Product

Product

class Product(product_path: Union[str, cloudpathlib.cloudpath.CloudPath, pathlib.Path], archive_path: Optional[Union[str, cloudpathlib.cloudpath.CloudPath, pathlib.Path]] = None, output_path: Optional[Union[str, cloudpathlib.cloudpath.CloudPath, pathlib.Path]] = None, remove_tmp: bool = False, **kwargs)[source]

Bases: object

Super class of EOReader Products

Methods

__init__(product_path[, archive_path, ...])

clean_tmp()

Clean the temporary directory of the current product

clear()

Clear this product's cache

get_band_paths(band_list[, resolution])

Return the paths of required bands.

get_date([as_date])

Get the product's acquisition date.

get_datetime([as_datetime])

Get the product's acquisition datetime, with format YYYYMMDDTHHMMSS <-> %Y%m%dT%H%M%S

get_default_band()

Get default band: Usually GREEN band for optical data and the first existing one between VV and HH for SAR data.

get_default_band_path(**kwargs)

Get default band path (among the existing ones).

get_existing_band_paths()

Return the existing band paths.

get_existing_bands()

Return the existing bands.

has_band(band)

Does this products has the specified band ?

has_bands(bands)

Does this products has the specified bands ?

load(bands[, resolution, size])

Open the bands and compute the wanted index.

read_mtd()

Read metadata and outputs the metadata XML root and its namespaces as a dict most of the time, except from L8-collection 1 data which outputs a pandas.DataFrame

stack(bands[, resolution, size, stack_path, ...])

Stack bands and index of a products.

Attributes

crs

Get UTM projection of the tile

default_transform

extent

Get UTM extent of the tile

footprint

Get UTM footprint of the products (without nodata, in french == emprise utile)

output

Output directory of the product, to write orthorectified data for example.

needs_extraction

Does this products needs to be extracted to be processed ? (True by default).

path

Usable path to the product, either extracted or archived path, according to the satellite.

filename

Product filename

name

Product true name (as specified in the metadata)

split_name

Split name, to retrieve every information from its true name (dates, tile, product type...).

archive_path

Archive path, same as the product path if not specified.

is_archived

Is the archived product is processed (a products is considered as archived if its products path is a directory).

date

Acquisition date.

datetime

Acquisition datetime.

tile_name

Tile if possible (for data that can be piled, for example S2 and Landsats).

sensor_type

Sensor type, SAR or optical.

product_type

Product type, satellite-related field, such as L1C or L2A for Sentinel-2 data.

band_names

Band mapping between band wrapping names such as GREEN and band real number such as 03 for Sentinel-2.

is_reference

If the product is a reference, used for algorithms that need pre and post data, such as fire detection.

corresponding_ref

The corresponding reference products to the current one (if the product is not a reference but has a reference data corresponding to it).

nodata

Product nodata, set to -9999 by default

platform

Product platform, such as Sentinel-2

resolution

Default resolution in meters of the current product.

condensed_name

Condensed name, the filename with only useful data to keep the name unique (ie.

sat_id

Satellite ID, i.e. S2 for Sentinel-2.

__init__(product_path: Union[str, cloudpathlib.cloudpath.CloudPath, pathlib.Path], archive_path: Optional[Union[str, cloudpathlib.cloudpath.CloudPath, pathlib.Path]] = None, output_path: Optional[Union[str, cloudpathlib.cloudpath.CloudPath, pathlib.Path]] = None, remove_tmp: bool = False, **kwargs) None[source]
clean_tmp()[source]

Clean the temporary directory of the current product

clear()[source]

Clear this product’s cache

abstract get_band_paths(band_list: list, resolution: Optional[float] = None, **kwargs) dict[source]

Return the paths of required bands.

