doc2tosca.py 10.6 KB
Newer Older
carignani's avatar
carignani committed
#!/bin/python3
'''
Generate tosca definitions from Docx specfication
'''

import sys
carignani's avatar
carignani committed
import os
import re
from io import StringIO
from docx.table import Table
from docx.text.paragraph import Paragraph
from example import generate_examples_between

BASE_FILENAME = "etsi_nfv_sol001_{}_{}_types.yaml"
TOSCA_VERSION = "tosca_simple_yaml_1_2"
DEFAULT_TOSCA_VERSION = "tosca_simple_yaml_1_2"
SPEC_VERSION = "v2.6.1"
allowed_versions = ["v2.6.1", "v2.6.3", "v2.7.1", "v2.8.1", "v3.3.1"]
MODEL_NAMES = ['vnfd', 'nsd', 'pnfd', 'common']
HDR = '''tosca_definitions_version: {tosca_version}
description: ETSI NFV SOL 001 {model} types definitions version {spec_version}
metadata:
carignani's avatar
carignani committed
  - template_name: etsi_nfv_sol001_{model}_types
  - template_name: ETSI_NFV
  - template_version: {spec_version}
  {imports}
carignani's avatar
carignani committed
SUBSECTIONS = [
    "Artifact Types",
    "Data Types",
    "Capability Types",
    "Interface Types",
    "Requirements Types",
    "Relationship Types",
    "Interface Types",
    "Node Types",
    "Group Types",
    "Policy Types"
]

MODELS = {}
EXAMPLES = {}

class Section():
    '''
    Defines a section of the base document
    '''

    def __init__(self, from_id, to_id, title):
        self.from_id = from_id
        self.to_id = to_id
        self.is_annex = title.strip().startswith("Annex")
        
        if not self.is_annex:
            cleaned_title = title.strip().split("\t")
            self.title = cleaned_title[1]
            self.number = int(cleaned_title[0])
        else:
            cleaned_title = title.strip().split(" ")
            self.title = " ".join(cleaned_title[3:])
            self.letter = cleaned_title[1]

    def __repr__(self):
        if self.is_annex:
            return "({}, Annex {}, {}-{})".format(self.title,self.letter, self.from_id, self.to_id)
        return "({}, {}, {}-{})".format(self.title, self.number, self.from_id, self.to_id)
def match_definition_incipit(txt):
    '''
    Returns tru if txt matches the incipit of a definition,
    identified by the word 'tosca' 
    '''
    return bool(re.match(r'^tosca\.[a-zA-Z\.:0-9\s]*$',txt.split("\n")[0].strip()))

def is_tosca_def(table):
    '''
    Returns true when a table contains TOSCA definitions, i.e.
    the table contains just one cell and text starts with an
    empty space ' '
    '''
    txt = table.rows[0].cells[0].text
        len(table.rows) == 1 and \
        len(table.columns) == 1 and \
        match_definition_incipit(txt)    
def tosca_model_info(name, version, imports):
    Returns a dictionary to hold information on the model
    '''
    return {
        'name' : name,
        'fn' : BASE_FILENAME.format(version.replace(".","-"), name),
        'fd' : None,
        'imports' : imports,
        'buf' :  StringIO()
    }
    '''
    Returns a list of all paragraphs and tables in the Document
    '''
    parag_count = 0
    table_count = 0
    for element in body._element:
        if isinstance(element, docx.oxml.text.paragraph.CT_P):
            ret.append(Paragraph(element, body))
            parag_count = parag_count + 1
        elif isinstance(element, docx.oxml.table.CT_Tbl):
            ret.append(Table(element, body))
            table_count = table_count + 1
carignani's avatar
carignani committed
        #else:
        #    print("Non paragraph or table " +  str(type(element)))
    print("Paragraphs: " + str(parag_count))
    print("Tables: " + str(table_count))
    return ret

