wegen/sosci2word/sosci2word.py

#!/usr/bin/python3
#
# sosci2word - Einlesen eines SoSci-Fragebogens und umwandeln in ein Word-Dokument
#
# (c) 2023 Hermann Voßeler
# This code is free software and can be used under the terms of the MIT License.
#

import sys
from xml.etree import ElementTree
from docx import Document
from demo import document



#-----------------------SETUP
SOSCI_XML='ThesisFragebogen.2023-04-19.xml'
WORD_TEMPLATE='Thesis-Fragebogen.docx'
WORD_OUTPUT='Thesis-Fragebogen-import.docx'
#-----------------------SETUP



def convertSosci2Word():
    # Schritt-1 : alle Frage-Definitionen aufsammeln
    questions = collectQuestionDefinitions()
    # Schritt-2 : die Fragen in der Reihenfolge vom Questionaire
    questionaire = buildQuestionaire(questions)
    ## TODO
    renderContentToWord(questionaire)
    ## TODO



class QuestionElement:
    def __init__(self, questionID, question):
        self.qid = questionID
        self.title = question.find('title').text
        self.desc = question.find('description').text
    
    def dumpBasis(self, type):
        return "##--%s--##\ndes=%s\ntitle=%s" % (type, self.desc, self.title)
    
    def dump(self):
        return self.dumpBasis("Basis")
    
    
    def render(self, document):
        pass # macht nichts in der Basisklasse



class LikertItem(QuestionElement):
    def __init__(self, questionID, question):
        QuestionElement.__init__(self, questionID, question)
        self.labels = []
        for entry in question.find('labels').findall('li'):
            code = int(entry.get('i'))
            text = entry.text
            self.labels.append((code,text))
        self.items = []
        for item in question.findall('item'):
            code = int(getSpecificID(item))
            text = item.find('text').text
            self.items.append((code,text))
    
    
    def dump(self):
        return ( self.dumpBasis('Likert')
               + "\nLabels: " + str(list("(%d,%s)" % (c,t) for c,t in self.labels))
               + "\nItems:  " + str(list("(%d,%s)" % (c,t) for c,t in self.items))
               ) 
             
    def render(self, document):
        questionText = self.desc + ": " + self.title
        document.add_paragraph(questionText, style='Fragebogen-Item')
        #
        # Tabelle mit den Antwortmöglichkeiten
        cntLabels = len(self.labels)
        cntItems = len(self.items)
        table = document.add_table(rows=cntItems+1, cols=cntLabels+1)
        # gesammte Tabelle mit dem korrekten Style auszeichnen
        for r in table.rows:
            for c in r.cells:
                c.paragraphs[0].style = 'Frabebogen-Wert'
        # alle Inhalts-Spalten ausfüllen
        # in die Kopfzeile kommt das richtige Label
        # in die Inhaltsfelder ein Symbol für das Antwort-Kästchen
        for c in range(1, cntLabels+1):
            table.cell(0,c).paragraphs[0].text = self.labels[c-1][1]
            for r in range(1, cntItems+1): 
                table.cell(r,c).paragraphs[0].text = '□'
        #
        for r in range(1, cntItems+1):
            table.cell(r,0).paragraphs[0].text = self.items[r-1][1]
        



class SelectItem(QuestionElement):
    def __init__(self, questionID, question):
        QuestionElement.__init__(self, questionID, question)
        self.items = []
        for item in question.findall('item'):
            code = int(getSpecificID(item))
            text = item.find('text').text
            self.items.append((code,text))
    
    def dump(self):
        return ( self.dumpBasis('Select')
               + "\nItems:  " + str(list("(%d,%s)" % (c,t) for c,t in self.items))
               ) 
    
    def render(self, document):
        questionText = self.title
        document.add_paragraph(questionText, style='Fragebogen-Item')
        #
        # Tabelle mit den Antwortmöglichkeiten
        cntItems = len(self.items)
        table = document.add_table(rows=cntItems, cols=2)
        # gesammte Tabelle mit dem korrekten Style auszeichnen
        for r in table.rows:
            for c in r.cells:
                c.paragraphs[0].style = 'Frabebogen-Wert'
        # alle Inhalts-Spalten ausfüllen
        # in die Inhaltsfelder links der Fragetext, rechts der Code-Wert
        for r in range(0, cntItems): 
            table.cell(r,0).paragraphs[0].text = self.items[r][1]        # text
            table.cell(r,1).paragraphs[0].text = str(self.items[r][0])   # code 




