airflow==${AIRFLOW_VERSION}" -r airflow-providers.txt -c airflow-constraints.txt
pip install dag-requirements.txt -c "airflow==${AIRFLOW_VERSION}"

The missing step is splitting up our current requirements file, where it's unclear which dependencies are copy-pasted constraints, and which are needed to maintain DAG functionality. This work can be incremental for a while - new dag dependencies should be added to requirements.in, and we can start to prune the old requirements file when we work on #318. Using uv pip tree while we do that can help us validate some of the links between dependencies in that file.