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

# # Creating Text-Fabric dataset (from LowFat XML trees)
# 
# <pre>
#     Code version: 0.7 (February 20, 2024)
#     Data version: February 10, 2024 (<a href="https://github.com/tonyjurg/Nestle1904LFT/tree/main/resources/xml/20240210#readme" target="_blank">Readme</a>)
# </pre>
# 
# ## Table of content <a class="anchor" id="TOC"></a>
# * <a href="#bullet1">1 - Introduction</a>
# * <a href="#bullet2">2 - Read LowFat XML data and store in pickle</a>
#     * <a href="#bullet2x1">2.1 - Required libraries</a>
#     * <a href="#bullet2x2">2.2 - Import various libraries</a>
#     * <a href="#bullet2x3">2.3 - Initialize global data</a>
#     * <a href="#bullet2x4">2.4 - Process the XML data and store dataframe in pickle</a>
# * <a href="#bullet3">3 - Optionaly export to aid investigation</a>
#     * <a href="#bullet3x1">3.1 - Export to Excel format</a> 
#     * <a href="#bullet3x2">3.2 - Export to CSV format</a>
# * <a href="#bullet4">4 - Text-Fabric dataset production from pickle input</a>
#     * <a href="#bullet4x1">4.1 - Explanation</a>
#     * <a href="#bullet4x1">4.2 - Running the TF walker function</a>
# * <a href="#bullet5">5 - Housekeeping</a>
#     * <a href="#bullet5x1">5.1 - Optionaly zip-up the pickle files</a>
#     * <a href="#bullet5x2">5.2 - Publishing on gitHub</a>

# # 1 - Introduction <a class="anchor" id="bullet1"></a>
# ##### [Back to TOC](#TOC)
# 
# The source data for the conversion are the LowFat XML trees files representing the macula-greek version of the Nestle 1904 Greek New Testment (British Foreign Bible Society, 1904). The starting dataset is formatted according to Syntax diagram markup by the Global Bible Initiative (GBI). The most recent source data can be found on github https://github.com/Clear-Bible/macula-greek/tree/main/Nestle1904/lowfat. 
# 
# Attribution: "MACULA Greek Linguistic Datasets, available at https://github.com/Clear-Bible/macula-greek/". 
# 
# The production of the Text-Fabric files consist of two phases. First one is the creation of piclke files (<a href="#bullet2">section 2</a>). The second phase is the the actual Text-Fabric creation process (<a href="#bullet3">section 3</a>). The process can be depicted as follows:
# 
# <img src='../images/highlevelconversion.png'>

# # 2 - Read LowFat XML data and store in pickle <a class="anchor" id="bullet2"></a>
# ##### [Back to TOC](#TOC)
# 
# This script harvests all information from the LowFat tree data (XML nodes), puts it into a Panda DataFrame and stores the result per book in a pickle file. Note: pickling (in Python) is serialising an object into a disk file (or buffer). See also the [Python3 documentation](https://docs.python.org/3/library/pickle.html).
# 
# Within the context of this script, the term 'Leaf' refers to nodes that contain the Greek word as data. These nodes are also referred to as 'terminal nodes' since they do not have any children, similar to leaves on a tree. Additionally, Parent1 represents the parent of the leaf, Parent2 represents the parent of Parent1, and so on. For a visual representation, please refer to the following diagram.
# 
# <img src="../images/leaves_and_root.png">
# 
# For a full description of the source data see document [MACULA Greek Treebank for the Nestle 1904 Greek New Testament.pdf](https://github.com/Clear-Bible/macula-greek/blob/main/doc/MACULA%20Greek%20Treebank%20for%20the%20Nestle%201904%20Greek%20New%20Testament.pdf)

# ## 2.1 Required libraries<a class="anchor" id="bullet2x1"></a>
# ##### [Back to TOC](#TOC)
# 
# The scripts in this notebook require (beside text-fabric) a number of Python libraries to be installed in the environment (see following section).
# You can install any missing library from within Jupyter Notebook using either `pip` or `pip3`. (eg.: !pip3 install pandas)

# ## 2.2 - Import various libraries<a class="anchor" id="bullet2x2"></a>
# ##### [Back to TOC](#TOC)
# 
# The following cell reads all required libraries by the scripts in this notebook.

# In[2]:


import pandas as pd
import sys                       # System
import os                        # Operating System
from os import listdir
from os.path import isfile, join
import time
import pickle
import re                        # Regular Expressions
from lxml import etree as ET
from tf.fabric import Fabric
from tf.convert.walker import CV
from tf.parameters import VERSION
from datetime import date
import pickle
import unicodedata
from unidecode import unidecode
import openpyxl


# ## 2.3 - Initialize global data<a class="anchor" id="bullet2x3"></a>
# ##### [Back to TOC](#TOC)
# 
# The following cell initializes the global data used by the various scripts in this notebook. Many of these global variables are shared among the scripts as they relate to common entities.
# 
# IMPORTANT: To ensure proper creation of the Text-Fabric files on your system, it is crucial to adjust the values of BaseDir, XmlDir, etc. to match the location of the data and the operating system you are using. In this Jupyter Notebook, Windows is the operating system employed.

# In[6]:


# set script version number
scriptVersion='0.7'
scriptDate='February 20, 2024'

# Define the source and destination locations 
BaseDir = '..\\'
XmlDir = BaseDir+'xml\\20240210\\'
PklDir = BaseDir+'pickle\\20240210\\'
XlsxDir = BaseDir+'excel\\20240210\\'
CsvDir = BaseDir+'csv\\20240210\\'
# note: create output directory prior running the scripts!