class TextElement(QuestionElement):
    def __init__(self, questionID, question):
        QuestionElement.__init__(self, questionID, question)
        lead = question.find('lead')
        if lead == None:
            self.text = ""
        else:
            self.text = lead.text

    def dump(self):
        return ( self.dumpBasis('Text')
               + "\nText:  " + self.text
               )
    
    def render(self, document):
        document.add_paragraph(self.desc, style='Überschrift-Anhang 3')
        if self.text != None:
            document.add_paragraph(self.text, style='Body Text')





def collectQuestionDefinitions():
    #
    questions = {}
    #
    print("Extracting Question Definitions...")
    root = ElementTree.parse(SOSCI_XML).getroot()
    for section in root.findall('section'):
        id_prefix = getSpecificID(section)
        
        for question in section.findall('question'):
            questionType = question.find('class').text
            id_suffix = getSpecificID(question)
            id_suffix = int(id_suffix)
            questionID = "%s%02d" % (id_prefix, id_suffix)
            if questionType == 'scale':
                element = LikertItem(questionID, question)
            elif questionType == 'select' or questionType == 'dropdown':
                element = SelectItem(questionID, question)
            elif questionType == 'TextContent':
                element = TextElement(questionID, question)
            elif  questionType == 'checkbox':
                pass # ignorieren wir
            else:
                __err("unbekannter QuestionType:"+questionType)
            
            questions[questionID] = element
            
    # Rückgabe: ein Dictionary mit allen bekannten Question-Elementen (keyed by ID)
    return questions



# Hilfsmittel für SoSci-XML: extrahiert ein ID-Element, das eingeschachtelt in <attributes.specific> liegt
def getSpecificID(element):
    attribspecific = element.find('attributes.specific')
    return next(attr.text for attr in attribspecific.findall('attr') if attr.get('id') == 'id')



def buildQuestionaire(questions):
    #
    print("Building Questionaire Pages...")
    #
    root = ElementTree.parse(SOSCI_XML).getroot()
    quest = root.find('questionnaire').find('attributes.specific')
    quest = next(a for a in quest.findall('attr') if a.get('id') == 'content')
    quest = quest.text
    
    questionaire = []
    
    questRoot = ElementTree.fromstring(quest)
    for page in questRoot.findall('page'):
        if page.get('ident') == 'I':
            continue                    # die erste Seite enthält den Einführungstext. Den fügen wir später manuell ein
        #
        info = page.find('info').text
        questionsOnThisPage = []
        for quest in page.findall('question'):
            questionID = quest.get('id')
            element = questions[questionID]
            questionsOnThisPage.append(element)
    
        questionaire.append((info, questionsOnThisPage))
    #
    return questionaire



def renderContentToWord(questionaire):
    #
    print('Integrate content in Word file "%s"....' % WORD_TEMPLATE)
    document = Document('Thesis-Fragebogen.docx')
    #
    renderQuestionaireTitle(document)
    for pageInfo, elements in questionaire:
        renderPageTitle(document, pageInfo)
        for element in elements:
            print(element.dump())
            element.render(document)
        
    print('Writing modified Word file as "%s"....' % WORD_OUTPUT)
    document.save(WORD_OUTPUT)
    #

# erzeugt ein neues Hauptkapitel im Anhang für den Fragebogen
def renderQuestionaireTitle(document):
    document.add_paragraph('Fragebogen (TODO Import)', style='Überschrift-Anhang 1')


# erzeugt einen Abschnitt im Anhang für die jeweilige Seite des Fragebogens
def renderPageTitle(document, title):
    document.add_paragraph(title, style='Überschrift-Anhang 2')




# ---- Hilfsfunktionen

def __err(text):
    print('---FAILURE----------------------')
    print(text)
    print('abort.')
    sys.exit(255)




if __name__=='__main__':
    convertSosci2Word()