def find_sect(sect_to_find, start_idx, doc_content):
    '''
    Returns the index in the doc_content list to the first paragraph
    or heading of the section with title sect_to_find,
    starting the research from start_idx
    '''
    while start_idx < len(doc_content):
        my_elem = doc_content[start_idx]
carignani's avatar
carignani committed
        if isinstance(my_elem, Paragraph) and my_elem.text.strip() == sect_to_find:
    print("FOUND " + sect_to_find + " at " + str(start_idx))
carignani's avatar
carignani committed
def is_lvl2_section_hdn(txt):
    ''' Returns true if txt is level 2 heading'''
    clean_txt = txt.strip()
    
    if not bool(re.match(r'^[0-9]\.[0-9]+\t[a-zA-Z\s]*$', clean_txt)):
        return False

    subtitle = clean_txt.split('\t')[1]
    
    return subtitle in SUBSECTIONS

def is_lvl1_section_hdn(txt):
    ''' Returns true if txt is level 1 heading'''
    clean_txt = txt.strip()
    return bool(re.match(r'^[0-9]+\t[a-zA-Z\s]*$', clean_txt)) or \
            bool(re.match(r'^Annex[\s]*[A-Z]+[\s\t]+[a-zA-Z\s\(\)]*', clean_txt))

def find_all_sections(doc_content):
    '''
    Scans the body of the document to find level 1 sections
    Returns a list of Section
    '''
    sections = []

    start_indx = 0
    end_indx = 1

    while end_indx < len(doc_content):
        my_elem = doc_content[end_indx]
        if isinstance(my_elem, Paragraph) and is_lvl1_section_hdn(my_elem.text):
            if start_indx != 0:
                sections.append(Section(start_indx, end_indx-1, doc_content[start_indx].text))
            
            start_indx = end_indx
            
        end_indx = end_indx + 1
    
    sections.append(Section(start_indx, end_indx-1, doc_content[start_indx].text))
    return sections

carignani's avatar
carignani committed
def write_subsection_to_file(txt, buf):
    '''
    Writes a subsection header in utf-8 encoding to file buf
    '''
    buf.write(slugify(txt)+":")
    if not txt.endswith('\n'):
        buf.write('\n')
    buf.write('\n')

def write_table_to_file(tab, buf):
    '''
    Writes content of table t in utf-8 encoding to file F
    '''
    txt = tab.rows[0].cells[0].text
    if not txt.endswith('\n'):
        buf.write('\n')
    buf.write('\n')
def gen_tables_btwn(a_id, b_id, content, buf):
    '''
    Loops over content and writes all tosca definitions to the
    fdesc file. Returns the number of written definitions
    '''
    definitions_count = 0
carignani's avatar
carignani committed

    for idx in range(a_id, b_id):
carignani's avatar
carignani committed
        if idx >= len(content):
carignani's avatar
carignani committed
            print("ERROR: Paragraph out of range (A: {}, B: {}, IDX: {}, LEN(CONTENT): {})".format(
                a_id, b_id,idx, len(content)))
            return definitions_count

        tmp_elem = content[idx]
carignani's avatar
carignani committed

        if isinstance(tmp_elem, Paragraph) and is_lvl2_section_hdn(tmp_elem.text):
            print(tmp_elem.text)
            write_subsection_to_file(tmp_elem.text.split("\t")[1], buf)
            definitions_count = definitions_count + 1

        if isinstance(tmp_elem, Table) and is_tosca_def(tmp_elem):
            write_table_to_file(tmp_elem, buf)
            definitions_count = definitions_count + 1
        elif isinstance(tmp_elem, Table):
            txt = tmp_elem.rows[0].cells[0].text
            if txt.strip().startswith("Name") or txt.strip().startswith("Shorthand") or \
                txt.strip().startswith("tosca_def"):
                continue
    return definitions_count

def generate_header(
        model_name, 
        buf, 
        spec_version=SPEC_VERSION, 
        imports=None, 
        tosca_version=DEFAULT_TOSCA_VERSION):
    '''
    Writes the header to the file for a specific model
    '''
    buf.write(HDR.format(
        tosca_version=tosca_version,
        model=model_name,
        spec_version=spec_version,
        imports=imports))