#         key: filename,       [0]=bookLong,   [1]=bookNum,  [3]=bookShort
bo2book = {'01-matthew':       ['Matthew',         '1',        'Matt'],
           '02-mark':          ['Mark',            '2',        'Mark'],
           '03-luke':          ['Luke',            '3',        'Luke'],
           '04-john':          ['John',            '4',        'John'],
           '05-acts':          ['Acts',            '5',        'Acts'],
           '06-romans':        ['Romans',          '6',        'Rom'],
           '07-1corinthians':  ['I_Corinthians',   '7',        '1Cor'],
           '08-2corinthians':  ['II_Corinthians',  '8',        '2Cor'],
           '09-galatians':     ['Galatians',       '9',        'Gal'],
           '10-ephesians':     ['Ephesians',       '10',       'Eph'],
           '11-philippians':   ['Philippians',     '11',       'Phil'],
           '12-colossians':    ['Colossians',      '12',       'Col'],
           '13-1thessalonians':['I_Thessalonians', '13',       '1Thess'],
           '14-2thessalonians':['II_Thessalonians','14',       '2Thess'],
           '15-1timothy':      ['I_Timothy',       '15',       '1Tim'],
           '16-2timothy':      ['II_Timothy',      '16',       '2Tim'],
           '17-titus':         ['Titus',           '17',       'Titus'],
           '18-philemon':      ['Philemon',        '18',       'Phlm'],
           '19-hebrews':       ['Hebrews',         '19',       'Heb'],
           '20-james':         ['James',           '20',       'Jas'],
           '21-1peter':        ['I_Peter',         '21',       '1Pet'],
           '22-2peter':        ['II_Peter',        '22',       '2Pet'],
           '23-1john':         ['I_John',          '23',       '1John'],
           '24-2john':         ['II_John',         '24',       '2John'],
           '25-3john':         ['III_John',        '25',       '3John'],     
           '26-jude':          ['Jude',            '26',       'Jude'],
           '27-revelation':    ['Revelation',      '27',       'Rev']}


# ## 2.4 - Process the XML data and store dataframe in pickle<a class="anchor" id="bullet2x4"></a>
# ##### [Back to TOC](#TOC)
# 
# This code processes all 27 books in the correct order.
# For each book, the following is done:
# 
# * create a parent-child map based upon the XML source (function buildParentMap).
# * loop trough the XML source to identify 'leaf' nodes and gather information regarding all its parents (function processElement) and store the results in a datalist.
# * After processing all the nodes the datalist is converted to a datframe and exported as a pickle file specific to that book.
# 
# Once the XML data is converted to PKL files, there is no need to rerun (unless the source XML data is updated).
# 
# Since the size of the pickle files can be rather large, it is advised to add the .pkl extention to the ignore list of gitHub (.gitignore)

# In[3]:


# Create the pickle files

# Set global variables for this script
WordOrder = 1
CollectedItems = 0

###############################################
#           The helper functions              #
###############################################

def buildParentMap(tree):
    """
    Builds a mapping of child elements to their parent elements in an XML tree.
    This function is useful for cases where you need to navigate from a child element
    up to its parent element, as the ElementTree API does not provide this functionality directly.

    Parameters:
      tree (ElementTree): An XML ElementTree object.

    Returns:
      dict: A dictionary where keys are child elements and values are their respective parent elements.
      
    Usage:
       To build the map:
          tree = ET.parse(InputFile)
          parentMap = buildParentMap(tree)
       Then, whenever you need a parent of an element:
          parent = getParent(someElement, parentMap)
    
    """
    return {c: p for p in tree.iter() for c in p}

def getParent(et, parentMap):
    """
    Retrieves the parent element of a given element from the parent map.

    Parameters:
      et (Element): The XML element whose parent is to be found.
      parentMap (dict): A dictionary mapping child elements to their parents.

    Returns:
      Element: The parent element of the given element. Returns None if the parent is not found.
    """
    return parentMap.get(et)

def processElement(elem, bookInfo, WordOrder, parentMap):
    """
    Processes an XML element to extract and augment its attributes with additional data.
    This function adds new attributes to an element and modifies existing ones based on the provided
    book information, word order, and parent map. It also collects hierarchical information
    about the element's ancestors in the XML structure.

    Parameters:
      elem (Element): The XML element to be processed.
      bookInfo (tuple): A tuple containing information about the book (long name, book number, short name).
      WordOrder (int): The order of the word in the current processing context.
      parentMap (dict): A dictionary mapping child elements to their parents.

    Returns:
      tuple: A tuple containing the updated attributes of the element and the next word order.
    """
    global CollectedItems
    LeafRef = re.sub(r'[!: ]', " ", elem.attrib.get('ref')).split()
    elemAttrib = dict(elem.attrib)    # Create a copy of the attributes using dict()

    # Adding new or modifying existing attributes
    elemAttrib.update({
        'wordOrder': WordOrder,
        'LeafName': elem.tag,
        'word': elem.text,
        'bookLong': bookInfo[0],
        'bookNum': int(bookInfo[1]),
        'bookShort': bookInfo[2],
        'chapter': int(LeafRef[1]),
        'verse': int(LeafRef[2]),
        'parents': 0  # Initialize 'parents' attribute
    })

    parentnode = getParent(elem, parentMap)
    index = 0
    while parentnode is not None:
        index += 1
        parent_attribs = {
            f'Parent{index}Name': parentnode.tag,
            f'Parent{index}Type': parentnode.attrib.get('type'),
            f'Parent{index}Appos': parentnode.attrib.get('appositioncontainer'),
            f'Parent{index}Class': parentnode.attrib.get('class'),
            f'Parent{index}Rule': parentnode.attrib.get('rule'),
            f'Parent{index}Role': parentnode.attrib.get('role'),
            f'Parent{index}Cltype': parentnode.attrib.get('cltype'),
            f'Parent{index}Unit': parentnode.attrib.get('unit'),
            f'Parent{index}Junction': parentnode.attrib.get('junction'),
            f'Parent{index}SN': parentnode.attrib.get('SN'),
            f'Parent{index}WGN': parentnode.attrib.get('WGN')
        }
        elemAttrib.update(parent_attribs)
        parentnode = getParent(parentnode, parentMap)

    elemAttrib['parents'] = index

    CollectedItems += 1
    return elemAttrib, WordOrder + 1

def fixAttributeId(tree):
    """
    Renames attributes in an XML tree that match the pattern '{*}id' to 'id'.

    Parameters:
      tree (lxml.etree._ElementTree): The XML tree to be processed.

