settings-example

###
# Binary paths
###

# What is the path to the ghostscript binary?  Usually you can just leave this
# set to "gs", but may need to be tweaked on some systems.
GHOSTSCRIPT="gs"

# What is the path to the opj binaries?  This may be opj_compress /
# opj_decompress, opj2_*, or a full path, depending on a variety of factors....
OPJ_COMPRESS="opj2_compress"
OPJ_DECOMPRESS="opj2_decompress"

###
# Web configuration
###

# Full URL to the NCA web app; though the hostname isn't currently used,
# it may be necessary for things like sending email notifications in the future
WEBROOT="https://internal.somewhere.edu/nca"

# The bind address is what the server listens on, often just a port string
# prefixed with a colon.  Apache would use this for reverse-proxying.
BIND_ADDRESS=":8080"

# Full URL to the IIIF server's base path - this is used to display issues'
# pages during metadata entry and review
IIIF_BASE_URL="https://my.server.com/iiif"

# Full URL to the live news site, for pulling information about issues which
# are in production, and may be used in the future to link from NCA to
# live batches' / issues' detail pages
NEWS_WEBROOT="https://news.somewhere.edu"

# Full URL to the staging news site for directing admins to batches which are
# ready for review. In theory this could point toward production, allowing a
# more loose "test it live" approach.
STAGING_NEWS_WEBROOT="https://news-staging.somewhere.edu"

# Locations (usually a URL) to pull marc records when a new newspaper title is
# added.  If a location begins with http/https, an HTTP request is made,
# otherwise it is treated as a path to a file.  The string "{{lccn}}" is
# replaced with the LCCN being looked up.  The second location is only
# considered if the first has no XML.
MARC_LOCATION_1="/var/local/marc/{{lccn}}/marc.xml"
MARC_LOCATION_2="https://chroniclingamerica.loc.gov/lccn/{{lccn}}/marc.xml"

###
# Database settings
###

DB_HOST="127.0.0.1"
DB_PORT=3306
DB_USER="nca"
DB_PASSWORD="nca"
DB_DATABASE="nca"

###
# SFTPGo settings.  These determine how NCA interacts with SFTPGo, not how
# SFTPGo behaves more generally; see
# https://github.com/drakkan/sftpgo/blob/main/docs/full-configuration.md for
# configuring the SFTPGo daemon.
#
# This section is optional.  If SFTP_API_URL is set to either blank or "-",
# connections to SFTPGo will not be initiated.  This will however mean that you
# have to manage publisher uploads yourself, and NCA will not have any
# integrated upload support.
###

# SFTPGo's API URL as NCA will use it; Set this to "-" to skip SFTPGo
# integration.  This must be the "internal" URL.  e.g., the default value below
# ("http://sftpgo:8080/api/v2") is how NCA sees SFTPGo in Docker.  This won't
# be the same URL you'd use in a browser for web-based administration.
SFTPGO_API_URL="http://sftpgo:8080/api/v2"

# The SFTPGo API key for API requests
# Reset to 'sftpgo_admin_api_key' surrounded by exclamation points
# and run sftpgo/get_admin_api_key.sh to get a new API key
SFTPGO_ADMIN_API_KEY=!sftpgo_admin_api_key!

# How much disk each new user is allotted for uploads.  Note that this is NOT
# retroactive.  This is simply the default used when a *new* title is sent to
# SFTPGo.  Changing quotas for existing titles requires manual editing of their
# title data in NCA.
SFTPGO_NEW_USER_QUOTA=5gb

###
# ONI Agent settings. These tell NCA what to connect to in order to start and
# monitor batch loading and purging tasks on staging and production servers.
###

STAGING_AGENT="oni-agent-staging:22"
PRODUCTION_AGENT="oni-agent-prod:22"

###
# Paths for PDFs, derivatives, etc.
###

# Uploaded PDFs are deposited here - these should be exactly what the publisher
# gives us.  There must be one subdirectory per title, named after the title's
# "SFTP Directory" value in the admin app, and each title must contain one
# subdirectory per issue, named by the issue's date, in YYYY-MM-DD format.
PDF_UPLOAD_PATH="/mnt/news/sftp"

# In-house scans are deposited here.  The folder structure must be precisely as follows:
#
#     <scan upload root>/<MARC org code>/<lccn>/<issue date with optional edition>
#
# For example, "/mnt/news/scans/oru/sn12345678/2018-01-01_02"
#
# The TIFF files should be at least 300dpi, and the PDFs should contain a
# 150dpi JPEG image encoded at about a quality of 40 (or "medium"), per the
# NDNP spec.  The PDF also needs to have the OCR text embedded for the
# derivative processing to create the proper page XML.
SCAN_UPLOAD_PATH="/mnt/news/scans"

