aleph.http
connect
(connect url)(connect url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a CONNECT request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
connection-pool
(connection-pool {:keys [connections-per-host total-connections target-utilization connection-options dns-options stats-callback control-period middleware max-queue-size], :or {connections-per-host 8, total-connections 1024, target-utilization 0.9, control-period 60000, middleware middleware/wrap-request, max-queue-size 65536}})
Returns a connection pool which can be used as an argument in `request`.
|:---|:---
| `connections-per-host` | the maximum number of simultaneous connections to any host
| `total-connections` | the maximum number of connections across all hosts
| `target-utilization` | the target utilization of connections per host, within `[0,1]`, defaults to `0.9`
| `stats-callback` | an optional callback which is invoked with a map of hosts onto usage statistics every ten seconds
| `max-queue-size` | the maximum number of pending acquires from the pool that are allowed before `acquire` will start to throw a `java.util.concurrent.RejectedExecutionException`, defaults to `65536`
| `control-period` | the interval, in milliseconds, between use of the controller to adjust the size of the pool, defaults to `60000`
| `dns-options` | an optional map with async DNS resolver settings, for more information check `aleph.netty/dns-resolver-group`. When set, ignores `name-resolver` setting from `connection-options` in favor of shared DNS resolver instance
| `middleware` | a function to modify request before sending, defaults to `aleph.http.client-middleware/wrap-request`
the `connection-options` are a map describing behavior across all connections:
|:---|:---
| `ssl-context` | an `io.netty.handler.ssl.SslContext` object, only required if a custom context is required
| `local-address` | an optional `java.net.SocketAddress` describing which local interface should be used
| `bootstrap-transform` | a function that takes an `io.netty.bootstrap.Bootstrap` object and modifies it.
| `pipeline-transform` | a function that takes an `io.netty.channel.ChannelPipeline` object, which represents a connection, and modifies it.
| `insecure?` | if `true`, ignores the certificate for any `https://` domains
| `response-buffer-size` | the amount of the response, in bytes, that is buffered before the request returns, defaults to `65536`. This does *not* represent the maximum size response that the client can handle (which is unbounded), and is only a means of maximizing performance.
| `keep-alive?` | if `true`, attempts to reuse connections for multiple requests, defaults to `true`.
| `idle-timeout` | when set, forces keep-alive connections to be closed after an idle time, in milliseconds.
| `epoll?` | if `true`, uses `epoll` when available, defaults to `false`
| `raw-stream?` | if `true`, bodies of responses will not be buffered at all, and represented as Manifold streams of `io.netty.buffer.ByteBuf` objects rather than as an `InputStream`. This will minimize copying, but means that care must be taken with Netty's buffer reference counting. Only recommended for advanced users.
| `max-initial-line-length` | the maximum length of the initial line (e.g. HTTP/1.0 200 OK), defaults to `65536`
| `max-header-size` | the maximum characters that can be in a single header entry of a response, defaults to `65536`
| `max-chunk-size` | the maximum characters that can be in a single chunk of a streamed response, defaults to `65536`
| `name-resolver` | specify the mechanism to resolve the address of the unresolved named address. When not set or equals to `:default`, JDK's built-in domain name lookup mechanism is used (blocking). Set to`:noop` not to resolve addresses or pass an instance of `io.netty.resolver.AddressResolverGroup` you need. Note, that if the appropriate connection-pool is created with dns-options shared DNS resolver would be used
| `proxy-options` | a map to specify proxy settings. HTTP, SOCKS4 and SOCKS5 proxies are supported. Note, that when using proxy `connections-per-host` configuration is still applied to the target host disregarding tunneling settings. If you need to limit number of connections to the proxy itself use `total-connections` setting.
| `response-executor` | optional `java.util.concurrent.Executor` that will execute response callbacks
| `log-activity` | when set, logs all events on each channel (connection) with a log level given. Accepts either one of `:trace`, `:debug`, `:info`, `:warn`, `:error` or an instance of `io.netty.handler.logging.LogLevel`. Note, that this setting *does not* enforce any changes to the logging configuration (default configuration is `INFO`, so you won't see any `DEBUG` or `TRACE` level messages, unless configured explicitly)
Supported `proxy-options` are
|:---|:---
| `host` | host of the proxy server
| `port` | an optional port to establish connection (defaults to 80 for http and 1080 for socks proxies)
| `protocol` | one of `:http`, `:socks4` or `:socks5` (defaults to `:http`)
| `user` | an optional auth username
| `password` | an optional auth password
| `http-headers` | (HTTP proxy only) an optional map to set additional HTTP headers when establishing connection to the proxy server
| `tunnel?` | (HTTP proxy only) if `true`, sends HTTP CONNECT to the proxy and waits for the 'HTTP/1.1 200 OK' response before sending any subsequent requests. Defaults to `false`. When using authorization or specifying additional headers uses tunneling disregarding this setting
| `connection-timeout` | timeout in milliseconds for the tunnel become established, defaults to 60 seconds, setting is ignored when tunneling is not used.
default-response-executor
delete
(delete url)(delete url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a DELETE request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
file
(file path)(file path offset length)(file path offset length chunk-size)
Specifies a file or a region of the file to be sent over the network.