def generate_templates(
        filename, 
        spec_ver=SPEC_VERSION, 
        yaml_root='uri', 
        tosc_ver=DEFAULT_TOSCA_VERSION):
carignani's avatar
carignani committed
    '''
carignani's avatar
carignani committed
    Takes a filename or file object and loads the definition into the MODELS dictionary
carignani's avatar
carignani committed
    '''
    if isinstance(filename, str):
        print("Opening " + filename)
carignani's avatar
carignani committed
        import_stmt = 'etsi_nfv_sol001_common_types.yaml'
        if yaml_root != 'local':
            import_stmt = \
                'https://forge.etsi.org/rep/nfv/SOL001/raw/{}/'.format(spec_ver) + import_stmt
        MODELS[mod] = tosca_model_info(
            mod,
            spec_ver, 
carignani's avatar
carignani committed
            '- ' + import_stmt
    try:
        sol_001 = docx.Document(filename)
    except:
        print("Error opening the submitted Docx file")
        raise ValueError("Cannot open the submitted Docx file")
        generate_header(
            MODELS[mod]['name'], 
            MODELS[mod]['buf'],
            spec_ver, 
            MODELS[mod]['imports'],
            tosc_ver
    content = get_content(sol_001)
    sections = find_all_sections(content)
    sections_to_models = {
        6 : 'vnfd',
        7 : 'nsd',
        8 : 'pnfd',
        9 : 'common'
    }
    for sect in sections:
        if not sect.is_annex:
            if sect.number in sections_to_models.keys():
                model = sections_to_models[sect.number]
                count = gen_tables_btwn(sect.from_id, sect.to_id, content, MODELS[model]['buf'])
                print("Printed " + str(count) + " types to " + model)
        else:
            if sect.letter == "A":
                count = generate_examples_between(sect.from_id, sect.to_id, content, EXAMPLES)
                print("Printed " + str(count) + " types to " + "Annex " + sect.letter)
carignani's avatar
carignani committed

carignani's avatar
carignani committed

def print_to_files(prefix=None):
carignani's avatar
carignani committed
    ''' 
carignani's avatar
carignani committed
    Prefix is a path to a folder to work into
    '''
    for key in MODELS:
        mod = MODELS[key]
        if prefix is not None:
            mod['fn'] = os.path.join(prefix, mod['fn'])
        print("Writing to " + mod['fn'])
        mod['fd'] = open(mod['fn'], 'w')
        mod['buf'].seek(0)
        mod['fd'].write(mod['buf'].read())
        mod['fd'].write('\n')
        mod['fd'].close()
    for k in EXAMPLES:
        if prefix is not None:
            fnm = os.path.join(prefix, "example_"+EXAMPLES[k].filename)
            fnm = EXAMPLES[k].filename
        print("Writing example file: " + fnm)
        with open(fnm, 'w') as newf:
            newf.write(EXAMPLES[k].text)
            newf.write("\n")
            newf.close()

def parse_version_from_filename(filename):
    '''
    Parses the version from the filename
    '''
    base_filename = os.path.basename(filename)

    if base_filename.startswith("gs_NFV-SOL001v"):         
        return "v" + base_filename.strip("gs_NFV-SOL001v") \
                .replace("0",".").strip(".").strip("p.docx")
    if base_filename.startswith("gs_nfv-sol001v"):         
        return "v" + base_filename.strip("gs_nfv-sol001v") \
                .replace("0",".").strip(".").strip("p.docx")
    return ""

carignani's avatar
carignani committed
def slugify(t):
    return t.replace(" ", "_").lower()

carignani's avatar
carignani committed
if __name__ == "__main__":

    try:
        SOL001_FN = sys.argv[1]
    except:
        print('Error: Filename missing or filename not a docx document')
        print('Usage: doc2tosca <docx-with-tosca-definitions>')
        sys.exit(1)

    ver = parse_version_from_filename(SOL001_FN)
    generate_templates(SOL001_FN, spec_ver=ver)
carignani's avatar
carignani committed

    print_to_files()