    Returns:
      None: The function modifies the tree in-place and does not return anything.
    """
    # Regex pattern to match attributes like '{...}id'
    pattern = re.compile(r'\{.*\}id')
    for element in tree.iter():
        attributes_to_rename = [attr for attr in element.attrib if pattern.match(attr)]
        for attr in attributes_to_rename:
            element.attrib['id'] = element.attrib.pop(attr)

###############################################
#             The main routine                #
###############################################

# Process books
print ('Extract data from XML files and store it in pickle files')
overalTime = time.time()
for bo, bookInfo in bo2book.items():
    CollectedItems = 0
    SentenceNumber = 0
    WordGroupNumber = 0
    dataList = []  # List to store data dictionaries

    InputFile = os.path.join(XmlDir, f'{bo}.xml')
    OutputFile = os.path.join(PklDir, f'{bo}.pkl')
    print(f'\tProcessing {bookInfo[0]} at {InputFile} ', end='')

    try:
        tree = ET.parse(InputFile)
        fixAttributeId(tree)
        parentMap = buildParentMap(tree)
    except Exception as e:
        print(f"Error parsing XML file {InputFile}: {e}")
        continue

    start_time = time.time()

    for elem in tree.iter():
        if elem.tag == 'sentence':
            SentenceNumber += 1
            elem.set('SN', str(SentenceNumber))
        elif elem.tag == 'error':  # workaround for one <error> node found in the source XML, which is in fact a <wg> node failing analysis 
            elem.tag = 'wg'
        if elem.tag == 'wg':
            WordGroupNumber += 1
            elem.set('WGN', str(WordGroupNumber))
        if elem.tag == 'w':
            elemAttrib, WordOrder = processElement(elem, bookInfo, WordOrder, parentMap)
            dataList.append(elemAttrib)

    fullDataFrame = pd.DataFrame(dataList)  # Create DataFrame once after processing all elements
    
    # Open the file using  a context manager
    with open(OutputFile, 'wb') as output:
        pickle.dump(fullDataFrame, output)

    print(f"Found {CollectedItems} items in {time.time() - start_time:.2f} seconds.")
    
print(f'Finished in {time.time() - overalTime:.2f} seconds.')


# # 3 - Optionaly export to aid investigation<a class="anchor" id="bullet3"></a>
# ##### [Back to TOC](#TOC)
# 
# This step is optional. It will allow for manual examining the input data to the Text-Fabric conversion script.
# 

# ## 3.1 - Export to Excel format<a class="anchor" id="bullet3x1"></a>
# ##### [Back to TOC](#TOC)
# 
# Warning: Exporting of pandas dataframes to Excel format is **very slow**.

# In[17]:


# Pre-construct the base paths for input and output since they remain constant
baseInputPath = os.path.join(PklDir, '{}.pkl')
baseOutputPath = os.path.join(XlsxDir, '{}.xlsx')

print('Exporting Pickle files to Excel format. Please be patient. This export takes significant time')
overalTime = time.time()
errorCondition=False

for directory in (PklDir,XlsxDir):
    if not os.path.exists(directory):
        print(f"Script aborted. The directory '{directory}' does not exist.")
        errorCondition=True

# Load books in order
if not errorCondition:
    for bo in bo2book:
        startTime = time.time()

        # Use formatted strings for file names
        inputFile = baseInputPath.format(bo)
        outputFile = baseOutputPath.format(bo)

        print(f'\tProcessing {inputFile} ...', end='')

        # Use context manager for reading pickle file
        with open(inputFile, 'rb') as pklFile:
            df = pickle.load(pklFile)

        # Export to Excel
        df.to_excel(outputFile, index=False)

        # Print the time taken for processing
        print(f' done in {time.time() - startTime:.2f} seconds.')
    
    print(f'Finished in {time.time() - overalTime:.2f} seconds.')


# ## 3.2 - Export to CSV format<a class="anchor" id="bullet3x2"></a>
# ##### [Back to TOC](#TOC)
# 
# Exporting the pandas datframes to CSV format is fast. This file can easily be loaded into excel.

# In[7]:


# Pre-construct the base paths for input and output since they remain constant
baseInputPath = os.path.join(PklDir, '{}.pkl')
baseOutputPath = os.path.join(CsvDir, '{}.csv')

print ('Exporting pickle files to CSV formated files')
overalTime = time.time()
errorCondition=False

for directory in (PklDir,CsvDir):
    if not os.path.exists(directory):
        print(f"Script aborted. The directory '{directory}' does not exist.")
        errorCondition=True

# Load books in order
if not errorCondition:
    for bo in bo2book:
        start_time = time.time()
 
        # Use formatted strings for file names
        inputFile = baseInputPath.format(bo)
        outputFile = baseOutputPath.format(bo)
    
        print(f'\tProcessing {inputFile} ...', end='')
    
        try:
            with open(inputFile, 'rb') as pklFile:
               df = pickle.load(pklFile)
            df.to_csv(outputFile, index=False)
            print(f' done in {time.time() - start_time:.2f} seconds.')
        except pickle.UnpicklingError as e:
            print(f"\n\tError while loading {inputFile}: {e}")
            continue
    
    print(f'\nFinished in {time.time() - overalTime:.2f} seconds.')


# # 4 - Text-Fabric dataset production from pickle input<a class="anchor" id="bullet4"></a>
# ##### [Back to TOC](#TOC)
# 
# 

# ## 4.1 - Explanation<a class="anchor" id="bullet4x1"></a>
# ##### [Back to TOC](#TOC)
# 
# This script creates the Text-Fabric files by recursive calling the TF walker function.
# API info: https://annotation.github.io/text-fabric/tf/convert/walker.html
# 
# The pickle files created by the script in  <a href="#bullet2x3">section 2.3</a> are stored on Github location [/resources/pickle](https://github.com/tonyjurg/Nestle1904LFT/tree/main/resources/pickle).
# 
# Explanatory notes about the data interpretation logic are incorporated within the Python code of the director function.

# ## 4.2 - Running the TF walker function<a class="anchor" id="bullet4x2"></a>
# ##### [Back to TOC](#TOC)

# In[9]:


# Load specific set of variables for the walker

from tf.fabric import Fabric
from tf.convert.walker import CV

# setting some TF specific variables
BASE = os.path.expanduser('~/github')
ORG = 'tonyjurg'
REPO = 'Nestle1904LFT'
RELATIVE = 'tf'
TF_DIR = os.path.expanduser(f'{BASE}//{ORG}//{REPO}//{RELATIVE}')
VERSION = f'{scriptVersion}'
TF_PATH = f'{TF_DIR}//{VERSION}'
TF = Fabric(locations=TF_PATH, silent=False)
cv = CV(TF)


# In[10]:


###############################################
#   Common helper functions                   #
###############################################

def sanitize(input):
    """
    Sanitizes the input data to handle missing or undefined values.