Accepts string path to the file, instance of `java.io.File` or instance of
`java.nio.file.Path`.
get
(get url)(get url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a GET request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
get-all
(get-all headers k)
Given a header map from an HTTP request or response, returns a collection of values associated with the key,
rather than a comma-delimited string.
head
(head url)(head url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a HEAD request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
options
(options url)(options url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a OPTIONS request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
patch
(patch url)(patch url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a PATCH request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
post
(post url)(post url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a POST request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
put
(put url)(put url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a PUT request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
register-connection-stats-callback
(register-connection-stats-callback c)
Registers a callback which will be called with connection-pool stats.
request
(request {:keys [pool middleware pool-timeout response-executor connection-timeout request-timeout read-timeout follow-redirects?], :or {pool default-connection-pool, response-executor default-response-executor, middleware identity, connection-timeout 60000.0}, :as req})
Takes an HTTP request, as defined by the Ring protocol, with the extensions defined
by [clj-http](https://github.com/dakrone/clj-http), and returns a deferred representing
the HTTP response. Also allows for a custom `pool` or `middleware` to be defined.
|:---|:---
| `pool` | a custom connection pool
| `middleware` | custom client middleware for the request
| `pool-timeout` | timeout in milliseconds for the pool to generate a connection
| `connection-timeout` | timeout in milliseconds for the connection to become established
| `request-timeout` | timeout in milliseconds for the arrival of a response over the established connection
| `read-timeout` | timeout in milliseconds for the response to be completed
| `follow-redirects?` | whether to follow redirects, defaults to `true`; see `aleph.http.client-middleware/handle-redirects`
start-server
(start-server handler options)
Starts an HTTP server using the provided Ring `handler`. Returns a server object which can be stopped
via `java.io.Closeable.close()`, and whose port can be discovered with `aleph.netty/port`.
|:---------|:-------------
| `port` | the port the server will bind to. If `0`, the server will bind to a random port.
| `socket-address` | a `java.net.SocketAddress` specifying both the port and interface to bind to.
| `bootstrap-transform` | a function that takes an `io.netty.bootstrap.ServerBootstrap` object, which represents the server, and modifies it.
| `ssl-context` | an `io.netty.handler.ssl.SslContext` object if an SSL connection is desired |
| `manual-ssl?` | set to `true` to indicate that SSL is active, but the caller is managing it (this implies `:ssl-context` is nil). For example, this can be used if you want to use configure SNI (perhaps in `:pipeline-transform`) to select the SSL context based on the client's indicated host name. |
| `pipeline-transform` | a function that takes an `io.netty.channel.ChannelPipeline` object, which represents a connection, and modifies it.
| `executor` | a `java.util.concurrent.Executor` which is used to handle individual requests. To avoid this indirection you may specify `:none`, but in this case extreme care must be taken to avoid blocking operations on the handler's thread.
| `shutdown-executor?` | if `true`, the executor will be shut down when `.close()` is called on the server, defaults to `true`.
| `request-buffer-size` | the maximum body size, in bytes, which the server will allow to accumulate before invoking the handler, defaults to `16384`. This does *not* represent the maximum size request the server can handle (which is unbounded), and is only a means of maximizing performance.
| `raw-stream?` | if `true`, bodies of requests will not be buffered at all, and will be represented as Manifold streams of `io.netty.buffer.ByteBuf` objects rather than as an `InputStream`. This will minimize copying, but means that care must be taken with Netty's buffer reference counting. Only recommended for advanced users.
| `rejected-handler` | a spillover request-handler which is invoked when the executor's queue is full, and the request cannot be processed. Defaults to a `503` response.
| `max-initial-line-length` | the maximum characters that can be in the initial line of the request, defaults to `8192`
| `max-header-size` | the maximum characters that can be in a single header entry of a request, defaults to `8192`
| `max-chunk-size` | the maximum characters that can be in a single chunk of a streamed request, defaults to `16384`
| `epoll?` | if `true`, uses `epoll` when available, defaults to `false`
| `compression?` | when `true` enables http compression, defaults to `false`
| `compression-level` | optional compression level, `1` yields the fastest compression and `9` yields the best compression, defaults to `6`. When set, enables http content compression regardless of the `compression?` flag value
| `idle-timeout` | when set, forces keep-alive connections to be closed after an idle time, in milliseconds
| `continue-handler` | optional handler which is invoked when header sends "Except: 100-continue" header to test whether the request should be accepted or rejected. Handler should return `true`, `false`, ring responseo to be used as a reject response or deferred that yields one of those.
