aleph.http — com.github.arnaudgeiser/aleph 0.4.8

connect clj

(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 | multipart | a vector of bodies

connection-pool clj

(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-connection-pool clj

default-response-executor clj

delete clj

(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 | multipart | a vector of bodies

file clj

(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 clj

(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 | multipart | a vector of bodies

get-all clj

(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 clj

(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 | multipart | a vector of bodies

options clj

(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 | multipart | a vector of bodies

patch clj

(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 | multipart | a vector of bodies

post clj

(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 | multipart | a vector of bodies

put clj

(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 | multipart | a vector of bodies

register-connection-stats-callback clj

(register-connection-stats-callback c)

Registers a callback which will be called with connection-pool stats.

request clj

(request {:keys [pool middleware pool-timeout response-executor
                 connection-timeout request-timeout read-timeout]
          :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, 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 | response-executor | optional java.util.concurrent.Executor that will handle the requests (defaults to a flow/utilization-executor of 256 max-threads and a queue-length of 0) | 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 clj

(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 clj

(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 | multipart | a vector of bodies

unregister-connection-stats-callback clj

(unregister-connection-stats-callback c)

Unregisters a previous connection-pool stats callback.

websocket-client clj

(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! clj

(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 clj

(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 clj

(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 clj

(wrap-ring-async-handler handler)

Converts given asynchronous Ring handler to Aleph-compliant handler.

More information about asynchronous Ring handlers and middleware: https://www.booleanknot.com/blog/2016/07/15/asynchronous-ring.html