    This function is used to ensure that float values and None types are converted to empty strings.
    Other data types are returned as-is. This is particularly useful in data processing and conversion
    tasks where missing data needs to be handled gracefully.

    Parameters:
      input: The data input which can be of any type.

    Returns:
      str: An empty string if the input is a float or None, otherwise returns the input as-is.
    """
    if isinstance(input, float) or isinstance(input, type(None)):
        return ''
    else:
        return input


def ExpandRole(input):
    """
    Expands syntactic role abbreviations into their full descriptive names.

    This function is particularly useful in parsing and interpreting syntactic structures, especially
    in the context of language processing. The expansion is based on the syntactic categories at the clause
    level as described in "MACULA Greek Treebank for the Nestle 1904 Greek New Testament" (page 5 & 6, section 2.4).

    Parameters:
      input (str): Abbreviated syntactic role label.

    Returns:
      str: The expanded, full descriptive name of the syntactic role. Returns an empty string for unrecognized inputs.
    """
    roleExpansions = {
        "adv": 'Adverbial',
        "io":  'Indirect Object',
        "o":   'Object',
        "o2":  'Second Object',
        "s":   'Subject',
        "p":   'Predicate',
        "v":   'Verbal',
        "vc":  'Verbal Copula',
        "aux": 'Auxiliar'
    }
    return roleExpansions.get(input, '')

def ExpandSP(input):
    """
    Expands Part of Speech (POS) label abbreviations into their full descriptive names.

    This function is utilized for enriching text data with clear, descriptive POS labels.
    The expansions are based on the syntactic categories at the word level as described in 
    "MACULA Greek Treebank for the Nestle 1904 Greek New Testament" (page 6 & 7, section 2.2).

    Parameters:
      input (str): Abbreviated POS label.

    Returns:
      str: The expanded, full descriptive name of the POS label. Returns an empty string for unrecognized inputs.
    """
    posExpansions = {
        'adj':  'Adjective',
        'conj': 'Conjunction',
        'det':  'Determiner',
        'intj': 'Interjection',
        'noun': 'Noun',
        'num':  'Numeral',
        'prep': 'Preposition',
        'ptcl': 'Particle',
        'pron': 'Pronoun',
        'verb': 'Verb'
    }
    return posExpansions.get(input, '')



def removeAccents(text):
    """
    Removes diacritical marks (accents) from Greek words or any text.

    This function is particularly useful in text processing where diacritical marks need to be
    removed, such as in search functionality, normalization, or comparison of strings. It leverages
    Unicode normalization to decompose characters into their base characters and diacritics, and then
    filters out the diacritics.

    Note: This function can be applied to any text where Unicode normalization is applicable.

    Parameters:
      text (str): The text from which accents/diacritical marks need to be removed.

    Returns:
      str: The input text with all diacritical marks removed.
    """
    return ''.join(c for c in unicodedata.normalize('NFD', text) if unicodedata.category(c) != 'Mn')

###############################################
#          The director routine               #
###############################################

def director(cv):
    
    ###############################################
    #   Innitial setup of data etc.               #
    ###############################################
    NoneType = type(None)      # needed as tool to validate certain data
    IndexDict = {}             # init an empty dictionary
    WordGroupDict={}           # init a dummy dictionary
    PrevWordGroupSet = WordGroupSet = []
    PrevWordGroupList = WordGroupList = []
    RootWordGroup = 0
    WordNumber=FoundWords=WordGroupTrack=0
    # The following is required to recover succesfully from an abnormal condition
    # in the LowFat tree data where a <wg> element is labeled as <error>
    # this number is arbitrary but should be high enough not to clash with 'real' WG numbers
    DummyWGN=200000  # first dummy WG number
    
    # Following variables are used for textual critical data 
    criticalMarkCharacters = "[]()—"
    punctuationCharacters = ",.;·"
    translationTableMarkers = str.maketrans("", "", criticalMarkCharacters)
    translationTablePunctuations = str.maketrans("", "", punctuationCharacters)
    punctuations=('.',',',';','·')
    
    for bo,bookinfo in bo2book.items(): 
        
        ###############################################
        #   start of section executed for each book   #
        ###############################################
        
        # note: bookinfo is a list! Split the data
        Book        = bookinfo[0]  
        BookNumber  = int(bookinfo[1])
        BookShort   = bookinfo[2]
        BookLoc     = os.path.join(PklDir, f'{bo}.pkl') 
        
        # load  data for this book into a dataframe. 
        # make sure wordorder is correct
        print(f'\tWe are loading {BookLoc}...')
        pkl_file = open(BookLoc, 'rb')
        df_unsorted = pickle.load(pkl_file)
        pkl_file.close()
        
        '''
        Fill dictionary of column names for this book 
        sort to ensure proper wordorder
        '''
        ItemsInRow=1
        for itemName in df_unsorted.columns.to_list():
            IndexDict.update({'i_{}'.format(itemName): ItemsInRow})
            # This is to identify the collumn containing the key to sort upon
            if itemName=="id": SortKey=ItemsInRow-1
            ItemsInRow+=1
        
        df=df_unsorted.sort_values(by=df_unsorted.columns[SortKey])
        del df_unsorted

        # Set up nodes for new book
        ThisBookPointer = cv.node('book')
        cv.feature(ThisBookPointer, book=Book, booknumber=BookNumber, bookshort=BookShort)
        
        ThisChapterPointer = cv.node('chapter')
        cv.feature(ThisChapterPointer, chapter=1, book=Book)
        PreviousChapter=1
        
        ThisVersePointer = cv.node('verse')
        cv.feature(ThisVersePointer, verse=1, chapter=1, book=Book)
        PreviousVerse=1
        
        ThisSentencePointer = cv.node('sentence')
        cv.feature(ThisSentencePointer, sentence=1, headverse=1, chapter=1, book=Book)
        PreviousSentence=1

