Transmogrifier source for reading files from the filesystem
Transmogrifier source for reading files from the filesystem
This package provides a Transmogrifier data source for reading files, images and directories from the filesystem. The output format is geared towards constructing Plone File, Image or Folder content. It is also possible to add arbitrary metadata (such as titles and descriptions) to the content items, by providing these in a separate CSV file.
First, add transmogrify.filesystem as a dependency of your product in your setup.py, and include its configuration in configure.zcml:
<include package="transmogrify.filesystem" />
Alternatively you can use transmogrify.filesystem with mr.migrator by adding transmogrifier.filesystem to the list of eggs in the mr.migrator section of your buildout.cfg. See: http://pypi.python.org/pypi/mr.migrator/1.0b4 for more information.
This is an experimental alternative approach to executing Transmogrifier pipelines, aimed at lowering the bar of entry to Transmogrifier migrations.
You can use transmogrify.filesystem in a pipeline like this:
[transmogrifier] pipeline = data constructor schema savepoint [data] blueprint = transmogrify.filesystem directory = my.package:data/root ignored = re:.*\.svn.* [constructor] blueprint = collective.transmogrifier.sections.constructor [schema] blueprint = plone.app.transmogrifier.atschemaupdater [savepoint] blueprint = collective.transmogrifier.sections.savepoint every = 50
This will scan the directory data inside the Python package my.package for files and directories, but will ignore any .svn directories and their contents. The source will yield items suitable for passing to the standard Transmogrifier content constructor and the schema updated from plone.app.transmogrifier.
Transmogrifier will re-create the directory structure and files in Plone, but there is no information here to set titles or descriptions for created items.
You can, however, add additional metadata to files by using a CSV file. For example, say you had a CSV file like this:
path,title,description /foo,Foo,A file /foo/bar,Bar,Another file
The headers in this file indicate the field names for additional data. One of the columns must have a heading of path, and should contain file paths relative to the Transmogrifier context (normally the Plone site root), with a leading slash, as shown here. These paths will be matched against the files loaded from the data directory.
With portal_type column you can pick a specific content type. When the portal type is “News Item” it checks whether the file is an image or not, and acts accordingly. When the portal type is Document or Event the content of the file will be written in the “text” field.
Subsequent columns are passed along as-is, so in this case, the title and description fields will be set as given in the CSV file.
To use this file, use the metadata key, e.g.:
[data] blueprint = transmogrify.filesystem directory = my.package:data/root metadata = my.package:data/metadata.csv ignored = re:.*\.svn.*
The available options are:
- directory (required)
- The directory from which files are read. Subdirectories should reflect the eventual path of images and files uploaded. May be given as an absolute path, a path relative to the current working directory, or a package reference (e.g. my.package:foo/bar).
- A CSV file containing metadata. See above. May be given as an absolute path, a path relative to the current working directory, or a package reference (e.g. my.package:foo/bar).
- If set to True, only files with matching metadata are included. Directories are always included regardless. Defaults to False.
- A character that delimits data in your Metadata CSV file.
- If this is set to True, transmogrify.filesystem will break if it finds a row of data in Metadata CSV which field-count does not match field-count of the first row in the CSV file.
- The portal type for folders to create (if required). Defaults to ‘Folder’.
- The portal type for files. Defaults to ‘File’.
- The name of the field to use for file (non-image) content. Defaults to file, which is the name for a standard ATFile.
- The portal type for images. Defaults to ‘Image’.
- The name of the field to use for image content. Defaults to image, which is the name for a standard ATImage.
- By default, file data will be wrapped into an OFS.File or OFS.Image object. Set this option to False to get the raw data, as a string.
- The default file type for content where the mimetype cannot be guessed. Defaults to application/octet-stream.
- A list of paths and/or regular expressions (prefixed with re: or regexp: to skip).
The yielded dictionaries will have the following keys:
- Portal type name. Will be one of folder-type (Folder), file-type (File) or image-type (Image).
- The ZODB path for the new item. This is based on the folder structure from which files were read.
For images and files, two additional keys are included:
- The mimetype, as guessed from the file extension. The default, if no adequate guess can be made, is application/octet-stream.
- Image field name (as set with file-field or image-field)
- The contents of the file.
In addition, any keys from matching rows in the metadata CSV file, if specified, will be included. The values will all be strings.
- Fix brown bag release. [zupo]
- Sort subfolders and filenames after you read it from the filesystem; this avoids different order of subdirs/files on different OSes. [zupo]
- When the portal type is “News Item” it checks whether the file is an image or not, and acts accordingly [sithmel]
- Add autoinclude entry targetting transmogrify to provide mr.migrator support. This means folks can list this package in their mr.migrator sections (in buildout.cfg) and mr.migrator will load its ZCML. [aclark]
- Added option to specify item’s portal_type inside metadata.csv. [zupo]
- Added option ‘strict’ that will make transmogrify.filesystem break if it finds a row of data in Metadata CSV which field-count does not match field-count of the first row in the CSV file. [zupo]
- Added option ‘delimiter’ so you can choose a character that delimits your Metadata CSV file. [zupo]
- Fixed test_empty_directory by ignoring .svn and similar directories. [zupo]
- Reverted toutpt’s change since resolvePackageReferenceOrFile is now available in a released version of transmogrifier. [zupo]
- update api resolvePackageReferenceOrFile -> resolvePackageReference [toutpt]
- Initial release [optilude]