Image Cache

[amywieliczka]

Addresses #1109

Stores thumbnail and media components on s3 at a configured CONTENT_COMPONENT_CACHE set to s3://<bucket>/<prefix>/ at the key: <collection_id>/<request.url>/<component_type: media|thumbnail>/<ETag>/<Last-Modified>

Where request.url, etag, and last-modified have all been run through urllib.parse.quote_plus to encode them.

Thumbnail and Media Components are defined as:

thumbnail_component = {
    'mimetype': 'image/jpeg',
    'path': content_s3_filepath,                # s3 filepath to generated thumbnail
    'dimensions': dimensions,                   # dimensions of generated thumbnail
    'component_content_harvest_metadata': {
        'md5': source_component['md5'],                                 # md5 of response content from mapped thumbnail source url
        'src_content-type': source_component['Content-Type'],           # content type response header from mapped thumbnail source url,
        'mapped_mimetype': mapped_thumbnail_source.get('mimetype'),     # mapped thumbnail source mimetype
        'src_size': source_component['size'],                           # size in bytes of response content from mapped thumbnail source url
        'date_content_component_created': datetime.now().isoformat()
    }
}

media_component = {
    'mimetype': mimetype,                                           # mimetype of generated media
    'path': content_s3_filepath,                                    # s3 path to generated media
    'format': NUXEO_MEDIA_TYPE_MAP.get(mapped_nuxeotype or ''),     # format, used by UI to determine display
    'component_content_harvest_metadata': {
        'md5': source_component['md5'],                             # md5 hash of source file
        'src_content-type': source_component['Content-Type'],       # content type response header from source file
        'mapped_mimetype': mapped_mimetype,                         # mimetype from mapper
        'src_size': source_component['size'],                       # size in bytes of source file
        'date_content_component_created': datetime.now().isoformat()
    }
}

[bibliotechy]

s3 implementation looks good. One small non-blocker comment

[barbarahui]

Neat!!