FediCrawl

Collect feeds to follow on Fediverse nodes.

App crawls from node to node (starting from the seed node) and searches for feeds (accounts or channels) and for other peered nodes.

Discovered nodes are added to database and are queued for next search. New nodes are processed preferentially. If there is no new node, data of oldest node are refreshed.

Node data are retrieved using node info api.

Node's public feeds and peering nodes are retrieved only on several supported softwares using thier's public APIs.

Supported Fediverse software

For now only two fediverse apps are supported:

• Mastodon
• Pleroma (hopefully after release of version 4.2)
• Peertube
• Misskey

Data providers for more apps will be probably added soon (Pull requests are welcomed)

Config

Configuration is done using environmental variables:

Variable	Description	Default value / Example value
ELASTIC_URL	Url address of ElasticSearch server	http://elastic:9200
ELASTIC_USER	Username for EalsticSearch server	elastic
ELASTIC_PASSWORD	Username for EalsticSearch server	empty
SEED_NODE_DOMAIN	Domain of the first node to search users and other nodes on	mastodon.social,mastodon.online
REATTEMPT_MINUTES	Optional, How many minutes should be waited for next node refresh attempt if the refresh fails	60
REFRESH_HOURS	Optional, How often (in hours) should be node info refreshed	120
WAIT_FOR_JOB_MINUTES	Optional, How many minutes should the thread sleep if there are no nodes to refresh	60
DEFAULT_TIMEOUT_MILLISECONDS	Optional, How many milliseconds should http wait for node api response on refresh	10000
SEED_TIMEOUT_MILLISECONDS	Optional, How many milliseconds should http wait for node api response on refresh of seed domains	value of DEFAULT_TIMEOUT_MILLISECONDS
BANNED_DOMAINS	Optional, Domains not to index (even with subdomains)	empty
CRAWLING_VERSION	Optional, Increasing this number can enforce recrawling of the whole index	0
MAX_CRAWLING_DEPTH	Optional, Limits how far is fediverse indexed from seed nodes	empty
TZ	Optional, Timezone	UTC

Deploy

App is designed to be run in docker container and deployed using docker-compose. More info can be found in FediSearch example docker-compose project

For searching in collected feeds there is a companion server app FediSearch