config_ipfs.yaml

# Path to a directory to where peer metadata and the overlay graph will be written.
output_directory_path: "output_data_crawls/ipfs"

# Path to a file to use as a node cache.
# The node cache is read at startup. All peers in the node cache will be
# contacted by the crawler. This should speed up the crawl, but only works if
# the cache is up-to-date. Due to churn, this is generally only the case for
# crawls that are performed immediately after one another.
#cache_file_path: nodes.cache

# Settings for the crawler
crawler:
  # The number of libp2p hosts to run.
  num_workers: 5

  # The maximum number of concurrent in-flight requests.
  concurrent_requests: 1000

  # Path to the (compressed) preimage file.
  preimage_file_path: "precomputed_hashes/preimages.csv.zst"

  # The bootstrap peers to connect to.
  bootstrap_peers:
    - /dnsaddr/bootstrap.libp2p.io/p2p/QmNnooDu7bfjPFoTZYxMNLWUQJyrVwtbZg5gBMjTezGAJN
    - /dnsaddr/bootstrap.libp2p.io/p2p/QmQCU2EcMqAqQPR2i9bChDtGNJchTbq5TbXJJ16u19uLTa
    - /dnsaddr/bootstrap.libp2p.io/p2p/QmbLHAnMoJPWSCR5Zhtx6BHJX9KiKNN6tpvbUcqanj75Nb
    - /dnsaddr/bootstrap.libp2p.io/p2p/QmcZf59bWwK5XFi76CZX8cbJ4BhTzzA3gU1ZjYZcYW3dwt
    - /ip4/104.131.131.82/tcp/4001/p2p/QmaCpDMGvV2BGHeYERUEnRQAwe3N8SzbUtfsmvsqQLuvuJ

  # Configuration of the libp2p hosts.
  worker_config:
    # The user agent to announce as.
    user_agent: "ipfs_crawler (https://github.com/trudi-group/ipfs-crawler)"

    # The timeout to establish a connection to a peer.
    connect_timeout: 180s

    # The number of times a connection attempt will be made.
    connection_attempts: 3

  # Configuration for the crawler "plugin"
  crawler_config:
    # The timeout for non-connection interactions.
    # These are steps of the crawling process, such as establishing a stream and
    # requesting a set of known peers. As such the total timeout to get all
    # neighbors of a peer is a multiple of this.
    interaction_timeout: 5s

    # The number of times each interaction is attempted.
    interaction_attempts: 10

    # The protocols to use for crawling.
    protocol_strings:
      - /ipfs/kad/1.0.0

  # Configuration for plugins.
  # Plugins are executed once a peer has been crawled completely, in the order
  # given here.
  plugins:

    # Configuration for the Bitswap probe plugin
#    - name: "bitswap-probe"
#      options:
#        # A list of CIDs to ask for
#        cids:
#          # CID of the IPFS logo
#          - "QmY7Yh4UquoXHLPFo2XbhXkhBvFoPwmQUSa92pxnxjQuPU"
#
#        # The timeout to use for requests
#        request_timeout: "5s"
#
#        # The period of time to wait for replies
#        response_period: "30s"