bdrc-volume-manifest-builder

New in Release 1.3

Release	Comment
1.3.2	Adds a missing integration
1.3.1	Corrects the setup.py
1.3.0	adds the ability to specify named image groups in the command line. It applies to fs and ``s3``modes.

ex: manifestforwork -w W23834 --image-group 3187,I123456 fs

Notes:

• the -i/–image-group argument is a comma-separated list of image groups (or one item) If it is not given, all the image groups in the work’s BUDA catalog will be processed.
• the –image-group flag cannot be given with the –work-list-file argument.
• The –image-group arguments do apply when the –work-rid argument is a file path.

New in Release 1.1

• Ability to use either file system or S3 for image repository

Intent

This project originated as a script to extract image dimensions from a work, and:

• write the dimensions to a json file
• report on images which broke certain rules.

Implementation

Archival Operations determined that this would be most useful to BUDA to implement as a service which could be injected into the current sync process. To do this, the system needed to:

• be more modular
• be distributable onto an instance which could be cloned in AWS.

This branch expands the original tool by:

• Adding the ability to use the eXist db as a source for the image dimensions.
• Use a pre-built BOM Bill of Materials) to derive the files which should be included in the dimesnsions file
• Read input from either S3 or local file system repositories
• Create and save log files.
• Manage input files.
• Run as a service on a Linux platform

Standalone tool

Internal tool to create json manifests of image format data for volumes present in S3 to support the BUDA IIIF presentation server.

Language

Python 3.7 or newer. It is highly recommended to use pip to install, to manage dependencies. If you must do it yourself, you can refer to setup.py for the dependency list.

Environment

1. Write access to /var/log/VolumeManifestBuilder which must exist.
2. systemctl service management, if you want to use the existing materials to install as a service.

Usage

Command line usage

The command line mode allows running one batch or one work at a time. Arguments specify the parameters, options.

You also must choose a repository mode which determines if the images are on a local file system (the fs mode), or on an AWS S3 system (the s3) mode.

Common parameters

This section describes the parameters which are independent of the repository mode.

  $ manifestforwork -h usage: manifestforwork [common
 options] { fs [fs options] \| s3 [s3 options]}

 Prepares an inventory of image dimensions

 optional arguments: -h, –help show this help message and exit -d
 {info,warning,error,debug,critical}, –debugLevel
 {info,warning,error,debug,critical} choice values are from python
 logging module -l LOG_PARENT, –logDir LOG_PARENT Path to log file
 directory -f WORK_LIST_FILE, –workListFile WORK_LIST_FILE File
 containing one RID per line. -w WORK_RID, –work-Rid WORK_RID name or
 partially qualified path to one work

 Repository Parser: Handles repository alternatives

 {s3,fs}

Common usage Notes:

`-f/--workListFile` is a file which contains a list of RIDS, **or a list of paths
to work RIDs, in the `fs` mode (see below.)**
`-w/--workRID` is a single work.

- The `--workListFile` and `--workRid` arguments are mutually exclusive

- The system logs its activity into a file named _yyyy-MM-DD_HH_MM_PID_.local_v_m_b.log`
  in the folder given in the `-l/--logDir` argument (default `/var/log`)
  mode.

  Before release 1.3.0, `manifestforwork` used an externally generated list of files (fileList.json) in the source directory to specify the population to process.
  After that, the entire directory is scanned (this was needed to be able to process arbitrary image groups.), and the file list is disregarded.

