fylr.example.yml
In this section you find most settings possible in fylr.yml and explanations as comments
This is NOT a coherent fylr.yml, do not use for an installation; use this instead.
If you start your hierarchy in fylr.yml with fylr+: instead of fylr:, then defaults are used where not explicilty overwritten. Defaults see fylr.default.yml
fylr.yml
## - Some of the regexs to check for versions might require modifications
## - All relative paths described in the file are relative to cmd/fylr unless otherwise
## stated
## - All ports described in the file are formatted to work with Golang net/http package
##
## Multiple config files can be parsed (use -c ... -c ... -c ...).
## The following rules apply:
## - For maps the whole map is replaced, unless the key which holds the map
## ends ins a plus. If it ends in a minus, values are removed from a map.
## - For slices the whole slice is replace. If the key ends in a +, values are
## are appended, if it ends in a "-" values are removed. If it ends in "<",
## values are prepended.
## - Other values are overwritten
##
## To set the externalURL and the sqlite database name in a minimum fylr.yml one could do:
##
## ---fylr.yml---
## fylr+:
## externalURL: "http://localhost"
## db+:
## dsn: mysqlite.db
## --------------
##
## The "+" on the top level and for "db" is needed to tell fylr to merge the map
## rather than overwriting it. If the .yml uses "fylr" instead (no "+") the
## default config will be replaced.
##
## Environemnt allows to overwrite the fylr.yml All variables are read in
## upper-case. For the path to deeper variables, use "_" as separator. Supported
## types: string, []string. To fill []string, use a JSON array.
##
## Example: FYLR_DB_DSN="host=localhost port=5432 user=fylr dbname=test"
## FYLR_DB_DRIVER="postgres"
##
## Use --env-prefix on the command line to overwrite the FYLR_ default
# All config is below "fylr"
fylr:
# Name of the instance (mandatory). It is used as prefix for the bucket
# names and elastic indexes. The name must match regexp [a-z][a-z0-9\-_]*[a-z0-9]:
# All small letters, start with a letter, followed by small letter, a number, dash or
# underscores, ending in letter or number. Must be between 3 and 32 bytes.
name: "myfylr"
# Public external url of the server. This url needs to be fully qualified. If
# you change this url for an existing database, a re-index is needed to update
# all object caches (which store the URL). This re-index can be started from
# /inspect/system.
#
# fylr will redirect all incoming requests to this externalURL, unless there is
# a reverse proxy source matching in fylr.services.webapp.reverseProxy.custom.
externalURL: "http://localhost"
# licenseFile (default: none). Path to license file. This is used as default
# if nothing is set in the baseconfig. This setting is mutually exclusive with
# fylr.license.
licenseFile: "license.json"
# licsense: Inline JSON of license. This is used as default if nothing is set
# in the baseconfig. This setting is mutually exclusive with fylr.licenseFile.
license: "<JSON>"
# encryptionKey is used to AES-encrypt sensitive information before writing it
# to the database. It must be 32 bytes long. Default is empty.
encryptionKey: ""
# stdErrFile redirects all output to stderr to the given file when fylr is run
# as a server. If the file cannot be created for writing, fylr will not start.
# The file path is relative to fylr's working directory.
stdErrFile: ""
# for development only
debug:
