Merge pull request #3 from smoke-trees/docs/export_keras

added docstings to export_keras.py file

forest_utils/export_keras.py

@@ -8,7 +8,33 @@
 
 
 class ModelFromH5(object):
+    """
+    A class for managing downloads and loading of .h5 models
+
+    Parameters
+    ----------
+    output : str
+        path to output file for downloading the model (by default it is 'model.h5')
+    config : str
+        path to the config file of the model (by default it is 'result.json')
     
+    Attributes
+    ----------
+    base_url : str
+        This is the bace url for downloading the .h5 model files
+    url_id 
+        Contains the link to the model extracted from the results.json files
+    output : str
+        Relative path to the output file for the model download
+
+    Methods
+    -------
+    get_complete_url(url)
+        method to get complete link from the given url
+    load_model()
+        download the model .h5 file from the url to output route and returns the loaded keras model
+    """
+
     def __init__(self, output = 'model.h5', config = 'result.json'):
         super().__init__()
 
@@ -17,10 +43,32 @@ def __init__(self, output = 'model.h5', config = 'result.json'):
         self.output = output
 
     def get_complete_url(self, url):
+        """
+        method (used internally inside class) to get complete link (including base_url) from the given url
+
+        Parameters
+        ----------
+        url : str
+            url to split and make complete url from
+        
+        Returns
+        -------
+        link : str
+            complete url to the model file
+
+        """
         split_url = url.split('/')
         return self.base_url + split_url[5]
 
     def load_model(self):
+        """
+        method to download model from the url to the output file and load it into keras
+
+        Returns 
+        -------
+        keras model
+            downloaded model loaded into keras model ready to use!
+        """
         try:
             gdown.download(self.url_id, self.output, quiet = False)
             return tf.keras.models.load_model(self.output)