>>> from eoreader.reader import Reader
>>> from eoreader.bands import *
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.get_band_paths([GREEN, RED])
{
    <OpticalBandNames.GREEN: 'GREEN'>: 'zip+file://S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip!/S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE/GRANULE/L1C_T30TTK_A027018_20200824T111345/IMG_DATA/T30TTK_20200824T110631_B03.jp2',
    <OpticalBandNames.RED: 'RED'>: 'zip+file://S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip!/S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE/GRANULE/L1C_T30TTK_A027018_20200824T111345/IMG_DATA/T30TTK_20200824T110631_B04.jp2'
}
Parameters
  • band_list (list) – List of the wanted bands

  • resolution (float) – Band resolution

  • kwargs – Other arguments used to load bands

Returns

Dictionary containing the path of each queried band

Return type

dict

get_date(as_date: bool = False) Union[str, datetime.date][source]

Get the product’s acquisition date.

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.get_date(as_date=True)
datetime.datetime(2020, 8, 24, 0, 0)
>>> prod.get_date(as_date=False)
'20200824'
Parameters

as_date (bool) – Return the date as a datetime.date. If false, returns a string.

Returns

Its acquisition date

Return type

str

abstract get_datetime(as_datetime: bool = False) Union[str, datetime.datetime][source]

Get the product’s acquisition datetime, with format YYYYMMDDTHHMMSS <-> %Y%m%dT%H%M%S

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.get_datetime(as_datetime=True)
datetime.datetime(2020, 8, 24, 11, 6, 31)
>>> prod.get_datetime(as_datetime=False)
'20200824T110631'
Parameters

as_datetime (bool) – Return the date as a datetime.datetime. If false, returns a string.

Returns

Its acquisition datetime

Return type

Union[str, datetime.datetime]

abstract get_default_band() eoreader.bands.bands.BandNames[source]

Get default band: Usually GREEN band for optical data and the first existing one between VV and HH for SAR data.

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.get_default_band()
<OpticalBandNames.GREEN: 'GREEN'>
Returns

Default band

Return type

str

abstract get_default_band_path(**kwargs) Union[cloudpathlib.cloudpath.CloudPath, pathlib.Path][source]

Get default band path (among the existing ones).

Usually GREEN band for optical data and the first existing one between VV and HH for SAR data.

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.get_default_band_path()
'zip+file://S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip!/S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE/GRANULE/L1C_T30TTK_A027018_20200824T111345/IMG_DATA/T30TTK_20200824T110631_B03.jp2'
Parameters

kwargs – Additional arguments

Returns

Default band path

Return type

Union[CloudPath, Path]

abstract get_existing_band_paths() dict[source]

Return the existing band paths.

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.get_existing_band_paths()
{
    <OpticalBandNames.CA: 'COASTAL_AEROSOL'>: 'zip+file://S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip!/S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE/GRANULE/L1C_T30TTK_A027018_20200824T111345/IMG_DATA/T30TTK_20200824T110631_B01.jp2',
    ...,
    <OpticalBandNames.SWIR_2: 'SWIR_2'>: 'zip+file://S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip!/S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE/GRANULE/L1C_T30TTK_A027018_20200824T111345/IMG_DATA/T30TTK_20200824T110631_B12.jp2'
}
Returns

Dictionary containing the path of each queried band

Return type

dict

abstract get_existing_bands() list[source]

Return the existing bands.

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.get_existing_bands()
[<OpticalBandNames.CA: 'COASTAL_AEROSOL'>,
<OpticalBandNames.BLUE: 'BLUE'>,
<OpticalBandNames.GREEN: 'GREEN'>,
<OpticalBandNames.RED: 'RED'>,
<OpticalBandNames.VRE_1: 'VEGETATION_RED_EDGE_1'>,
<OpticalBandNames.VRE_2: 'VEGETATION_RED_EDGE_2'>,
<OpticalBandNames.VRE_3: 'VEGETATION_RED_EDGE_3'>,
<OpticalBandNames.NIR: 'NIR'>,
<OpticalBandNames.NNIR: 'NARROW_NIR'>,
<OpticalBandNames.WV: 'WATER_VAPOUR'>,
<OpticalBandNames.CIRRUS: 'CIRRUS'>,
<OpticalBandNames.SWIR_1: 'SWIR_1'>,
<OpticalBandNames.SWIR_2: 'SWIR_2'>]
Returns

List of existing bands in the products

Return type

list

has_band(band: Union[eoreader.bands.bands.BandNames, Callable]) bool[source]

Does this products has the specified band ?

