a0d44bbb61
with a expanded profiles length check that is more accurate |
||
---|---|---|
.. | ||
src | ||
build.rs | ||
Cargo.toml | ||
README.md |
ipfs-http crate
HTTP api on top of ipfs
crate. The end binary has some rudimentary ipfs CLI
functionality but mostly in the aim of testing the rust-ipfs
via:
The vision for this crate is to eventually provide warp filters and async methods suitable to providing the Ipfs HTTP API in other applications as well instead of having to write application specific debug and introspection APIs.
HTTP specs:
Status: Pre-alpha, most of the functionality is missing or 501 Not Implemented
. See the repository level README for more information.
Getting started
This tutorial will demonstrate how to run a rust-ipfs node using the ipfs-http crate. If you haven't already, you'll need to install Rust. You should also install the go-ipfs CLI as this will make it easier to interact with the node.
By default ipfs-http stores the configuration for the node in the .rust-ipfs
directory. Should you want to override this, you can do so by setting the
IPFS_PATH
environment variable. For this example it's a good idea to set the
path to .rust-ipfs
so that the go-ipfs CLI knows to use that directory as
well (default is .ipfs
).
You can initialise the directory with:
cargo run -p ipfs-http -- init --profile test --bits 2048
The --profile
option allows the user to use a set of defaults. Currently two
profiles are supported:
test
: runs the daemon using ephemeral portsdefault
runs the daemon on port4004
The --bits
option specifies the length of the RSA keys to be used. The output
should return a peer id and confirm the path of the newly initialised node is
either the default or the IPFS_PATH
.
The .rust-ipfs
directory now contains a configuration file, config
:
{
"Identity": {
"PeerID": "QmTETy4bmL44fwkvbkMzXMVmiUDTvEcupsfpM8BCgNERUe",
"PrivKey": "CAASpgkwggSiAgEAAoIBAQCyFR6pKSRt62WLJ6fi2MeG0pn [...]"
},
"Addresses": {
"Swarm": [
"/ip4/127.0.0.1/tcp/0"
]
}
}
It stores the peer id, private key (shortened for brevity) and swarm addresses for the node. Let's run the node as a daemon:
cargo run -p ipfs-http -- daemon
This exposes the node as an HTTP API. The config directory has also grown to
include a blockstore
, a datastore
and an api
file:
.rust-ipfs
├── api
├── blockstore
├── config
└── datastore
└── pins
The blockstore
and datastore
are empty, as we haven't yet added any data to
the ipfs node. The api
file keeps track of the node's address.
The node can now be queried using the go-ipfs CLI. In another terminal window run:
ipfs id
This returns the information about the node.
{
"ID": "QmTETy4bmL44fwkvbkMzXMVmiUDTvEcupsfpM8BCgNERUe",
"PublicKey": "CAASpgIwggEiMA0GCSqGSIb3D [...]",
"Addresses": [
"/ip4/127.0.0.1/tcp/58807/p2p/QmTETy4bmL44fwkvbkMzXMVmiUDTvEcupsfpM8BCgNERUe"
],
"AgentVersion": "rust-ipfs/version",
"ProtocolVersion": "ipfs/version",
"Protocols": null
}
The query is logged by the node and shows the /api/v0/id
endpoint handled the request:
INFO ipfs-http: 127.0.0.1:58811 "POST /api/v0/id HTTP/1.1" 200 "-" "go-ipfs-cmds/http" 2.795971ms