        ###############################################
        # Iterate through words and construct objects #
        ###############################################
        
        for row in df.itertuples():
            WordNumber += 1
            FoundWords +=1
                       
            # Detect and act upon changes in sentences, verse and chapter 
            # the order of terminating and creating the nodes is critical: 
            # close verse - close chapter - open chapter - open verse 
            NumberOfParents = sanitize(row[IndexDict.get("i_parents")])
            ThisSentence=int(row[IndexDict.get("i_Parent{}SN".format(NumberOfParents-1))])
            ThisVerse = sanitize(row[IndexDict.get("i_verse")])
            ThisChapter = sanitize(row[IndexDict.get("i_chapter")])
            
            if (ThisVerse!=PreviousVerse):
                cv.terminate(ThisVersePointer)
            
            if (ThisSentence!=PreviousSentence):
                cv.terminate(ThisSentencePointer)
                             
    
            if (ThisChapter!=PreviousChapter):
                cv.terminate(ThisChapterPointer)
                PreviousChapter = ThisChapter
                ThisChapterPointer = cv.node('chapter')
                cv.feature(ThisChapterPointer, chapter=ThisChapter, book=Book)
                
            if (ThisVerse!=PreviousVerse):
                PreviousVerse = ThisVerse  
                ThisVersePointer = cv.node('verse')
                cv.feature(ThisVersePointer, verse=ThisVerse, chapter=ThisChapter, book=Book)
                
            if (ThisSentence!=PreviousSentence):
                PreviousSentence=ThisSentence
                ThisSentencePointer = cv.node('sentence')
                cv.feature(ThisSentencePointer, sentence=ThisSentence, headverse=ThisVerse, chapter=ThisChapter, book=Book)       
        
            ###############################################
            #         analyze and process <WG> tags       #
            ###############################################
                    
            PrevWordGroupList=WordGroupList
            WordGroupList=[]  # stores current active WordGroup numbers
            
            for i in range(NumberOfParents-2,0,-1): # important: reversed itteration!
                
                _WGN=int(row[IndexDict.get("i_Parent{}WGN".format(i))])                             
                if _WGN!='':
                     WGN=int(_WGN)
                if WGN!='':
                     WGclass=sanitize(row[IndexDict.get("i_Parent{}Class".format(i))])
                     WGrule=sanitize(row[IndexDict.get("i_Parent{}Rule".format(i))])
                     WGtype=sanitize(row[IndexDict.get("i_Parent{}Type".format(i))])
                     if WGclass==WGrule==WGtype=='':
                         WGclass='empty'
                     else:
                         #print ('---',WordGroupList)
                         if WGN not in WordGroupList:
                             WordGroupList.append(WGN) 
                             #print(f'append WGN={WGN}')
                         WordGroupDict[(WGN,0)]=WGN
                         if WGrule[-2:]=='CL' and WGclass=='':  
                             WGclass='cl*'  # to simulate the way Logos presents this condition
                         WordGroupDict[(WGN,6)]=WGclass
                         WordGroupDict[(WGN,1)]=WGrule
                         WordGroupDict[(WGN,8)]=WGtype
                         WordGroupDict[(WGN,3)]=sanitize(row[IndexDict.get("i_Parent{}Junction".format(i))])
                         WordGroupDict[(WGN,2)]=sanitize(row[IndexDict.get("i_Parent{}Cltype".format(i))])
                         WordGroupDict[(WGN,7)]=sanitize(row[IndexDict.get("i_Parent{}Role".format(i))])
                         WordGroupDict[(WGN,9)]=sanitize(row[IndexDict.get("i_Parent{}Appos".format(i))])  # appos is not pressent any more in the newer dataset. kept here for the time being...
                         WordGroupDict[(WGN,10)]=NumberOfParents-1-i  # = number of parent wordgroups     
            if not PrevWordGroupList==WordGroupList:
                #print ('##',PrevWordGroupList,WordGroupList,NumberOfParents)
                if RootWordGroup != WordGroupList[0]:
                    RootWordGroup = WordGroupList[0]
                    SuspendableWordGoupList = []
                    # we have a new sentence. rebuild suspendable wordgroup list
                    # some cleaning of data may be added here to save on memmory... 
                    #for k in range(6): del WordGroupDict[item,k]
                for item in reversed(PrevWordGroupList):
                   if (item not in WordGroupList):
                         # CLOSE/SUSPEND CASE
                         SuspendableWordGoupList.append(item)
                         #print ('\n close: '+str(WordGroupDict[(item,0)])+' '+ WordGroupDict[(item,6)]+' '+ WordGroupDict[(item,1)]+' '+WordGroupDict[(item,8)],end=' ')                     
                         cv.terminate(WordGroupDict[(item,4)])
                for item in WordGroupList:
                    if (item not in PrevWordGroupList):
                         if (item in SuspendableWordGoupList):
                              # RESUME CASE
                              #print ('\n resume: '+str(WordGroupDict[(item,0)])+' '+ WordGroupDict[(item,6)]+' '+WordGroupDict[(item,1)]+' '+WordGroupDict[(item,8)],end=' ') 
                              cv.resume(WordGroupDict[(item,4)])
                         else:
                              # CREATE CASE
                              #print ('\n create: '+str(WordGroupDict[(item,0)])+' '+ WordGroupDict[(item,6)]+' '+ WordGroupDict[(item,1)]+' '+WordGroupDict[(item,8)],end=' ')
                              WordGroupDict[(item,4)]=cv.node('wg')
                              WordGroupDict[(item,5)]=WordGroupTrack
                              WordGroupTrack += 1
                              cv.feature(WordGroupDict[(item,4)], wgnum=WordGroupDict[(item,0)], junction=WordGroupDict[(item,3)], 
                                         clausetype=WordGroupDict[(item,2)], wgrule=WordGroupDict[(item,1)], wgclass=WordGroupDict[(item,6)], 
                                         wgrole=WordGroupDict[(item,7)],wgrolelong=ExpandRole(WordGroupDict[(item,7)]),
                                         wgtype=WordGroupDict[(item,8)],wglevel=WordGroupDict[(item,10)])
     
