dataiku_bundle_modify.py

import dataiku
from dataiku import pandasutils as pdu
import pandas as pd
import shutil
import zipfile


def extract_bundle(zip_filepath, save_location):
    if not os.path.isdir(save_location):
        # path to zipped bundle
        zfile = zipfile.ZipFile(zip_filepath)

        # extract bundle to new location
        zfile.extractall(save_location)

        
def build_bundle(new_directory):
    os.chdir(new_directory)
    file_name = os.path.basename(new_directory)
    os.system('zip -r ../'+ file_name + '.zip *')

    
def get_new_directory_name(new_name, original_name, save_location):
    # replace original name with new name in directory name
    _replace_re = re.compile(original_name)
    return _replace_re.sub(new_name, save_location)

def remove_directory(directory):
    shutil.rmtree(directory)

def replace_project_name(new_name, original_name, save_location):
    # sets up new directory name
    new_directory = get_new_directory_name(new_name, original_name, save_location)
    
    # copies data to new directory
    if os.path.isdir(new_directory):
        remove_directory(new_directory)
        
    shutil.copytree(save_location, new_directory)
    
    # in new directory, replace all occurences of original name with new name
    _replace_re = re.compile(original_name)
    for dirpath, dirnames, filenames in os.walk(new_directory):
        for file in filenames:
            file = os.path.join(dirpath, file)
            tempfile = file + ".temp"
            with open(tempfile, "w") as target:
                with open(file) as source:
                    for line in source:
                        line = _replace_re.sub(new_name, line)
                        target.write(line)
            os.rename(tempfile, file)
    return new_directory

def renamed_bundle(new_name, original_name, zip_filepath, save_location):
    extract_bundle(zip_filepath, save_location)
    new_directory = replace_project_name(new_name, original_name, save_location)
    build_bundle(new_directory)
    remove_directory(new_directory)
    return new_directory + ".zip"

def project_exists(project_key):
    client = dataiku.api_client()
    return project_key in client.list_project_keys()


def create_or_update_project(project_name, bundle_path):
    print bundle_path
    if project_exists(project_name):
        client = dataiku.api_client()
        project = client.get_project(project_name)
        project.import_bundle_from_archive(bundle_path)
    else:
        client = dataiku.api_client()
        client.create_project_from_bundle_local_archive(bundle_path)
        project = client.get_project(project_name)       
    return project