Defining Source¶

The source parameter in Earth Data Kit can be defined in multiple ways depending on your needs and the engine you’re using.

Source Types¶

Earth Data Kit supports three main ways to define a source:

Single File Path: A direct path to a single data file:
```
source = "s3://my-bucket/path/to/file.tif"
```

Multiple Files as Array: A list of specific file paths:

source = ["s3://my-bucket/file1.tif", "s3://my-bucket/file2.tif"]

Pattern-based Source: A template with placeholders for matching multiple files:
```
source = "s3://modis-pds/MCD43A4.006/{h}/{v}/%Y%j/*.TIF"
```

Note that you cannot mix patterns within an array - if using an array, each element must be a direct file path.

The format of your source definition depends on the engine you’re using:

Storage-based Engines (AWS S3, filesystem, http(s), FTP): These engines support pattern-based sources that can include time and spatial components.

Earth Engine: Requires the collection name as defined in Google Earth Engine:

source = "COPERNICUS/S2_SR_HARMONIZED"  # Example: Sentinel-2 Surface Reflectance Harmonized

For storage-based engines, you can include time and spatial components in your source pattern:

Time Components: Use standard Python datetime format codes (e.g., %Y for 4-digit year, %m for month) (see https://docs.python.org/3/library/datetime.html#format-codes for reference):
```
source = "s3://my-bucket/data/%Y/%m/%d/*.tif"  # Matches year/month/day folders
```
Spatial Components: Use placeholders like {h} and {v} for horizontal and vertical grid coordinates. Note that you will have to provide a grid dataframe with the appropriate columns that match your placeholders:
```
source = "s3://landsat-pds/c1/L8/{p}/{r}/%Y/%m/%d/*.TIF"  # {p} for path, {r} for row
```

These patterns work in conjunction with the set_timebounds() and set_spacebounds() methods to filter data based on your specified criteria.

For a detailed example of using pattern-based sources with MODIS data, see Example: MODIS PDS.