# When an uploaded PDF has been split, this is where we back it up.  Originals
# are stored in the batches, so if you back up the batch, this can be a short-
# term backup location.
ORIGINAL_PDF_BACKUP_PATH="/mnt/news/backup/originals"

# This is where split PDFs are moved for manual reordering / renaming (e.g.,
# using Adobe Bridge)
PDF_PAGE_REVIEW_PATH="/mnt/news/page-review"

# Once processing happens, full batches are put here. They will be moved to the
# archive path upon QC approval and successful push to production.
BATCH_OUTPUT_PATH="/mnt/news/outgoing"

# When a batch is built, this location is where the ONI-required files live
# (basically everything but TIFFs and born-digital original file backups).
# Basically the final home for the always-online files. Both your production
# and staging Open ONI servers must be able to read files from this location.
BATCH_PRODUCTION_PATH="/mnt/news/production-batches"

# When a batch is live, its source files, including TIFF and backups, move to
# an archival location. This can be a dark archive, a "transfer" location for
# prepping bulk DA moves, or a location you manually manage in some way.
BATCH_ARCHIVE_PATH="/mnt/news/batch-archive"

# This is where scanned issues and SFTPed issues go after all manual processing
# is done.  Issues moved here shouldn't be accessible to anybody for manual
# modification, and all metadata will live in the database.
WORKFLOW_PATH="/mnt/news/workflow"

# This is where you want issues to be dropped off when they're reported as
# unfixable and then an "Issue Manager" user chooses to remove them permanently
# from NCA.  This location should be someplace curators / reviewers can access
# so rescanning or other manual fixes can take place.
ERRORED_ISSUES_PATH="/mnt/news/errors"

# This is where cached files are stored for the issue data.  This is extremely
# important for issues which live on the live website, as those are very
# expensive to pull each time a given process is run.
ISSUE_CACHE_PATH="/var/local/news/nca/cache"

# This is where the core application lives, and is important for finding the
# HTML template files as well as static files like JS and CSS
APP_ROOT="/usr/local/nca"

# Where is the template for building issue XML?  You can point this to the
# XML in the repo (templates/xml/mets.go.html) for ease, but some XML values
# may be wrong.
METS_XML_TEMPLATE_PATH="/usr/local/nca/templates/xml/mets.go.html"

# Where is the template for building the batch XML?  This should be safe to use
# as-is, but it could be changed if necessary.
BATCH_XML_TEMPLATE_PATH="/usr/local/nca/templates/xml/batch.go.html"

###
# General rules for issue processing and batch creation
###

# How many pages must an issue have?  Setting this to 2 can avoid processing
# obviously unfinished uploads.
MINIMUM_ISSUE_PAGES=2

# MARC organization code you want to use in chronam to represent batches from
# PDF sources.  e.g., "oru" results in batch names like
# "batch_oru_20150101120000", attributed to the UO Knight Library.  It is best
# to choose a code that chronam already knows about.
PDF_BATCH_MARC_ORG_CODE="oru"

# How many PDFs do we allow in a single batch?  This limit is a lot less
# necessary now that fixing a batch is a simpler process.  Historically, it was
# a lot easier to deal with smaller batches when they had to be rebuilt each
# time an issue had an error.
#
# NOTE: this limit will *not* split an issue.  i.e., this is the maximum number
# of PDFs, not necessarily the precise number a batch will contain.
MAX_BATCH_SIZE=10000

# What is the minimum size of a batch?  A batch won't be queued up until this number of pages is reached.
#
# NOTE: this limit will be ignored if issues have been ready to go live for too
# long in order to avoid issues being "stranded"
MIN_BATCH_SIZE=5000

# How long we require an issue to be untouched prior to anybody queueing it. If
# set to zero, people can queue an issue *immediately* after upload. This is
# not recommended if uploads aren't tightly restricted, but can make sense if
# NCA curators are also the ones doing the uploading.
#
# Duration format follows the Go ["ParseDuration" function][1] To summarize
# what you'd likely need, though: "1h" is one hour, "24h" would be one day,
# "30m" would be 30 minutes, "1s" would be one second, etc. The default is 48h,
# or two days.
#
# [1]: <https://pkg.go.dev/time#ParseDuration>
DURATION_ISSUE_CONSIDERED_DANGEROUS=48h

# How long we warn users that the issue is new - but it can be queued before
# that warning goes away so long as DAYS_ISSUE_CONSIDERED_DANGEROUS has elapsed
#
# Duration format follows the Go "ParseDuration" function - see above. The
# default of 336h is two weeks.
DURATION_ISSUE_CONSIDERED_NEW=336h

###
# Derivative settings
###

# DPI for ghostscript to use on the PDF-to-PNG conversion
DPI=200

# JP2 quality value for graphicsmagick
QUALITY=62.5

# Set this to the value your scanned images use to embed JPGs.  This should be
# 150 per the NDNP spec, but could be changed if scanned images aren't under
# your control.
SCANNED_PDF_DPI=150