# Skip term creation
skipTerms: false
# Don't announce plugins bundle on /api/plugin
noPluginsBundle: false
# Simulate local status
localStatus: "" # purge, reindex, startup, ready
# This setting skips the sometime expensive permission check
# on the /api/eas/download endpoint. However, there is still
# some security left, the link is checked against a valid hash
# of the file which needs to be provided.
easDownloadSkipCheckRights: false
# Allow management of indices in Elasticsearch cluster, such as
# viewing and deleting an index.
# The indices visible there are not limited to those associated
# to the fylr instance.
inspectEnableElasticIndices: false
# Outputs object loader timing and load depth information. This setting can also
# be set in the base config.
logTimings: true
# tempDir is used by api/system/backup and the export to during a TAR
# production. With no tempDir given the system defaults temp dir is used. This
# tempDir gets not cleaned up automatically by FYLR.
tempDir: _tmp_workdir
# Settings for the "logger" (we use https://github.com/rs/zerolog)
logger:
# Set to "json" or "console". Default: "console"
format: "console"
# Set zerolog log level: trace, debug, info, warn, error, fatal, panic
# default to "info".
level: "info"
# timeFormat is the Go representation to format the time in the log output.
# zerolog's time keeping resolution is always set to milliseconds by FYLR.
# example values and their effect:
# "2006-01-02 15:04:05" a template (recommended. 1 to 6 mark the fields)
# see https://pkg.go.dev/time#example-Time.Format for details
# "" results in time only (console format) or seconds since 1970 (json format)
# "UNIXMS" milliseconds since 1970 (json format only)
# "UNIXMICRO" microseconds since 1970 (json format only)
# Defaults to ""
timeFormat: "2006-01-02 15:04:05"
# turn off color for zerolog's underlying ConsoleWriter
# format: "console" only.
noColor: false
# add hostname to log output
addHostname: false
# add remote addr to log output
addRemoteAddr: false
# wrong password log level, set the level for wrong password output
wrongPasswordLevel: debug
# data is stored in a database, FYLR supports sqlite & postgres
db:
# driver: sqlite3
# dsn: "../../_data/sqlite/fylr.db"
driver: postgres
# specify a PostgreSQL user owning the given db (with LOGIN and INHERIT, which are defaults for CREATE USER anyway).
dsn: "host=localhost port=5432 user=fylr password=fylr dbname=fylr sslmode=disable"
# https://golang.org/pkg/database/sql/#DB.SetMaxOpenConns
# default: 0 which is unlimited
# recommended: same as postgreSQL max_connections (which defaults to 100)
# At least: 4 + execserver.parallel + execserver.parallelHigh + elastic.parallel.
# Two of these connections will be dedicated to a separate connection pool
# managing the sequences. (Postgres only)
maxOpenConns: 100
# https://golang.org/pkg/database/sql/#DB.SetMaxIdleConns, default: 0
# This has to be not more than maxOpenConns
maxIdleConns: 10
# The init block is used to pre-fill the database when its created or purged.
init:
# Path to base config file. If set, on a fresh install or after a purge this
# file is used and preloaded into the base config. The format of the file is
# JSON in API format. Defaults to "", which starts with an empty base
# config. The empty base config uses defaults which are defined in
# resources/baseconfig/baseconfig.yml. Unset keys (from a partial base
# config file), are set to their defaults.
configFile: ""
# Inline base config. Works like configFile but used the inlined
# baseconfig setting in this .yml file. Runs after configFile, so everything
# in config overwrites settings from configFile. Default is empty.
config: null
# To setup a purgeable system with default storage locations
# locfile & locs3, configure the following. The location_default map
# accepts location ids as well as names:
#
# config:
# system:
# config:
# purge:
# allow_purge: true
# purge_storage: true
# location_defaults:
# originals: file
# versions: s3
# backups: file
# preconfigure locations for empty databases
locations:
# the location's name can be any string which you choose
#
# Inside the storage location fylr will create a structure like this:
#
# [prefix/]fylr-UUID/originals
# [prefix/]fylr-UUID/versions
# [prefix/]fylr-UUID/backups
#
# The driver "file" will create directories inside the configured
# top level directory but never remove them. So, if you use purge
# a lot, FYLR will leave empty directories on disk.
file:
# The kind is either "file" or "s3" or "azure" (see below)
kind: file
# Each location can configure a prefix which will
# be attached before the file to be created
prefix: "apitest/"