            # These roles are performed either by a WG or just a single word.
            Role=row[IndexDict.get("i_role")]
            ValidRoles=["adv","io","o","o2","s","p","v","vc","aux"]
            DistanceToRoleClause=0
            if isinstance (Role,str) and Role in ValidRoles: 
                # Role is assign to this word (uniqely)
                WordRole=Role
                WordRoleLong=ExpandRole(WordRole)
            else:
                # Role details needs to be taken from some uptree wordgroup 
                WordRole=WordRoleLong=''
                for item in range(1,NumberOfParents-1):
                    Role = sanitize(row[IndexDict.get("i_Parent{}Role".format(item))])
                    if isinstance (Role,str) and Role in ValidRoles: 
                        WordRole=Role        
                        WordRoleLong=ExpandRole(WordRole)
                        DistanceToRoleClause=item
                        break
                        
            # Find the number of the WG containing the clause definition
            for item in range(1,NumberOfParents-1):
                WGrule = sanitize(row[IndexDict.get("i_Parent{}Rule".format(item))])
                if row[IndexDict.get("i_Parent{}Class".format(item))]=='cl' or WGrule[-2:]=='CL':  
                    ContainedClause=sanitize(row[IndexDict.get("i_Parent{}WGN".format(item))])
                    break

            ###############################################
            #         analyze and process <W> tags        #
            ###############################################
                
            # Determine syntactic categories at word level. 
            PartOfSpeech=sanitize(row[IndexDict.get("i_class")])
            PartOfSpeechFull=ExpandSP(PartOfSpeech)
            
            # The folling part of code reproduces feature 'word' and 'after' that are
            # currently containing incorrect data in a few specific cases.
            # See https://github.com/tonyjurg/Nestle1904LFT/blob/main/resources/identifying_odd_afters.ipynb
            # Get the word details and detect presence of punctuations
            # it also creates the textual critical features

            rawWord=sanitize(row[IndexDict.get("i_unicode")])
            cleanWord= rawWord.translate(translationTableMarkers)
            rawWithoutPunctuations=rawWord.translate(translationTablePunctuations)
            markBefore=markAfter=PunctuationMarkOrder=''
            if cleanWord[-1] in punctuations:
                punctuation=cleanWord[-1]
                after=punctuation+' '
                word=cleanWord[:-1]
            else:
                after=' '
                word=cleanWord
                punctuation=''
            if rawWithoutPunctuations!=word:
                markAfter=markBefore=''
                if rawWord.find(word)==0:
                    markAfter=rawWithoutPunctuations.replace(word,"")
                    if punctuation!='':
                        if rawWord.find(markAfter)-rawWord.find(punctuation)>0:
                            PunctuationMarkOrder="3" # punct. before mark
                        else:
                            PunctuationMarkOrder="2" # punct. after mark.
                    else:
                        PunctuationMarkOrder="1" #no punctuation, mark after word
                else:
                    markBefore=rawWithoutPunctuations.replace(word,"")
                    PunctuationMarkOrder="0" #mark is before word
                    
            # Some attributes are not present inside some (small) books. The following is to prevent exceptions.
            degree='' 
            if 'i_degree'  in IndexDict: 
                degree=sanitize(row[IndexDict.get("i_degree")]) 
            subjref=''
            if 'i_subjref' in IndexDict: 
                subjref=sanitize(row[IndexDict.get("i_subjref")]) 

                    
            # Create the word slots
            this_word = cv.slot()
            cv.feature(this_word, 
                    after=         after,
                    unicode=       rawWord,
                    word=          word,
                    wordtranslit=  unidecode(word),
                    wordunacc=     removeAccents(word),
                    punctuation=   punctuation,
                    markafter=     markAfter,
                    markbefore=    markBefore,
                    markorder=     PunctuationMarkOrder,
                    monad=         FoundWords,
                    orig_order=    sanitize(row[IndexDict.get("i_wordOrder")]),
                    book=          Book,
                    booknumber=    BookNumber,
                    bookshort=     BookShort,
                    chapter=       ThisChapter,
                    ref=           sanitize(row[IndexDict.get("i_ref")]),
                    sp=            PartOfSpeech,
                    sp_full=       PartOfSpeechFull,
                    verse=         ThisVerse,
                    sentence=      ThisSentence,
                    normalized=    sanitize(row[IndexDict.get("i_normalized")]),
                    morph=         sanitize(row[IndexDict.get("i_morph")]),
                    strongs=       sanitize(row[IndexDict.get("i_strong")]),
                    lex_dom=       sanitize(row[IndexDict.get("i_domain")]),
                    ln=            sanitize(row[IndexDict.get("i_ln")]),
                    gloss=         sanitize(row[IndexDict.get("i_gloss")]),
                    gn=            sanitize(row[IndexDict.get("i_gender")]),
                    nu=            sanitize(row[IndexDict.get("i_number")]),
                    case=          sanitize(row[IndexDict.get("i_case")]),
                    lemma=         sanitize(row[IndexDict.get("i_lemma")]),
                    person=        sanitize(row[IndexDict.get("i_person")]),
                    mood=          sanitize(row[IndexDict.get("i_mood")]),
                    tense=         sanitize(row[IndexDict.get("i_tense")]),
                    number=        sanitize(row[IndexDict.get("i_number")]),
                    voice=         sanitize(row[IndexDict.get("i_voice")]),
                    degree=        degree,
                    type=          sanitize(row[IndexDict.get("i_type")]),
                    reference=     sanitize(row[IndexDict.get("i_ref")]),     
                    subj_ref=      subjref,
                    nodeID=        sanitize(row[IndexDict.get("i_id")]),
                    wordrole=      WordRole,
                    wordrolelong=  WordRoleLong,
                    wordlevel=     NumberOfParents-1,
                    roleclausedistance = DistanceToRoleClause,
                    containedclause = ContainedClause
                    )
            cv.terminate(this_word)   
   
        
        '''
        wrap up the book. At the end of the book we need to close all nodes in proper order.
        '''   
        # close all open WordGroup nodes
        for item in WordGroupList:
            #cv.feature(WordGroupDict[(item,4)], add some stats?)
            cv.terminate(WordGroupDict[item,4])

        cv.terminate(ThisSentencePointer)
        cv.terminate(ThisVersePointer)
        cv.terminate(ThisChapterPointer)      
        cv.terminate(ThisBookPointer)

