Parser/OVI/EVT/xmlgat_to_EVT_parser.py

# %%
# Imports
import xml.etree.ElementTree as ET
import re
import json
# %%
# This is to handle the xmnls attribute in the TEI element in the templates
uri1 = "{http://www.tei-c.org/ns/1.0}"
namespaces = {
    '':         "http://www.tei-c.org/ns/1.0",
}
for prefix, uri in namespaces.items():
    ET.register_namespace(prefix, uri)
# Reference directories
basedir = '../../DATA/'
baseindir = basedir + 'OVI/datiniXML/xmlgat/'
baseoutdir = basedir + 'OVI/datiniXML/xmlevt/'
# %%
# Import lems list + xml info file

# get lem list as a json object
lemfile = basedir + 'OVI/datiniXML/power_lemmarioB.json'
lems = json.load(open(lemfile, 'r'))

# Get BiblioDatini.xml, extract a list of the <Biblio> nodes with ElementTree
infofile = basedir + 'OVI/datiniXML/BiblioDatini.xml'
infotree = ET.parse(infofile)
inforoot = infotree.getroot()
infoBiblioNodeList = list(inforoot.iter('Biblio'))
# %%
# FUNCTIONS TO PROCESS THE XMLGAT FILEs

# Get a lem index
def lemIndex(lem):
    for item in lems:
        if lem.attrib['n'] in item['coordinate']:
            return item['id']
    raise ValueError("code " + lem.attrib['n'] + " not found")

# Get the ElementTree node ('Element' class object) associated to an OVI sigla in the BiblioDatini.xml file
def getBiblioNodeBySigla(sigla):
    for node in infoBiblioNodeList:
        for child in node:
            if child.tag=='sigla' and child.text==sigla:
                return node

# Get the ElemenTree object of the whole xmlgat file corresponding to a given OVI sigla
def getLetterRootFromFile(filecode, inputdirectory=baseindir):
    fileName = inputdirectory + 'xmlgat.' +  filecode + '.xml'
    try:
        letterRoot = ET.parse(fileName).getroot()
    except ET.ParseError:
        with open(fileName, encoding="ISO-8859-1") as fp:
            xml_string = fp.read()
            xml_string = xml_string.replace('&Csic&c', "<hi rend='italic'>sic</hi>")
#            return xml_string
            letterRoot = ET.fromstring(xml_string)
    return letterRoot


##################################
# ELABORATING LEMS IN XMLGAT FILES
##################################

# PREMISE:
# in the xmlgat files, the <lem> tag doesn't surround lems:
# 
# 1. Single-word lems are in the 'tail' of the corr. lem tags, as in:
#  <lem>A_LEM
# with no closing </lem>
#
# 2. Multiple-word lems are inside <w> tags immediately following the <lem>, as in:
#  <lem><w>A MULTIWORD LEM</w>
# The body of the text is inside a single <div>

# This functions puts all lems inside a <lem> tag to make xmlgat files more standard xml-compliant and easier to process, also dropping the <w> tag.
# Basically:
#
# <lem>A_LEM  -->  <lem>A_LEM</lem>
# <lem><w>A MULTIWORD LEM</w>  -->  <lem>A MULTIWORD LEM</lem>
def surroundLems(letterRoot):

    textRoot = list(letterRoot.iter('div'))[0]

    texttags = [node for node in textRoot if node.tag == 'lem' or node.tag == 'w']

    doit = False
    for node in texttags:
        if doit and node.tag=='w':
            node.tag = 'NEWlem'
            node.attrib = prev_node.attrib
            prev_node.tag = 'OLDlem'
        if node.tag == 'lem' and node.tail != None:
            thelem = re.findall(r'\w+', node.tail)[0] # First word
            node.text = thelem
            node.tail = node.tail.replace(thelem, '')
            doit = False
        else:
            doit = True
        prev_node = node
    for node in textRoot.findall('OLDlem'):
        textRoot.remove(node)
    for node in textRoot.findall('NEWlem'):
        node.tag = 'lem'
    return textRoot