| `continue-executor` | optional `java.util.concurrent.Executor` which is used to handle requests passed to :continue-handler. To avoid this indirection you may specify `:none`, but in this case extreme care must be taken to avoid blocking operations on the handler's thread.
trace
(trace url)(trace url {:keys [pool middleware headers body multipart], :or {pool default-connection-pool, middleware identity}, :as options})
Makes a TRACE request, returns a deferred representing
the response.
|:---|:---
| `pool` | the `connection-pool` that should be used, defaults to the `default-connection-pool`
| `middleware` | any additional middleware that should be used for handling requests and responses
| `headers` | the HTTP headers for the request
| `body` | an optional body, which should be coercable to a byte representation via [byte-streams](https://github.com/ztellman/byte-streams)
| `multipart` | a vector of bodies
unregister-connection-stats-callback
(unregister-connection-stats-callback c)
Unregisters a previous connection-pool stats callback.
websocket-client
(websocket-client url)(websocket-client url options)
Given a url, returns a deferred which yields a duplex stream that can be used to
communicate with a server over the WebSocket protocol.
|:---|:---
| `raw-stream?` | if `true`, the connection will emit raw `io.netty.buffer.ByteBuf` objects rather than strings or byte-arrays. This will minimize copying, but means that care must be taken with Netty's buffer reference counting. Only recommended for advanced users.
| `insecure?` | if `true`, the certificates for `wss://` will be ignored.
| `ssl-context` | an `io.netty.handler.ssl.SslContext` object, only required if a custom context is required
| `extensions?` | if `true`, the websocket extensions will be supported.
| `sub-protocols` | a string with a comma seperated list of supported sub-protocols.
| `headers` | the headers that should be included in the handshake
| `compression?` | when set to `true`, enables client to use permessage-deflate compression extension, defaults to `false`.
| `pipeline-transform` | an optional function that takes an `io.netty.channel.ChannelPipeline` object, which represents a connection, and modifies it.
| `max-frame-payload` | maximum allowable frame payload length, in bytes, defaults to `65536`.
| `max-frame-size` | maximum aggregate message size, in bytes, defaults to `1048576`.
| `bootstrap-transform` | an optional function that takes an `io.netty.bootstrap.Bootstrap` object and modifies it.
| `epoll?` | if `true`, uses `epoll` when available, defaults to `false`
| `heartbeats` | optional configuration to send Ping frames to the server periodically (if the connection is idle), configuration keys are `:send-after-idle` (in milliseconds), `:payload` (optional, empty frame by default) and `:timeout` (optional, to close the connection if Pong is not received after specified timeout).
websocket-close!
(websocket-close! conn)(websocket-close! conn status-code)(websocket-close! conn status-code reason-text)(websocket-close! conn status-code reason-text deferred)
Closes given websocket endpoint (either client or server) sending Close frame with provided
status code and reason text. Returns a deferred that will yield `true` whenever the closing
handshake was initiated with given params or `false` if the connection was already closed.
Note, that for the server closes the connection right after Close frame was flushed but the
client waits for the connection to be closed by the server (no longer than close handshake
timeout, see websocket connection configuration for more details).
websocket-connection
(websocket-connection req)(websocket-connection req options)
Given an HTTP request that can be upgraded to a WebSocket connection, returns a
deferred which yields a duplex stream that can be used to communicate with the
client over the WebSocket protocol.
|:---|:---
| `raw-stream?` | if `true`, the connection will emit raw `io.netty.buffer.ByteBuf` objects rather than strings or byte-arrays. This will minimize copying, but means that care must be taken with Netty's buffer reference counting. Only recommended for advanced users.
| `headers` | the headers that should be included in the handshake
| `compression?` | when set to `true`, enables permessage-deflate compression extention support for the connection, defaults to `false`.
| `pipeline-transform` | an optional function that takes an `io.netty.channel.ChannelPipeline` object, which represents a connection, and modifies it.
| `max-frame-payload` | maximum allowable frame payload length, in bytes, defaults to `65536`.
| `max-frame-size` | maximum aggregate message size, in bytes, defaults to `1048576`.
| `allow-extensions?` | if true, allows extensions to the WebSocket protocol, defaults to `false`.
| `heartbeats` | optional configuration to send Ping frames to the client periodically (if the connection is idle), configuration keys are `:send-after-idle` (in milliseconds), `:payload` (optional, empty uses empty frame by default) and `:timeout` (optional, to close the connection if Pong is not received after specified timeout).
websocket-ping
(websocket-ping conn)(websocket-ping conn d')(websocket-ping conn d' data)
Takes a websocket endpoint (either client or server) and returns a deferred that will
yield true whenever the PONG comes back, or false if the connection is closed. Subsequent
PINGs are supressed to avoid ambiguity in a way that the next PONG trigger all pending PINGs.
wrap-ring-async-handler
(wrap-ring-async-handler handler)