Filecoin v1 retrievals
V1 retrievals on Filecoin are done using HTTP range requests to Storage providers hosting the files. Currently, this requires that you provide multiple pieces of information:
A metadata.yml file that specifies the way that the full epoch car file was split into Filecoin pieces
A deals.csv file that maps the individual Filecoin pieces to storage providers
A sample config file for an epoch using Filecoin v1 retrievals would look like this:
version: 1
epoch: 100
data:
car:
from_pieces:
deals:
uri: '/of1/100/deals.csv'
metadata:
uri: '/of1/100/metadata.yaml'
indexes:
cid_to_offset_and_size:
uri: /of1/100/epoch-100-bafyreibqt2nvroysxlxctgb52xxn27ectsllv2xyka4qar7ga6vupmbs3i-mainnet-cid-to-offset-and-size.index
slot_to_cid:
uri: /of1/100/epoch-100-bafyreibqt2nvroysxlxctgb52xxn27ectsllv2xyka4qar7ga6vupmbs3i-mainnet-slot-to-cid.index
sig_to_cid:
uri: /of1/100/epoch-100-bafyreibqt2nvroysxlxctgb52xxn27ectsllv2xyka4qar7ga6vupmbs3i-mainnet-sig-to-cid.index
sig_exists:
uri: /of1/100/epoch-100-bafyreibqt2nvroysxlxctgb52xxn27ectsllv2xyka4qar7ga6vupmbs3i-mainnet-sig-exists.index
gsfa:
uri: /of1/100/epoch-100-bafyreibqt2nvroysxlxctgb52xxn27ectsllv2xyka4qar7ga6vupmbs3i-mainnet-gsfa.indexdir
All indexes except gsfa can use either a local file system URL or an HTTP URL.
Last updated