By band, we mean:

  • satellite band

  • index

  • DEM band

  • cloud band

>>> from eoreader.reader import Reader
>>> from eoreader.bands import *
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.has_band(GREEN)
True
>>> prod.has_band(TIR_2)
False
>>> prod.has_band(NDVI)
True
>>> prod.has_band(SHADOWS)
False
>>> prod.has_band(HILLSHADE)
True
Parameters

band (Union[BandNames, Callable]) – EOReader band (optical, SAR, clouds, DEM)

Returns

True if the products has the specified band

Return type

bool

has_bands(bands: Union[list, eoreader.bands.bands.BandNames, Callable]) bool[source]

Does this products has the specified bands ?

By band, we mean:

  • satellite band

  • index

  • DEM band

  • cloud band

See has_bands for a code example.

Parameters

bands (Union[list, BandNames, Callable]) – EOReader bands (optical, SAR, clouds, DEM)

Returns

True if the products has the specified band

Return type

bool

load(bands: Union[list, eoreader.bands.bands.BandNames, Callable], resolution: Optional[float] = None, size: Optional[Union[list, tuple]] = None, **kwargs) dict[source]

Open the bands and compute the wanted index.

The bands will be purged of nodata and invalid pixels, the nodata will be set to 0 and the bands will be masked arrays in float.

Bands that come out this function at the same time are collocated and therefore have the same shapes. This can be broken if you load data separately. Its is best to always load DEM data with some real bands.

>>> from eoreader.reader import Reader
>>> from eoreader.bands import *
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> bands = prod.load([GREEN, NDVI], resolution=20)
>>> bands
{
    <function NDVI at 0x000001EFFFF5DD08>: <xarray.DataArray 'NDVI' (band: 1, y: 5490, x: 5490)>
    array([[[0.949506  , 0.92181516, 0.9279379 , ..., 1.8002278 ,
             1.5424857 , 1.6747767 ],
            [0.95369846, 0.91685396, 0.8957871 , ..., 1.5847116 ,
             1.5248713 , 1.5011379 ],
            [2.9928885 , 1.3031474 , 1.0076253 , ..., 1.5969834 ,
             1.5590671 , 1.5018653 ],
            ...,
            [1.4245619 , 1.6115025 , 1.6201663 , ..., 1.2387121 ,
             1.4025431 , 1.800678  ],
            [1.5627214 , 1.822388  , 1.7245892 , ..., 1.1694248 ,
             1.2573677 , 1.5767351 ],
            [1.653781  , 1.6424649 , 1.5923225 , ..., 1.3072611 ,
             1.2181134 , 1.2478763 ]]], dtype=float32)
    Coordinates:
      * band         (band) int32 1
      * y            (y) float64 4.5e+06 4.5e+06 4.5e+06 ... 4.39e+06 4.39e+06
      * x            (x) float64 2e+05 2e+05 2e+05 ... 3.097e+05 3.098e+05 3.098e+05
        spatial_ref  int32 0,
    <OpticalBandNames.GREEN: 'GREEN'>: <xarray.DataArray (band: 1, y: 5490, x: 5490)>
    array([[[0.0615  , 0.061625, 0.061   , ..., 0.12085 , 0.120225,
             0.113575],
            [0.061075, 0.06045 , 0.06025 , ..., 0.114625, 0.119625,
             0.117625],
            [0.06475 , 0.06145 , 0.060925, ..., 0.111475, 0.114925,
             0.115175],
            ...,
            [0.1516  , 0.14195 , 0.1391  , ..., 0.159975, 0.14145 ,
             0.127075],
            [0.140325, 0.125975, 0.131875, ..., 0.18245 , 0.1565  ,
             0.13015 ],
            [0.133475, 0.1341  , 0.13345 , ..., 0.15565 , 0.170675,
             0.16405 ]]], dtype=float32)
    Coordinates:
      * band         (band) int32 1
      * y            (y) float64 4.5e+06 4.5e+06 4.5e+06 ... 4.39e+06 4.39e+06
      * x            (x) float64 2e+05 2e+05 2e+05 ... 3.097e+05 3.098e+05 3.098e+05
        spatial_ref  int32 0
}
Parameters
  • bands (Union[list, BandNames, Callable]) – Band list

  • resolution (float) – Resolution of the band, in meters

  • size (Union[tuple, list]) – Size of the array (width, height). Not used if resolution is provided.

