Upload files to "/"

extract_attachments.py

@@ -0,0 +1,101 @@
+import base64
+from lxml import etree
+import gzip
+import shutil, os, io
+
+
+def write_file_in_new_directory(directory, file_name, content):
+    # Create the new directory if it doesn't exist
+    os.makedirs(directory, exist_ok=True)
+    
+    # Define the file path
+    file_path = os.path.join(directory, file_name)
+    
+    # Write content to the file
+    with open(file_path, 'wb') as file:
+        file.write(content)
+    
+    print(f"File '{file_name}' has been written in the directory '{directory}'")
+
+def write_gunzip_bytestream_to_file(bytestream, output_directory, output_file):
+    # Create the directory if it doesn't exist
+    os.makedirs(output_directory, exist_ok=True)
+    
+    # Define the file path
+    file_path = os.path.join(output_directory, output_file)
+    
+    # Decompress the GZIP bytestream
+    with gzip.GzipFile(fileobj=io.BytesIO(bytestream), mode='rb') as gzip_file:
+        decompressed_data = gzip_file.read()
+    
+    # Write the decompressed data to the file
+    with open(file_path, 'wb') as file:
+        file.write(decompressed_data)
+    
+    print(f"Decompressed data has been written to '{file_path}'")
+
+# Parse the XML file
+tree = etree.parse('incident.xml')
+root = tree.getroot()
+
+# Extract and save attachments
+for attachment in root.findall(".//sys_attachment"):
+    file_name_element = attachment.find('file_name')
+    if file_name_element is None:
+        continue
+    file_name = file_name_element.text
+    sys_id = attachment.find('sys_id').text
+
+    previous_element =  None
+    for sibling in attachment.itersiblings(preceding=True):
+        if sibling.tag == 'incident':
+            previous_element = sibling
+            break
+
+    incident = previous_element.find('number').text
+    print(f"Processing attachment: {file_name}, sys_id: {sys_id} for incident: {incident}")
+
+    # Find the corresponding attachment data chunks
+    attachment_data_chunks = root.findall(f".//sys_attachment_doc[sys_attachment='{sys_id}']")
+    combined_data = ""
+
+    if not attachment_data_chunks:
+        print(f"No data chunks found for sys_id: {sys_id}")
+        continue
+
+    # Save the attachment to a file
+    # try:
+    #     with open(f'./output/{file_name}.gz', 'wb') as file:
+    #         for chunk in attachment_data_chunks:
+    #             encoded_data = chunk.find('data').text
+
+    #             # Decode the combined base64-encoded data
+    #             try:
+    #                 decoded_data = base64.b64decode(encoded_data)
+    #                 file.write(decoded_data)
+    #                 # write_file_in_new_directory(incident, file_name + ".gz", decoded_data)
+    #             except Exception as e:
+    #                 print(f"Error decoding data for sys_id: {sys_id}, Error: {e}")
+    #                 continue
+
+    try:
+        # with open(f'./output/{file_name}.gz', 'wb') as file:
+        decoded_data = b''
+        for chunk in attachment_data_chunks:
+            encoded_data = chunk.find('data').text
+
+            # Decode the combined base64-encoded data
+            try:
+                decoded_data += base64.b64decode(encoded_data)
+        
+            except Exception as e:
+                print(f"Error decoding data for sys_id: {sys_id}, Error: {e}")
+                continue
+
+        # write_file_in_new_directory(incident, file_name + ".gz", decoded_data)
+        write_gunzip_bytestream_to_file(decoded_data, incident, file_name)
+
+    except Exception as e:
+        print(f"Error saving attachment: {file_name}, Error: {e}")
+
+print("All attachments processed.")