  The use of a file list was in response to many badly formed entries in `dimensions.json` due to random files being scanned. In Release 1.3.0, these files
  are now explicitly tagged in the `dimensions.json`

{
  "filename":"SomeFile.ext",
  "error":"UnidentifiedImageError"
}

fs Mode Usage

 ❯ manifestforwork fs -h usage: manifestforwork [common
 options] { fs [fs options] \| s3 [s3 options]} fs [-h] [-c CONTAINER]
 [-i IMAGE_FOLDER_NAME]

 optional arguments: -h, –help show this help message and exit -c
 CONTAINER, –container CONTAINER container for all work_rid archives.
 Prefixes entries in –source_rid or –workList -i IMAGE_FOLDER_NAME,
 –image-folder-name IMAGE_FOLDER_NAME name of parent folder of image
 files

Notes:

+ the `-c/--container` defines a path to the RIDS (or the RID subpaths) given.
  It is optional. It prepends its value to the WorkRID paths or individual workRIDs
  in the input file (`-f`) or to the individual work (`-w`)

In the `-w` or `-f` options above. The system supports user expansion
(`~[uid]/path...` in Linux) and environment variable expansion in both the `-c`
and the `-f` options. That is, the file given in the `-f` option can contain

- Environment variables
- User alias pathnames (`~[user]/...`)
- Fully qualified pathnames

e.g.

> pwd
/data
>ls
Works
>ls ~/tmp
/home/me/tmp/Works
> export THISWORK="Works/FromThom"
> cat workList
$WORKS/W12345
~/tmp/$WORKS/W12345
/home/me/tmp/Works/W89012

using this list in

manifestforwork -f worklist fs

will process files from

• /data/Works/FromThom
• /home/me/tmp/Works/FromThom
• /home/me/tmp/Works/W89012 if the --container argument is not given. (-c defaults to the current working directory)

s3 mode usage

 ❯ manifestforwork s3 –help usage: manifestforwork
 [common options] { fs [fs options] \| s3 [s3 options]} s3 [-h] [-b
 BUCKET]

 optional arguments: -h, –help show this help message and exit -b BUCKET,
 –bucket BUCKET Bucket - source and destination

The S3 mode uses a bucket named with the optional `-b/--bucket` argument. The default bucket
is closely held. note that the `--container` argument is not applicable in this mode, and
that if a worklist is given, it must contain only RIDs, not paths.

The bucket example takes the aws s3api form, e.g. `--bucket somewhere.over.the.rainbow`

Installation

PIP

PyPI contains bdrc-volume-manifest-builder

Global installation

Install is simply sudo python3 -m pip install --upgrade bdrc-volume-manifest-builder to install system-wide (which is needed to run as a service)

Local installation

To install and run locally, python3 -m pip install --upgrade bdrc-volume-manifest-builder will do. Best to do this in a virtual python environment, see [venv](https://docs.python.org/3/library/venv.html)

When you install volume-manifest-builder three entry points are defined in /usr/local/bin (or your local environment):

• manifestforlist the command mode, which operates on a list of RIDs
• manifestforwork alternate command line mode, which works on one path

## Service

See [Service Readme](service/README.md) for details on installing manifestFromS3 as a service on systemctl supporting platforms.

Development

volume-manifest-builder is hosted on [BUDA Github volume-manifest-builder](https://github.com/buda-base/volume-manifest-builder/)

• Credentials: you must have the input credentials for a specific AWS user installed to deposit into the archives on s3.

Usage

volume-manifest-builder has two use cases:

• command line, which allows using a list of workRIDS on a local system
• service, which continually polls a well-known location, s3://manifest.bdrc.org/processing/todo/ for a file.

Building a distribution

Be sure to check PyPI for current release, and update accordingly. Use PEP440 for naming releases.

Prerequisites

pip3 install setuptools
pip3 install wheel
pip3 install twine

Building

python3 setup.py sdist bdist_wheel
twine upload dist/<thing you built>

Project changelog

Release	Commit	Changes
1.3.0	30a3b2c3	use specific image groups
	82adb9f	Use only image files in search
1.2.10		Clean up S3 error message
1.2.9		Error diags in generateManifest
1.2.8		Update changelog to readme
1.2.7		Use bdrc-util logging
1.2.6		Use BUDA only for resolution
		Use BUDA first for resolution
1.2.0		Sort all output by filename