# Set to true if files in this location can be
# purged by FYLR (if the purge api call is used)
allow_purge: true
# Set to true if that storage location allows its download URLs to be
# exposed. For the kind "file" storage backend this is a dangerous
# setting and should be used for development purposes only. Default to
# false.
allow_redirect: false
config:
file:
dir: "_files"
s3:
kind: s3
prefix: "apitest/"
allow_purge: true
config:
s3:
bucket: apitest
endpoint: "127.0.0.1:9000"
accesskey: "minioadmin"
secretkey: "minioadmin"
region: "us-east-1"
ssl: false
azure:
kind: azure
allow_purge: true
config:
azure:
# The name of the container is required. It is created if it doesn't exist.
container: fylr
# You can provide a connection string OR the single settings account_name, account_key and endpoint_suffix
connection_string: "DefaultEndpointsProtocol=https;AccountName=azure;AccountKey=l/fnF4kp...==;EndpointSuffix=core.windows.net"
account_name: "azure"
account_key: "l/fnF4kp..."
# optional endpoint suffix (default to core.windows.net)
endpoint_suffix: "core.windows.net"
# DEPRECATED, will be removed in next version
# files are stored in S3. Buckets are created by FYLR automatically
s3:
endpoint: "127.0.0.1:9000"
accessKeyID: "minio"
secretAccessKey: "minio123"
bucketLocation: "us-east-1"
bucketName: fylr-census-dev
ssl: false
# allowPurge controls if a purge also purges the storage
# or not. Defaults to false
allowPurge: false
plugin:
# load plugins at startup. the loader crawls the given directories
# and loads given files for plugin config files, ending in ".yml".
# "*" and "?" are allowed as placeholders, unmatched directories or
# files are silently skipped.
paths:
- ../../../easydb-plugins
- ../../../fylr-plugins/fylr_example
urls:
- https://github.com/programmfabrik/fylr-plugin-formula-columns/releases/download/v0.1.2/fylr-plugin-formula-columns.zip
# defaults is a map setting defaults for the plugin registration.
# This is configured as map with the plugin name as key.
defaults:
# default for fylr_example plugin
fylr_example:
# enable, set to false to disable the plugin, defaults to true
enabled: false
# update_policy: automatic, always, never, defaults to automatic
update: "never"
# Set to true to allow /api/settings/purge. dont use on production systems!
allowpurge: true
# Optional path to resources, that are all files needed by FLYR during runtime
# (inspect, pages, etc..) If unset, fylr uses binary embedded resources (so no
# resource folder needed in that case). Defaults to embedded.
resources: ""
# Elastic is the indexer for FYLR
elastic:
# Logger used for the elastic client
# "Text": TextLogger prints the log message in plain text.
# "Color": ColorLogger prints the log message in a terminal-optimized plain text.
# "Curl": CurlLogger prints the log message as a runnable curl command.
# "JSON": JSONLogger prints the log message as JSON.
logger: ""
# Where to find the Elastic search index settings. If you provide your own
# file make sure to base it on the default
# resources/index/index_settings.json which is included in the distribution.
settings: ""
# number of parallel workers to index documents, default to 1, set to 0 to disable
parallel: 1
# number of objects per job passed to the indexer process
objectsPerJob: 100
# limit of payloads sent to Elastic
maxMem: 100mb
# fielddata (debug feature). if set to true, fields are mapped including their fielddata
# in the reverse index. with that, the inspect view of the indexed version of
# the object shows a per field list of stored terms. This can be useful for debugging
# of analyzer settings.
fielddata: false
# addresses of the elastic servers
addresses:
- "http://localhost:9200"
# username
username: ""
# password
password: ""
# Client configuration of execserver is used
# for syncing of files, metadata generation and plugin execution
execserver:
# number of parallel file workers, default to 2, set to 0 to disable.
parallel: 2
# number of parallel file workers only taking high priority tasks. Currently
# producing of all standard versions is a high priority task.
parallelHigh: 2
# addresses of the execserver. they are tried in round robin.
# if a server reports to be busy, the next server is tried.
# if the server URL contains a /job/{service} path it is only used for the given service
# example to only match service "node": http://localhost:8083/job/node?pretty=true
addresses:
- http://localhost:8083/?pretty=true
# the maximum a callback is allowed to run
pluginJobTimeoutSec: 2400
# the maximum the server will wait until a worker gets a job
connectTimeoutSec: 120
# callbackBackendInternalURL will be included in execserver jobs, this is
# used for plugin installation (loaded from the backend into the execserver)
# and progress updates.
callbackBackendInternalURL: "http://localhost:8081"
# callbackApiInternalURL will be presented to execserver plugin jobs. This
# can be used by plugins to call back into the API.
callbackApiInternalURL: "http://localhost:8080"
# services which will be started. It is possible to configure a standalone
# execserver or webapp. backend & server can only be configured together.
services:
# service "server" serves the API & oauth2
api:
# address of the api listener (with authentication)
# if omitted, this server is not started
addr: :8080
# for tls support ("addr" only), provide a cert and key file
tls:
certFile: ""
keyFile: ""
# hotfolder path, if set this folder will be created by fylr (if not
# exists).
webDAVHotfolderPath: "_hotfolder"
# oauth2 server settings
oauth2Server:
# This secret is used to sign authorize codes, access and refresh
# tokens. If unset, a random string is used. If multiple fylr server are
# serving the same instance, the signingSecret needs to be configured.
# The format is a 32-byte string.
signingSecret: ""
# allowHttpRedirects will disable the check whether a redirect URL is
# secure. If set to true, this setting http urls are allowed to be used
# for redirects. Default: false.
allowHttpRedirects: false
# The default clients are built into the web app and other apps.
# They are public and thus do neither need nor have a secret.
# see https://docs.fylr.io/for-system-administrators/configuration/fylr.default
clients:
web-client:
# secret must be given as bcrypt hash
secret: $2y$04$81xGNnm8PS1uiIzjbos6Le3NzFaNB0goNqnBpOx7S/EyrayzJCNAq # foo
# this must point to the public URL or the webapp (/oauth2/callback)
redirectURIs:
- http://localhost/oauth2/callback
# the scopes this client needs to be granted
# offline is mandatory to issue refresh tokens
# others will be used in rights management
scopes:
- "read"
- "write"
- "offline"
# public, set to true if this client is publicly known
public: false
# service backend is used for /inspect and by the execserver as callback
# to get installed plugins and to update progress during execution
backend:
# address of the server listener
# if omitted, this server is not started
addr: :8081
# for tls support ("addr" only), provide a cert and key file
tls:
certFile: ""
keyFile: ""
# provide the /inspect URL endpoint
inspect+:
# backup: if configured, /inspect/migration can be used to migrate
# instances.
backup+:
# Path is needed if you want to use /inspect/migration of this
# instance. Disk path to backup. Is created if not present. Each
# backup gets its own subdirectory.
path: /tmp/migration
# oauth2 client + web client
webapp:
# listener for the webapp
addr: ":80"
# For tls support ("addr" only), provide a cert and key file
tls:
# certFile contains the certificate in pem Format and optionally intermediate chain certificates below in the same file.
certFile: "server.crt"
keyFile: "server.key"
# forward requests to fylr.services.webapp.addr (the port of the) webserver.
# Typically you use ":80" for the forward http and ":443" for webapp.addr.
# Default is "" (off)
# With letsEncrypt configured below, this is used for:
#forwardHttpAddr: "" # use port 443, letsencrypt challenge_type=tls-alpn-01
#forwardHttpAddr: ":80" # use port 80, letsencrpyt challenge_type=http-01
forwardHttpAddr: ":80"
# Automatic certificate management can be enabled using the "letsencrypt"
# property. With this setting, fylr automatically retrieves and renews
# an certificate for the domain of fylr.externalURL.
letsEncrypt:
# email is used to register the certificate with Let's Encrypt
email: you@example.com
# useStagingCA sets the staging server of Let's Encrypt which has a higher
# quota than the production server. However, these certificates are for
# testing purposes only. They are not signed for official use, so browser
# will recognise them as being insecure.
useStagingCA: false
# additionalDomains can be given to get additional certificates for the given
# domains.
# additionalDomains:
# - "www.database.de"
# javascript files for the webapp
path: ../easydb-webfrontend/build
# oauth2 client configuration to connect to
# server. This must match the configuration above.
oauth2:
# clientID used to connect to oauth2 server
clientID: web-client
# clientSecret must be provided in clear
clientSecret: foo
# Optional setting of the URL used to build the token url which is used
# for server to server communication to exchange the auth code for a
# token defaults to fylr.externalURL
# Needs to be set to the port of fylr.services.api.addr
internalURL: "http://localhost:8080/"
# optional reverse proxy. redirect endpoint
# /api and /inspect to target
reverseProxy:
api: "http://localhost:8080"
backend: "http://localhost:8081"
# custom map for reverse proxy endpoints, use pathPrefix as the key and
# the targetURL as the value. Internally fylr uses
# https://pkg.go.dev/net/http/httputil#NewSingleHostReverseProxy. It is
# possible to give a full URL here and set the source host
# (domain:port). A match of such a source URL will precede the redirect
# which would normally go to fylr.externalURL
custom:
"/system2": "http://localhost:7890"
"//other-domain/": http://localhost:9999/
# http basic auth for all access to the webfrontend
basicAuth:
root: admin
# Optional redirect. The requested path & query & hash is matched against
# the "from" and redirected to "to" using the given statuscode as http
# status. Variables can be replaced and take the form of "{...}". The
# variables can contain a-z and A-Z characters and match all characters
# (non-greedy til the next) given character. Variables at the end match
# until the end of the URL.
redirects:
- from: "/easydb?myeasydb4ID={sid}"
to: "/#/detail/{sid}"
statuscode: 301
# service execserver executes binaries and used by FYLR
# to generate previews, execute plugins and to get metadata
# of files.
execserver:
# addr of the execserver listener
# if omitted no server is started
addr: :8083
# tokenResponseSendServerIP is the IP which can be used by
# a client to send a job to. This IP is sent back to the client
# in the token response
tokenResponseSendServerIP: ""
# Mandatory path to a directory the execserver - can work in. The
# execserver will create a sub-directory per job and leave the space to
# the worker. After the work is done, the sub-directory is removed. At
# startup directories cache, job, plugin are deleted. Make sure that these
# directories do not get deleted by the OS or other processes. The
# subdirectories get cleaned by a built-in janitor process.
# Defaults to fylr.tempDir
tempDir: ""
# jobRemovalPolicy can be "none", "done", "all". "none" never removes job
# files after the job ran. However, these files are removed by the
# janitor. "done" removes succesful jobs and leaves errornous job files
# for inspection. "all" removes all job files after they ran. Defaults to
# "done".
jobRemovalPolicy: "done"
# janitorFileAge sets the age for files before they are removed. https://pkg.go.dev/time#ParseDuration
# is used to parse this value. Minimum duration is one minute. Defaults to "24h".
janitorFileAge: "24h"
waitgroups:
a:
processes: 4
b:
processes: 2
c:
processes: 4
# env can be set for all programs started by the execserver
# this is overwritten by the env set for the specific command and by the
# os environment
env:
- FYLR_METADATA_BLURHASH=1g
# set env to set threads used by ffmpeg for mp4
- FYLR_CONVERT_VIDEO_MP4_THREADS=2
# overwrite to use a different binary, defaults to "chromium" for the PDF plugin
- SERVER_PDF_CHROME=chromium
# common command defintion for all services. Also used to set FYLR_CMD_<PROG> environment
# for helper programs which are started by "fylr SUBCOMMAND". Like "exiftool" or "magick"
commands:
exiftool:
prog: exiftool
magick:
prog: magick
services:
node:
waitgroup: b
commands:
node:
prog: "node"
python3:
waitgroup: b
commands:
python3:
prog: "python3"
convert:
waitgroup: a
commands:
fylr_convert:
prog: "fylr"
args:
- "convert"
convert:
prog: "convert"
env:
- "OMP_NUM_THREADS=1"
# if startupCheck is omitted, fylr only checks if it is in PATH
# an empty startupCheck will execute the prog without any arguments
startupCheck:
# args are optional
args:
- "-version"
# If a regexp is given, the stdout of the command is checked against it
regex: "Version: ImageMagick 7..*?https://imagemagick.org"
composite:
prog: "composite"
env:
- "OMP_NUM_THREADS=1"
startupCheck:
args:
- "-version"
regex: "Version: ImageMagick 7..*?https://imagemagick.org"
fylr_metadata:
# For the blurhash production a maximum size can be set via env.
# Settings this to "-" turns the blurhash off. Blurhash needs to
# be calculated in RAM so the bigger the image is to produce a
# blurhash, the more RAM is needed. More info about blurhashes can
# be found here: https://blurha.sh/. The default for this setting
# is unlimited.
env:
- FYLR_METADATA_BLURHASH=1g
prog: "fylr"
args:
- "metadata"
ffmpeg:
waitgroup: a
commands:
ffmpeg:
prog: ffmpeg
startupCheck:
args:
- "-version"
regex: "ffmpeg version 5[\\.0-9]+ Copyright"
ffprobe:
prog: ffprobe
startupCheck:
args:
- "-version"
regex: "ffprobe version 5[\\.0-9]+ Copyright"
convert:
prog: "convert"
env:
- "OMP_NUM_THREADS=1"
startupCheck:
args:
- "-version"
regex: "Version: ImageMagick 7..*?https://imagemagick.org"
composite:
prog: "composite"
env:
- "OMP_NUM_THREADS=1"
startupCheck:
args:
- "-version"
regex: "Version: ImageMagick 7..*?https://imagemagick.org"
fylr_metadata:
env:
- FYLR_METADATA_BLURHASH=1g
prog: "fylr"
args:
- "metadata"
ffmpegthumbnailer:
prog: ffmpegthumbnailer
startupCheck:
args:
- "-v"
regex: "ffmpegthumbnailer version: 2\\..*"
soffice:
waitgroup: c
commands:
soffice:
prog: soffice
startupCheck:
args:
- "--headless"
- "--invisible"
- "--version"
regex: "LibreOffice 7[.0-9]+"
convert:
prog: "convert"
env:
- "OMP_NUM_THREADS=1"
startupCheck:
args:
- "-version"
regex: "Version: ImageMagick 7..*?https://imagemagick.org"
composite:
prog: "composite"
env:
- "OMP_NUM_THREADS=1"
startupCheck:
args:
- "-version"
regex: "Version: ImageMagick 7..*?https://imagemagick.org"
fylr_metadata:
env:
- FYLR_METADATA_BLURHASH=1g
prog: "fylr"
args:
- "metadata"
metadata:
waitgroup: a
commands:
fylr_metadata:
env:
- FYLR_METADATA_BLURHASH=1g
prog: "fylr"
args:
- "metadata"
ffprobe:
prog: ffprobe
startupCheck:
args:
- "-version"
regex: "ffprobe version 4[\\.0-9]+ Copyright"
pdf2pages:
waitgroup: a
commands:
fylr_pdf2pages:
# fylr_* utils use other programs to do their job. These
# programs must be either found in the $PATH of the OS or
# passed in by environment in the form of FYLR_CMD_<prog>
# The <prog> is the program name (upper case)
#
# pdf2pages needs
# - exiftool
# - magick
prog: "fylr"
args:
- "pdf2pages"
fylr_metadata:
env:
- FYLR_METADATA_BLURHASH=1g
prog: "fylr"
args:
- "metadata"
xslt:
waitgroup: a
commands:
xsltproc:
prog: "xsltproc"
iiif:
waitgroup: a
commands:
convert:
prog: convert
fylr_iiif:
prog: "fylr"
args:
- "iiif"
Last updated