  • kwargs – Other arguments used to load bands

Returns

{band_name, band xarray}

Return type

dict

read_mtd() Any[source]

Read metadata and outputs the metadata XML root and its namespaces as a dict most of the time, except from L8-collection 1 data which outputs a pandas.DataFrame

>>> from eoreader.reader import Reader
>>> path = r"S1A_IW_GRDH_1SDV_20191215T060906_20191215T060931_030355_0378F7_3696.zip"
>>> prod = Reader().open(path)
>>> prod.read_mtd()
(<Element product at 0x1832895d788>, '')
Returns

Metadata XML root and its namespace or pd.DataFrame

Return type

Any

stack(bands: list, resolution: Optional[float] = None, size: Optional[Union[list, tuple]] = None, stack_path: Optional[Union[str, cloudpathlib.cloudpath.CloudPath, pathlib.Path]] = None, save_as_int: bool = False, **kwargs) xarray.core.dataarray.DataArray[source]

Stack bands and index of a products.

>>> from eoreader.reader import Reader
>>> from eoreader.bands import *
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> stack = prod.stack([NDVI, MNDWI, GREEN], resolution=20)  # In meters
>>> stack
<xarray.DataArray 'NDVI_MNDWI_GREEN' (z: 3, y: 5490, x: 5490)>
array([[[ 0.949506  ,  0.92181516,  0.9279379 , ...,  1.8002278 ,
          1.5424857 ,  1.6747767 ],
        [ 0.95369846,  0.91685396,  0.8957871 , ...,  1.5847116 ,
          1.5248713 ,  1.5011379 ],
        [ 2.9928885 ,  1.3031474 ,  1.0076253 , ...,  1.5969834 ,
          1.5590671 ,  1.5018653 ],
        ...,
        [ 1.4245619 ,  1.6115025 ,  1.6201663 , ...,  1.2387121 ,
          1.4025431 ,  1.800678  ],
        [ 1.5627214 ,  1.822388  ,  1.7245892 , ...,  1.1694248 ,
          1.2573677 ,  1.5767351 ],
        [ 1.653781  ,  1.6424649 ,  1.5923225 , ...,  1.3072611 ,
          1.2181134 ,  1.2478763 ]],
       [[ 0.27066118,  0.23466069,  0.18792598, ..., -0.4611526 ,
         -0.49751845, -0.4865216 ],
        [ 0.22425456,  0.28004232,  0.27851456, ..., -0.5032771 ,
         -0.501796  , -0.502669  ],
        [-0.07466951,  0.06360884,  0.1207174 , ..., -0.50617427,
         -0.50219285, -0.5034222 ],
        [-0.47076276, -0.4705828 , -0.4747971 , ..., -0.32138503,
         -0.36619243, -0.37428448],
        [-0.4826967 , -0.5032287 , -0.48544118, ..., -0.278925  ,
         -0.31404778, -0.36052078],
        [-0.488381  , -0.48253912, -0.4697526 , ..., -0.38105175,
         -0.30813277, -0.27739233]],
       [[ 0.0615    ,  0.061625  ,  0.061     , ...,  0.12085   ,
          0.120225  ,  0.113575  ],
        [ 0.061075  ,  0.06045   ,  0.06025   , ...,  0.114625  ,
          0.119625  ,  0.117625  ],
        [ 0.06475   ,  0.06145   ,  0.060925  , ...,  0.111475  ,
          0.114925  ,  0.115175  ],
        ...,
        [ 0.1516    ,  0.14195   ,  0.1391    , ...,  0.159975  ,
          0.14145   ,  0.127075  ],
        [ 0.140325  ,  0.125975  ,  0.131875  , ...,  0.18245   ,
          0.1565    ,  0.13015   ],
        [ 0.133475  ,  0.1341    ,  0.13345   , ...,  0.15565   ,
          0.170675  ,  0.16405   ]]], dtype=float32)
Coordinates:
  * y            (y) float64 4.5e+06 4.5e+06 4.5e+06 ... 4.39e+06 4.39e+06
  * x            (x) float64 2e+05 2e+05 2e+05 ... 3.097e+05 3.098e+05 3.098e+05
    spatial_ref  int32 0
  * z            (z) MultiIndex
  - variable     (z) object 'NDVI' 'MNDWI' 'GREEN'
  - band         (z) int64 1 1 1
-Attributes:
    long_name:  ['NDVI', 'MNDWI', 'GREEN']
Parameters
  • bands (list) – Bands and index combination

  • resolution (float) – Stack resolution. . If not specified, use the product resolution.

  • size (Union[tuple, list]) – Size of the array (width, height). Not used if resolution is provided.

  • stack_path (Union[str, CloudPath, Path]) – Stack path

  • save_as_int (bool) – Convert stack to uint16 to save disk space (and therefore multiply the values by 10.000)

  • **kwargs – Other arguments passed to load or rioxarray.to_raster() (such as compress)

Returns

Stack as a DataArray

Return type

xr.DataArray

archive_path

Archive path, same as the product path if not specified. Useful when you want to know where both the extracted and archived version of your product are stored.

band_names

Band mapping between band wrapping names such as GREEN and band real number such as 03 for Sentinel-2.

condensed_name

Condensed name, the filename with only useful data to keep the name unique (ie. 20191215T110441_S2_30TXP_L2A_122756). Used to shorten names and paths.

corresponding_ref

The corresponding reference products to the current one (if the product is not a reference but has a reference data corresponding to it). A list because of multiple ref in case of non-stackable products (S3, S1…)

abstract property crs: rasterio.crs.CRS

Get UTM projection of the tile

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.utm_crs()
CRS.from_epsg(32630)
Returns

CRS object

Return type

crs.CRS

date

Acquisition date.

datetime

Acquisition datetime.

default_transform = <methodtools._LruCacheWire object>[source]
abstract property extent: geopandas.geodataframe.GeoDataFrame

Get UTM extent of the tile

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.utm_extent()
                                            geometry
0  POLYGON ((309780.000 4390200.000, 309780.000 4...
Returns

Footprint in UTM

Return type

gpd.GeoDataFrame

filename

Product filename

property footprint: geopandas.geodataframe.GeoDataFrame

Get UTM footprint of the products (without nodata, in french == emprise utile)

>>> from eoreader.reader import Reader
>>> path = r"S2A_MSIL1C_20200824T110631_N0209_R137_T30TTK_20200824T150432.SAFE.zip"
>>> prod = Reader().open(path)
>>> prod.footprint
   index                                           geometry
0      0  POLYGON ((199980.000 4500000.000, 199980.000 4...
Returns

Footprint as a GeoDataFrame

Return type

gpd.GeoDataFrame

is_archived

Is the archived product is processed (a products is considered as archived if its products path is a directory).

is_reference

If the product is a reference, used for algorithms that need pre and post data, such as fire detection.

name

Product true name (as specified in the metadata)

needs_extraction

Does this products needs to be extracted to be processed ? (True by default).

nodata

Product nodata, set to -9999 by default

property output: Union[cloudpathlib.cloudpath.CloudPath, pathlib.Path]

Output directory of the product, to write orthorectified data for example.

path

Usable path to the product, either extracted or archived path, according to the satellite.

platform

Product platform, such as Sentinel-2

product_type

Product type, satellite-related field, such as L1C or L2A for Sentinel-2 data.

resolution

Default resolution in meters of the current product. For SAR product, we use Ground Range resolution as we will automatically orthorectify the tiles.

sat_id

Satellite ID, i.e. S2 for Sentinel-2

sensor_type

Sensor type, SAR or optical.

split_name

Split name, to retrieve every information from its true name (dates, tile, product type…).

tile_name

Tile if possible (for data that can be piled, for example S2 and Landsats).