        # clear dataframe for this book, clear the index dictionary
        del df
        IndexDict.clear()
        #gc.collect()
        
        ###############################################
        #    end of section executed for each book    #
        ###############################################

    ###############################################
    #      end of director function               #
    ###############################################
        
###############################################
#            Output definitions               #
###############################################

# define TF dataset granularity
slotType = 'word'  

# dictionary of config data for sections and text formats
otext = { 
        'fmt:text-orig-full':     '{word}{after}',
        'fmt:text-normalized':    '{normalized}{after}',
        'fmt:text-unaccented':    '{wordunacc}{after}',
        'fmt:text-transliterated':'{wordtranslit}{after}', 
        'fmt:text-critical':      '{unicode} ',
        'sectionTypes':'book,chapter,verse',
        'sectionFeatures':'book,chapter,verse',
        'structureFeatures': 'book,chapter,verse',
        'structureTypes': 'book,chapter,verse',
        }

# configure provenance metadata
generic = {  # dictionary of metadata meant for all features
        'textFabricVersion': '{}'.format(VERSION),  #imported from tf.parameter
        'xmlSourceLocation': 'https://github.com/tonyjurg/Nestle1904LFT/tree/main/resources/xml/20240210',
        'xmlSourceDate': 'February 10, 2024',
        'author': 'Evangelists and apostles',
        'availability': 'Creative Commons Attribution 4.0 International (CC BY 4.0)',
        'converters': 'Tony Jurg',
        'converterSource': 'https://github.com/tonyjurg/Nestle1904LFT/tree/main/resources/converter',
        'converterVersion': '{} ({})'.format(scriptVersion,scriptDate),
        'dataSource': 'MACULA Greek Linguistic Datasets, available at https://github.com/Clear-Bible/macula-greek/tree/main/Nestle1904/nodes',
        'editors': 'Eberhart Nestle (1904)',
        'sourceDescription': 'Greek New Testment (British Foreign Bible Society, 1904)',
        'sourceFormat': 'XML (Low Fat tree XML data)',
        'title': 'Greek New Testament (Nestle1904LFT)'
         }

# set datatype of feature (if not listed here, they are ususaly strings)
intFeatures = {  
         'booknumber',
         'chapter',
         'verse',
         'sentence',
         'wgnum',
         'orig_order',
         'monad',
         'wglevel'
         }

# per feature dicts with metadata
# icon provides guidance on feature maturity (✅ = trustworthy, 🆗  = usable, ⚠️ = be carefull when using)
featureMeta = {  
                 'after':       {'description': '✅ Characters (eg. punctuations) following the word'},
                 'book':        {'description': '✅ Book name (in English language)'},
                 'booknumber':  {'description': '✅ NT book number (Matthew=1, Mark=2, ..., Revelation=27)'},
                 'bookshort':   {'description': '✅ Book name (abbreviated)'},
                 'chapter':     {'description': '✅ Chapter number inside book'},
                 'verse':       {'description': '✅ Verse number inside chapter'},
                 'headverse':   {'description': '✅ Start verse number of a sentence'},
                 'sentence':    {'description': '✅ Sentence number (counted per chapter)'},
                 'type':        {'description': '✅ Wordgroup type information (e.g.verb, verbless, elided, minor)'},
                 'wgrule':      {'description': '✅ Wordgroup rule information (e.g. Np-Appos, ClCl2, PrepNp)'},
                 'orig_order':  {'description': '✅ Word order (in source XML file)'},
                 'monad':       {'description': '✅ Monad (smallest token matching word order in the corpus)'},
                 'word':        {'description': '✅ Word as it appears in the text (excl. punctuations)'},
                 'wordtranslit':{'description': '🆗 Transliteration of the text (in latin letters, excl. punctuations)'},
                 'wordunacc':   {'description': '✅ Word without accents (excl. punctuations)'},
                 'unicode':     {'description': '✅ Word as it apears in the text in Unicode (incl. punctuations)'},
                 'punctuation': {'description': '✅ Punctuation after word'},
                 'markafter':   {'description': '🆗 Text critical marker after word'},
                 'markbefore':  {'description': '🆗 Text critical marker before word'},
                 'markorder':   {'description': ' Order of punctuation and text critical marker'},
                 'ref':         {'description': '✅ Value of the ref ID (taken from XML sourcedata)'},
                 'sp':          {'description': '✅ Part of Speech (abbreviated)'},
                 'sp_full':     {'description': '✅ Part of Speech (long description)'}, 
                 'normalized':  {'description': '✅ Surface word with accents normalized and trailing punctuations removed'},
                 'lemma':       {'description': '✅ Lexeme (lemma)'},
                 'morph':       {'description': '✅ Morphological tag (Sandborg-Petersen morphology)'},
                                 # see also discussion on relation between lex_dom and ln 
                                 # @ https://github.com/Clear-Bible/macula-greek/issues/29
                 'lex_dom':     {'description': '✅ Lexical domain according to Semantic Dictionary of Biblical Greek, SDBG (not present everywhere?)'},
                 'ln':          {'description': '✅ Lauw-Nida lexical classification (not present everywhere?)'},
                 'strongs':     {'description': '✅ Strongs number'},
                 'gloss':       {'description': '✅ English gloss'},
                 'gn':          {'description': '✅ Gramatical gender (Masculine, Feminine, Neuter)'},
                 'nu':          {'description': '✅ Gramatical number (Singular, Plural)'},
                 'case':        {'description': '✅ Gramatical case (Nominative, Genitive, Dative, Accusative, Vocative)'},
                 'person':      {'description': '✅ Gramatical person of the verb (first, second, third)'},
                 'mood':        {'description': '✅ Gramatical mood of the verb (passive, etc)'},
                 'tense':       {'description': '✅ Gramatical tense of the verb (e.g. Present, Aorist)'},
                 'number':      {'description': '✅ Gramatical number of the verb (e.g. singular, plural)'},
                 'voice':       {'description': '✅ Gramatical voice of the verb (e.g. active,passive)'},
                 'degree':      {'description': '✅ Degree (e.g. Comparitative, Superlative)'},
                 'type':        {'description': '✅ Gramatical type  of noun or pronoun (e.g. Common, Personal)'},
                 'reference':   {'description': '✅ Reference (to nodeID in XML source data, not yet post-processes)'},
                 'subj_ref':    {'description': '🆗 Subject reference (to nodeID in XML source data, not yet post-processes)'},
                 'nodeID':      {'description': '✅ Node ID (as in the XML source data)'},
                 'junction':    {'description': '✅ Junction data related to a wordgroup'},
                 'wgnum':       {'description': '✅ Wordgroup number (counted per book)'},
                 'wgclass':     {'description': '✅ Class of the wordgroup (e.g. cl, np, vp)'},
                 'wgrole':      {'description': '✅ Syntactical role of the wordgroup (abbreviated)'},
                 'wgrolelong':  {'description': '✅ Syntactical role of the wordgroup (full)'},
                 'wordrole':    {'description': '✅ Syntactical role of the word (abbreviated)'},
                 'wordrolelong':{'description': '✅ Syntactical role of the word (full)'},
                 'wgtype':      {'description': '✅ Wordgroup type details (e.g. group, apposition)'},
                 'clausetype':  {'description': '✅ Clause type details (e.g. Verbless, Minor)'},
                 'wglevel':     {'description': '🆗 Number of the parent wordgroups for a wordgroup'},
                 'wordlevel':   {'description': '🆗 Number of the parent wordgroups for a word'},
                 'roleclausedistance': {'description': '⚠️ Distance to the wordgroup defining the syntactical role of this word'},
                 'containedclause': {'description': '🆗 Contained clause (WG number)'}
                 }