# This function tries to match a lem inside <lem> node (ElementTree Element object 'node'), by its attribute 'n',
# the a lem in the lem list, the json object 'lems'
def getLemByCode(lem):
    for item in lems:
        if lem.attrib['n'] in item['coordinate']:
            return item
    raise ValueError("code " + lem.attrib['n'] + " not found")


# Dictionary assigning to each OVI lem type a tag useful for the final TEI output
lemTypeDict = {'s.m.': "sostantivo maschile", 's.f.': 'sostantivo femminile', 'antr.': 'antroponimo', 'agg.': 'aggettivo', 'n.g.': 'nome di luogo', 'v.': 'verbo'}


# This function processes each lem attributes and adds more tags around the lem useful for the final TEI output
def redefineLems(textRoot, fileCode):
    for node in textRoot.iter('lem'):
        node.attrib['n'] = fileCode + '_' + node.attrib['n']
        thisLem = getLemByCode(node)
        lemRef = '#' + str(thisLem['id'])
        node.attrib.pop('n')
        lemType = thisLem['lemma']['categoria']
        #node.attrib['type'] = lemType
        #sub = ET.SubElement(node, 'lem2')
        if lemType=='antr.':
            node.tag = 'persName'
            node.attrib['ref'] = lemRef
            node.attrib['type'] = lemTypeDict[lemType]
        elif lemType=='n.g.':
            node.tag = 'placeName'
            node.attrib['ref'] = lemRef
            node.attrib['type'] = lemTypeDict[lemType]
        else:
            node.tag = 'lem'
            node.attrib['ref'] = lemRef
            for lemType in lemTypeDict:
                node.attrib['type'] = lemType
        #sub.text = node.text
        node.text = node.text
    for node in textRoot.iter('lem'):
        node.tag = 'lem'
        node.attrib['ref'] = lemRef
        node.attrib['type'] = lemTypeDict[lemType]
        
def replacepbcode(textRoot, fileCode):
    for ii, node in enumerate(textRoot.iter('pb')):
        node.attrib['n'] = fileCode + ' ' + str(ii + 1)

def surroundPages(textRoot):
    # Create a new, 'clean', root
    newRoot = ET.fromstring("<div/>")

    # Add a <p/> to the new root for each page in the old one
    for node in textRoot.iter('pb'):
        ET.SubElement(newRoot, 'p') 

    # Fill the pages in the new root
    page = None
    elementInPage = None
    for child in textRoot:
        if child.tag=='pb' and page is None:
            page = 0
            elementInPage = 0
        elif child.tag=='pb':
            page = page+1
            elementInPage = 0
        if page is not None and elementInPage is not None and child.tag!='milestone':
            newRoot[page].append(child)
            newRoot[page][elementInPage].tail = child.tail
            elementInPage = elementInPage+1
    
    return newRoot



# Get the letter template as a string
def getTemplateString():
    preLetterTemplateTree = ET.parse('pre_letter_template.xml')
    where = list( list( preLetterTemplateTree.getroot().iter(uri1+'body') )[0].iter(uri1+'div') )[0]
    #
    ET.SubElement(where, 'letterBody')
    #
    letterString = ET.tostring(preLetterTemplateTree.getroot(), encoding='unicode', method='xml')
    return letterString

# Get a file by OVI sigla 'filecode' as an ElementTree object, process its lem, transform it to string and format it
def newProcessFile(filecode, inputdirectory=baseindir):
    tree1 = getLetterRootFromFile(filecode, inputdirectory)
    #ET.dump(tree1)
    #
    textRoot1 = surroundLems(tree1)
    #
    redefineLems(textRoot1, filecode)
    #
    replacepbcode(textRoot1, filecode)
    #
    textRoot2 = surroundPages(textRoot1)
    #
    indent1 = "               "
    textString1 = ET.tostring(textRoot2, encoding='unicode', method='xml')
    textString2 = textString1.replace("<lb />\n", "<lb />\n"+indent1)
    return textString2
# %%
letterTemplateString = getTemplateString()
# %%
# Example
filecodeExample = 'j34'
with open('test.xml', 'w') as f1:
    newString = letterTemplateString.replace('<letterBody />', newProcessFile(filecodeExample))
    f1.write(newString)
# %%

# %%