Merge pull request #18 from marxt/geotiff-zip

add zipped geotiff reading

src/veranda/raster/native/geotiff.py

@@ -2,6 +2,7 @@
 
 import os
 import struct
+from zipfile import ZipFile
 import numpy as np
 from osgeo import gdal
 from typing import List
@@ -128,8 +129,16 @@ def is_file_bigtiff(filepath) -> bool:
             True if the given file is a BigTIFF, else False.
 
         """
-        with open(filepath, 'rb') as f:
-            header = f.read(4)
+        if '.zip' in filepath:
+            if filepath.startswith('/vsizip/'):
+                filepath = filepath[len('/vsizip/'):]  # removes gdal virtual file system prefix to open w ZipFile
+            zip_filepath, inzip_filepath = filepath.split('.zip')
+            with ZipFile(zip_filepath + '.zip') as zip:
+                with zip.open(inzip_filepath[1:], 'r') as f:
+                    header = f.read(4)  # tests reads bytes not str, no need covert to bytearray
+        else:
+            with open(filepath, 'rb') as f:
+                header = f.read(4)
         byteorder = {b'II': '<', b'MM': '>', b'EP': '<'}[header[:2]]
         version = struct.unpack(byteorder + "H", header[2:4])[0]
         return version == 43
@@ -171,7 +180,20 @@ def _open(self):
 
         """
         if self.mode == 'r':
-            if not os.path.exists(self.filepath):
+            if '.zip' in self.filepath:
+                if self.filepath.startswith('/vsizip/'):
+                    self.filepath = self.filepath[len('/vsizip/'):]  # removes gdal virtual file system prefix if its there
+                zip_filepath, inzip_filepath = self.filepath.split('.zip')
+                zip_filepath += '.zip'
+                if not os.path.exists(zip_filepath):  # checks zip file
+                    err_msg = f"File '{zip_filepath}' does not exist."
+                    raise FileNotFoundError(err_msg)
+                with ZipFile(zip_filepath) as zip:
+                    if inzip_filepath[1:] not in zip.namelist():  # checks file in zip
+                        err_msg = f"File '{self.filepath}' does not exist."
+                        raise FileNotFoundError(err_msg)
+                self.filepath = '/vsizip/' + self.filepath  # adds the gdal vsi prefix
+            elif not os.path.exists(self.filepath):
                 err_msg = f"File '{self.filepath}' does not exist."
                 raise FileNotFoundError(err_msg)
             self.src = gdal.Open(self.filepath, gdal.GA_ReadOnly)
@@ -621,4 +643,4 @@ def _fill_vrt_file_per_band(et_root, filepath, bands, band_idx, entry_idx, n_col
 
 
 if __name__ == '__main__':
-    pass
+    pass

tests/raster/native/geotiff/test_geotiff.py

@@ -3,6 +3,7 @@
 import numpy as np
 from osgeo import gdal
 from tempfile import mkdtemp
+from zipfile import ZipFile
 
 from veranda.raster.native.geotiff import GeoTiffFile
 
@@ -137,3 +138,27 @@ def test_sref(filepath):
             sref_val = src.sref_wkt
 
         assert sref_val == sref_ref
+
+
+def test_read_one_band_zip(filepath):
+
+    data = np.ones((1, 100, 100), dtype=np.float32)
+
+    filename = os.path.basename(filepath)
+    zip_filepath = os.path.splitext(filepath)[0] + r'.zip'
+
+    with ZipFile(zip_filepath, 'w') as zip:
+        with GeoTiffFile(filepath, mode='w') as src:
+            src.write(data)
+        zip.write(filepath)
+
+        in_zip_filename = [name for name in zip.namelist() if name.endswith('.tif')][0]
+
+    read_filepath = os.path.join(zip_filepath, in_zip_filename)
+
+    with GeoTiffFile(read_filepath) as src:
+        ds = src.read()
+
+    np.testing.assert_array_equal(ds[1], data[0, :, :])
+
+    os.remove(zip_filepath)