###############################################
#            the main function                #
###############################################

good = cv.walk(
    director,
    slotType,
    otext=otext,
    generic=generic,
    intFeatures=intFeatures,
    featureMeta=featureMeta,
    warn=True,
    force=True
)

if good:
  print ("done")


# ## 5 - Housekeeping<a class="anchor" id="bullet5"></a>
# ##### [Back to TOC](#TOC)

# ## 5.1 - Optionaly zip-up the pickle files<a class="anchor" id="bullet5x1"></a>
# ##### [Back to TOC](#TOC)
# 
# In order to save filespace, the pickle files can be zipped. The following will zip the pickle files and remove the original file. The removal of pickle files is important if their size is too large (i.e., more than 100Mb in size), leading to issues when uploaded to gitHub. Hence it is adviced to include the .pkl extention in the ignore list (.gitignore).

# In[8]:


# set variable if original pickle file needs to be removed or not
removeOriginal=False

import zipfile
import os

def zipTheFile(sourceFile, destinationFile, removeOriginal):
    """
    Create a zip file from the specified source file and optionally remove the source file.

    Parameters:
       sourceFile      (str) : The file path of the source file to be zipped.
       destinationFile (str) : The file path for the resulting zip file.
       removeOriginal  (bool): If True, the source file will be deleted after zipping.
    """
    # check for existance of the file to zip
    if not os.path.exists(sourceFile):
        print(f"\tSource file does not exist: {sourceFile}")
        return False
    # Get only the file name, not the full path
    fileNameOnly = os.path.basename(sourceFile)
    # Creating a zip file from the source file
    with zipfile.ZipFile(destinationFile, 'w', zipfile.ZIP_DEFLATED) as zipArchive:
        zipArchive.write(sourceFile,arcname=fileNameOnly)
    # Removing the source file if required
    if removeOriginal: 
        os.remove(sourceFile)
    return True

# Pre-construct the base paths for input and output since they remain constant
baseInputPath = os.path.join(PklDir, '{}.pkl')
baseOutputPath = os.path.join(PklDir, '{}.zip')

print('Zipping up all pickle files' + (' and removing them afterwards.' if removeOriginal else '.'))
overallTime = time.time()
errorOccurred = False

# Load books in order
for bo in bo2book:
    startTime = time.time()
 
    # Use formatted strings for file names
    inputFile = baseInputPath.format(bo)
    outputFile = baseOutputPath.format(bo)
    
    if not zipTheFile(inputFile, outputFile, removeOriginal):
        errorOccurred = True
        break
    else: 
        print(f'\tloading {inputFile}...', end='')
        print(f' Zipping done in {time.time() - startTime:.2f} seconds.')
    
if errorOccurred:
    print("Operation aborted due to an error (are all pickle files already zipped?).")
else:
    print(f'\nFinished in {time.time() - overallTime:.2f} seconds.')


# ## 5.2 - Inspect the created dataset<a class="anchor" id="bullet5x2"></a>
# ##### [Back to TOC](#TOC)
# 
# Perform some inspections on the newly created datase.

# In[12]:


# Load TF code

from tf.fabric import Fabric
from tf.app import use


# In[13]:


# load the app and data
N1904 = use ("tonyjurg/Nestle1904LFT", version=scriptVersion, checkout="clone", hoist=globals())


# ### 5.2.1 - Dump otype<a class="anchor" id="bullet5x2x1"></a>
# ##### [Back to TOC](#TOC)

# In[14]:


with open(f'{TF_PATH}/otype.tf') as fh:
  print(fh.read())


# ### 5.2.2 - Dump otext<a class="anchor" id="bullet5x2x2"></a>
# ##### [Back to TOC](#TOC)

# In[15]:


with open(f'{TF_PATH}/otext.tf') as fh:
  print(fh.read())


# ## 5.3 - Publish it on GitHub<a class="anchor" id="bullet5x3"></a>
# ##### [Back to TOC](#TOC)
# 
# The following section will first load the created Text-Fabric dataset. Then it will publish it on gitHub.

# In[16]:


# Define the repository
ORG = "tonyjurg"
REPO = "Nestle1904LFT"

# Added details for the release
MESSAGE = "New release"
DESCRIPTION = """
This release uses a new dataset. 

The main difference is in feature Strongs:
  * Some errors were corrected
  * composite words are now with two or more Strong values
  
This release has been published with the command `A.publish()`, a function in Text-Fabric.
"""


# In[18]:


N1904.publishRelease(3, message=MESSAGE, description=DESCRIPTION)


# In[ ]: