Update documentation.

docs/source/CompressionSpecificationFormat.rst

 Compression Specification Format
 ================================
 
-In order to define compression specifications in a way that makes it readable and easy to use by different applications in differen programming languages, we defined our own Compression Specification Format.
-One example looks like this:
+We created a Compression Specification Format, which can be easily read and utilized by various applications written in different programming languages.
+An example of this format would be:
 
     >>> lossy,zfp,rate,4.0
 
-Its purpose is to represent the specifications in a way that are easy to understand and use.
-
 The currently implemented compressors include `Blosc <https://www.blosc.org/>`_ for lossless compression,
-and `ZFP <https://computing.llnl.gov/projects/zfp>`_ and `SZ <https://szcompressor.org/>`_ for lossy compression.
+and `ZFP <https://computing.llnl.gov/projects/zfp>`_, `SZ <https://szcompressor.org/>`_ and `SZ3 <https://szcompressor.org/>`_ for lossy compression.
 In order to use lossless compression, one can simply write:
 
     >>> lossless
@@ -30,18 +28,23 @@ One example with a different backend and compression level could be:
 
 
 For lossy compression, it is mandatory to include the compressor, the mode and the parameter.
-At the moment two lossy compressors are supported: **SZ** and **ZFP**.
+At the moment the lossy supported compressors are: **ZFP**, **SZ** and **SZ3**.
 
-The two compressors have different methods and different names for them:
-    -**SZ**:
-        - **abs**: absolute threshold mode
-        - **rel**: relative threshold mode
-        - **pw_rel**: point-wise relative threshold mode
+The compressors have different methods and different names for them:
     -**ZFP**:
         - **accuracy**: absolute threshold mode
         - **rate**: number of bits-per-value
         - **precision**: keep a certain bits of precision
 
+    -**SZ**:
+        - **abs**: absolute threshold mode
+        - **rel**: relative threshold mode
+        - **pw_rel**: point-wise relative threshold mode
+    -**SZ3**:
+        - **abs**: absolute threshold mode
+        - **rel**: relative threshold mode
+        - **norm2**: using norm2.
+        - **psnr**: using the peak signal to noise ratio mode.
 
 Few examples of lossy compression specifications:
 
@@ -49,6 +52,7 @@ Few examples of lossy compression specifications:
 
     lossy,sz,abs,0.01
     lossy,zfp,rate,4.0
+    lossy,sz3,psnr,40
     lossy,sz,rel,1e-3
     lossy,zfp,precision,10
     lossy,sz,pw_rel,0.05

docs/source/api.rst

@@ -7,8 +7,8 @@ Encodings
 .. autosummary::
    :toctree: Classes
 
-        enstools.encoding.api.FilterEncodingForXarray
-        enstools.encoding.api.FilterEncodingForH5py
+        enstools.encoding.api.DatasetEncoding
+        enstools.encoding.api.VariableEncoding
 
 
 Compressors and Compression modes

docs/source/index.rst

@@ -4,7 +4,7 @@ Welcome to Enstools-Encoding's documentation! |docsbadge|
 
 .. attention::
     This package is a building block of `enstools-compression <https://github.com/wavestoweather/enstools-compression>`_.
-    The recomendation is to directly use **enstools-compression**.
+    The recommendation is to directly use **enstools-compression**.
     The only purpose of splitting the package was to provide a way of using compression filters with **xarray** without adding too much dependencies.
     It might be the main sound reason to use this package separately.
     

docs/source/quick-start/basic_examples.rst

@@ -5,10 +5,10 @@ Using **lossy compression** with xarray can be as easy as adding a **single line
 
 .. code::
 
-    from enstools.encoding.api import FilterEncodingForXarray
+    from enstools.encoding.api import DatasetEncoding
 
     ...
-    encoding = FilterEncodingForXarray(dataset, "lossy,sz,rel,1.e-4")
+    encoding = DatasetEncoding(dataset, "lossy,sz,rel,1.e-4")
     dataset.to_netcdf(dummy_output_file, encoding=encoding, engine="h5netcdf")
     ...
 
@@ -18,9 +18,9 @@ Save an **xarray** dataset using losslessly compression:
 
 .. code::
 
-    from enstools.encoding.api import FilterEncodingForXarray
+    from enstools.encoding.api import DatasetEncoding
     ...
-    encoding = FilterEncodingForXarray(dataset, "lossless")
+    encoding = DatasetEncoding(dataset, "lossless")
     dataset.to_netcdf(dummy_output_file, encoding=encoding, engine="h5netcdf")
 
 
@@ -40,11 +40,11 @@ If we want to directly use **h5py** we can do the following:
 
 .. code::
 
-    from enstools.encoding.api import FilterEncodingForH5py
+    from enstools.encoding.api import VariableEncoding
 
     ...
 
-    encoding = FilterEncodingForH5py.from_string("lossless")
+    encoding = VariableEncoding("lossless")
     f = h5py.File('test.h5', 'w')
     f.create_dataset('lossless_compression_using_blosc',
                      data=numpy.arange(100),
@@ -56,11 +56,11 @@ Or without using specification strings:
 
 .. code::
 
-    from enstools.encoding.api import FilterEncodingForH5py, Compressors, CompressionModes
+    from enstools.encoding.api import VariableEncoding
 
     ...
 
-    encoding = FilterEncodingForH5py(Compressors.ZFP, CompressionModes.RATE, 4.0)
+    encoding = VariableEncoding(compressor="zfp", mode="rate",parameter="4.0")
 
     f = h5py.File('test.h5', 'w')
     f.create_dataset('lossy_compression_with_zfp_rate_4.0',