web2py : downloading zip file created by cStringIO text strings in controller

问题

I have a web2py based portal where a group of annotators provide text labels and bounding-boxes related information for the various images shown to them. I'd like to get this information in the form of XML files (one-per-image) and add a download-annotations functionality to the portal which serves a ZIP file containing all these XML files. I am able to create a zip-download from the portal but unzipping the same throws the following error :

mohit@nightfury13:~/Downloads$ unzip Arabic\ Set1-Arabic_annotations.zip    
Archive:  Arabic Set1-Arabic_annotations.zip
      End-of-central-directory signature not found.  Either this file is not
      a zipfile, or it constitutes one disk of a multi-part archive.  In the
      latter case the central directory and zipfile comment will be found on
      the last disk(s) of this archive.
    unzip:  cannot find zipfile directory in one of Arabic Set1-Arabic_annotations.zip or
            Arabic Set1-Arabic_annotations.zip.zip, and cannot find Arabic Set1-Arabic_annotations.zip.ZIP, period.

Following is the code I've written to perform this task. Can someone point out what am I doing wrong? - First half of the code is preparation of the dataset-XML-string (you may skip that). Second half is the zip pipeline.

@auth.requires_login()
def prepare_dataset():
    import os
    from PIL import Image
    import zipfile, cStringIO

    # Check if a valid data-id was passed.
    if not request.vars.data_id:
        session.flash = 'No dataset selected for download'
        redirect(URL('default', 'select_db?redirect=1'))

    # Create the annotation-data in a proper format.
    else:
        data_id = int(request.vars.data_id)
        dataset = db(db.Datasets.id==data_id).select()[0]
        root_path = dataset['data_path'].split('cropped')[0]
        root_images = [i for i in os.listdir(root_path) if i.endswith('.jpg') or i.endswith('.jpeg') or i.endswith('.png')]
        content = {}
        imgs_data = db((db.Images.data_id==data_id)&(db.FinalLabels.img_id==db.Images.id)).select()
        for img_data in imgs_data:
            label = img_data['FinalLabels']['label']
            if 'bad' not in [i.lower() for i in label.split()]:
                img_name = img_data['Images']['img_name']
                root_img_name = img_name.split('_')[0]
                xmin, ymin, xmax, ymax = img_name.split('.')[0].split('_')[2:]

                if root_img_name not in content:
                    r_im_name = [i_name for i_name in root_images if root_img_name in i_name][0] # This one also has the extension
                    root_im = Image.open(os.path.join(root_path, r_im_name))
                    r_depth = 3
                    if not root_im.mode=='RGB':
                        r_depth = 1
                    r_width, r_height = root_im.size

                    content[root_img_name] = {'name':r_im_name, 'depth':r_depth, 'width':r_width, 'height':r_height, 'crops':[{'label':label, 'xmin':xmin, 'ymin':ymin, 'xmax':xmax, 'ymax':ymax}]}
                else:
                    content[root_img_name]['crops'].append({'label':label, 'xmin':xmin, 'ymin':ymin, 'xmax':xmax, 'ymax':ymax})

        # Compress img-annotation data (content) to zip and export
        zip_chunks = cStringIO.StringIO()
        zipf = zipfile.ZipFile(zip_chunks, "w", compression=zipfile.ZIP_DEFLATED)

        for im_name in content:
            root_im = content[im_name]
            root_folder = filter(None, root_path.split('/'))[-1]
            xml_str = """<annotation>
    <folder>%s</folder>
    <filename>%s</filename>
    <path>%s</path>
    <source>
        <database>Unknown</database>
    </source>
    <size>
        <width>%s</width>
        <height>%s</height>
        <depth>%s</depth>
    </size>
    <segmented>0</segmented>""" % (root_folder, im_name, root_im['name'], root_im['width'], root_im['height'], root_im['depth'])

            for crop in root_im['crops']:
                xml_str+="""
    <object>
        <name>%s</name>
        <pose>Unspecified</pose>
        <truncated>0</truncated>
        <difficult>0</difficult>
        <bndbox>
            <xmin>%s</xmin>
            <ymin>%s</ymin>
            <xmax>%s</xmax>
            <ymax>%s</ymax>
        </bndbox>
    </object>""" % (crop['label'], crop['xmin'], crop['ymin'], crop['xmax'], crop['ymax'])


            xml_str+='\n</annotation>'
            zipf.writestr(im_name+'.xml', xml_str)

        zipf.close()
        zip_name = dataset['data_name']+'_annotations.zip'
        file_header = 'attachment; filename='+zip_name
        response.headers['Content-Type'] = 'application/zip'
        response.headers['Content-Disposition'] = file_header
        return zipf

回答1:

ZipFile is writing the data to the zip_chunks StringIO object, so you must return zip_chunks.getvalue(), not zipf.

来源：https://stackoverflow.com/questions/44804747/web2py-downloading-zip-file-created-by-cstringio-text-strings-in-controller