#!/usr/bin/env python
# coding: utf-8

# # Compare two XML files (Nestle1904GBI)

# Jupiter Notebook intended to quicly compare the content and structure of two versions of the input XML files.

# In[3]:


# Following variables should contain the relative path and name of the two files to compare
book="01-matthew.xml"
version1="oct_27_2022"
version2="apr_6_2023"
file1_path = "sourcedata/"+version1+"/"+book
file2_path = "sourcedata/"+version2+"/"+book
# Set to True if detailed reporting is required
details = True


# In[4]:


import xml.etree.ElementTree as ET

def normalize_text(text):
    return ' '.join(text.strip().split())

def compare_xml_files(file1, file2, details):
    # Parse the XML files
    tree1 = ET.parse(file1)
    tree2 = ET.parse(file2)

    # Get the root elements of both trees
    root1 = tree1.getroot()
    root2 = tree2.getroot()

    # Compare the two root elements and their descendants recursively
    differences=compare_elements(root1, root2, details)

    if not differences:
        print("The XML files are identical.")
    else:
        print("The XML files are different.")
        if details:
            print("\nDetails:")
            for diff in differences:
                print(diff)

def compare_elements(elem1, elem2, details):
    differences = []
    ReportTag=ReportElement=ReportChildren=ReportChildDiff=True
    
    # Compare element tags and attributes
    if elem1.tag != elem2.tag or elem1.attrib != elem2.attrib:
        if details:
            differences.append(f"\n")
            for attr in elem1.attrib:
                if attr not in elem2.attrib:
                    differences.append(f"Attribute: {attr} is present in the first file but not in the second file")

            for attr in elem2.attrib:
                if attr not in elem1.attrib:
                   differences.append(f"Attribute: {attr} is present in the second file but not in the first file")
                
            for attr in elem1.attrib:
                if attr in elem2.attrib and elem1.attrib[attr] != elem2.attrib[attr]:
                   differences.append(f"Attribute: {attr} - Value in first file: {elem1.attrib[attr]}, Value in second file: {elem2.attrib[attr]}")
                
            differences.append(f"Tag/Attributes details: {elem1.tag} != {elem2.tag} or {elem1.attrib} != {elem2.attrib}")
            
        else:
            if ReportTag==True:
                differences.append("Differences in Tag/Attributes found")
                ReportTag=False

    # Normalize and compare element text
    text1 = normalize_text(elem1.text)
    text2 = normalize_text(elem2.text)
    
    if text1 != text2:
        if details:
            differences.append("Differences in element Text found")
            differences.append(f"Text: {elem1.text} != {elem2.text}\n")
        else:
            if ReportElement==True:
                differences.append("Differences in element Text found")
                ReportElement=False

    # Compare element children recursively
    children1 = list(elem1)
    children2 = list(elem2)
    if len(children1) != len(children2):
        if details:
            differences.append("Differences in number of children")
            differences.append(f"Number of Children: {len(children1)} != {len(children2)}\n")
        else:
            if ReportChildren==True:
                differences.append("Differences in number of children")
                ReportChildren=False    
    else:
        for child1, child2 in zip(children1, children2):
            child_diff = compare_elements(child1, child2, details)
            differences.extend(child_diff)
            
    return differences

# this is the main part
print ("Comparing file ",file1_path," with ",file2_path,"\n\nResult:",end="")
compare_xml_files(file1_path, file2_path, details)


# In[ ]:





# In[ ]: