sttp.tapir

An attribute is arbitrary data that is attached to an endpoint or endpoint input/output. The data is not interpreted by tapir's core in any way, but might be used by interpreters.

Typically, you'll add attributes using Endpoint.attribute and EndpointTransput.Basic.attribute. The attribute keys should be defined by the interpreters which are using them, and made available for import.

A bi-directional mapping between low-level values of type L and high-level values of type H. Low level values are formatted as CF.

The mapping consists of a pair of functions, one to decode (L => H), and one to encode (H => L) Decoding can fail, and this is represented as a result of type DecodeResult.

A codec also contains optional meta-data on the schema of the high-level value, as well as an instance of the format (which determines the media type of the low-level value).

Codec instances are used as implicit values, and are looked up when defining endpoint inputs/outputs. Depending on a particular endpoint input/output, it might require a codec which uses a specific format, or a specific low-level value.

Codec instances can be derived basing on other values (e.g. such as json encoders/decoders when integrating with json libraries). Or, they can be defined by hand for custom types, usually customising an existing, simpler codec.

Codecs can be chained with Mapping s using the map function.

Specifies the format of the encoded values. Each variant must be a proper type so that it can be used as a discriminator for different (implicit) instances of Codec values.

A description of an endpoint with the given inputs & outputs. The inputs are divided into two parts: security (A) and regular inputs (I). There are also two kinds of outputs: error outputs (E) and regular outputs (O).

In case there are no security inputs, the PublicEndpoint alias can be used, which omits the A parameter type.

An endpoint can be interpreted as a server, client or documentation. The endpoint requires that server/client interpreters meet the capabilities specified by R (if any).

When interpreting an endpoint as a server, the inputs are decoded and the security logic is run first, before decoding the body in the regular inputs. This allows short-circuiting further processing in case security checks fail. Server logic can be provided using EndpointServerLogicOps.serverSecurityLogic variants for secure endpoints, and EndpointServerLogicOps.serverLogic variants for public endpoints.

A concise description of an endpoint can be generated using the EndpointMetaOps.show method.

A transput is EITHER an input, or an output (see: https://ell.stackexchange.com/questions/21405/hypernym-for-input-and-output). The transput traits contain common functionality, shared by all inputs and outputs.

Note that implementations of EndpointIO can be used both as inputs and outputs.

The hierarchy is as follows:

                       /---> `EndpointInput`  >---\
`EndpointTransput` >---                            ---> `EndpointIO`
                       \---> `EndpointOutput` >---/

Inputs and outputs additionally form another hierarchy:

                                     /--> `Single` >--> `Basic` >--> `Atom`
`EndpointInput` / `EndpointOutput` >--
                                     \--> `Pair`

where the intuition behind the traits is:

• single: represents a single value, as opposed to a pair (tuple); a pair, but transformed using .map is also a single value
• basic: corresponds to an input/output which is encoded/decoded in one step; always single
• atom: corresponds to a single component of a request or response. Such inputs/outputs contain a Codec and Info meta-data, and are always basic

Lower-priority codec implicits, which transform other codecs. For example, when deriving a codec List[T] <-> Either[A, B], given codecs ca: T <-> A and cb: T <-> B, we want to get listHead(eitherRight(ca, cb)), not eitherRight(listHead(ca), listHead(cb)) (although they would function the same).

A bi-directional mapping between values of type L and values of type H.

Low-level values of type L can be decoded to a higher-level value of type H. The decoding can fail; this is represented by a result of type DecodeResult.Failure. Failures might occur due to format errors, wrong arity, exceptions, or validation errors. Validators can be added through the validate method.

High-level values of type H can be encoded as a low-level value of type L.

Mappings can be chained using one of the map functions.

Information needed to handle a multipart body. Individual parts are decoded as described by rawBodyType, which contains codecs for each part, as well as an optional default codec. The sequence of decoded parts can be further decoded into a high-level T type using codec.

Information needed to read a single part of a multipart body: the raw type (rawBodyType), and the codec which further decodes it.

The raw format of the body: what do we need to know, to read it and pass to a codec for further decoding.

Describes the type T: its low-level representation, meta-data and validation rules.

The type of the low-level representation of a T values. Part of Schemas.

Mixin containing aliases for top-level types and modules in the tapir package.

A body in any format, read using the given codec, from a raw string read using charset.

A body in any format, read using the given codec, from a raw string read using UTF-8.

Inputs which describe authentication credentials with metadata.

Usage:

binaryBody(RawBodyType.FileBody)[MyType]

, given that a codec between a file and MyType is available in the implicit scope.

binaryBody(RawBodyType.FileBody)[MyType] }}} scope.

Requires an implicit Codec.JsonCodec in scope. Such a codec can be created using Codec.json.

However, json codecs are usually derived from json-library-specific implicits. That's why integrations with various json libraries define jsonBody methods, which directly require the library-specific implicits.

Unless you have defined a custom json codec, the jsonBody methods should be used.

An empty output. Useful if one of the oneOf branches of a coproduct type is a case object that should be mapped to an empty body.

Extract a value from a server request. This input is only used by server interpreters, it is ignored by documentation interpreters and the provided value is discarded by client interpreters.

A server endpoint, which exposes a single file from local storage found at systemPath, using the given path.

fileGetServerEndpoint("static" / "hello.html")("/home/app/static/data.html")

Create pair of endpoints (head, get) for particular file

A server endpoint, which exposes files from local storage found at systemPath, using the given prefix. Typically, the prefix is a path, but it can also contain other inputs. For example:

filesGetServerEndpoint("static" / "files")("/home/app/static")

A request to /static/files/css/styles.css will try to read the /home/app/static/css/styles.css file.

A server endpoint, used to verify if sever supports range requests for file under particular path Additionally it verify file existence and returns its size

An output which contains a number of variant outputs. Each variant can contain different outputs and represent different content. To describe an output which represents same content, but with different content types, use oneOfBody.

All possible outputs must have a common supertype (T). Typically, the supertype is a sealed trait, and the variants are implementing case classes.

When encoding to a response, the first matching output is chosen, using the following rules:

1. the variant's appliesTo method, applied to the output value (as returned by the server logic) must return true.
2. when a fixed content type is specified by the output, it must match the request's Accept header (if present). This implements content negotiation.

When decoding from a response, the first output which decodes successfully is chosen.

The outputs might vary in status codes, headers (e.g. different content types), and body implementations. However, for bodies, only replayable ones can be used, and they need to have the same raw representation (e.g. all byte-array-base, or all file-based).

Note that exhaustiveness of the variants (that all subtypes of T are covered) is not checked.

See oneOfBody.

Allows explicitly specifying the content type range, for which each body will be used, instead of defaulting to the exact media type as specified by the body's codec. This is only used when choosing which body to decode.

A body input or output, which can be one of the given variants. All variants should represent T instances using different content types. Hence, the content type is used as a discriminator to choose the appropriate variant.

Should be used to describe an input or output which represents the same content, but using different content types. For an output which describes variants including possibly different outputs (representing different content), see oneOf.

The server behavior is as follows:

• when encoding to a response, the first variant matching the request's Accept header is chosen (if present). Otherwise, the first variant is used. This implements content negotiation.
• when decoding a request, the variant corresponding to the request's Content-Type header is chosen (if present). Otherwise, a decode failure is returned, which by default results in an 415 Unsupported Media Type response.

The client behavior is as follows:

• when encoding a request, the first variant is used.
• when decoding a response, the variant corresponding to the response's Content-Type header is chosen (if present). Otherwise, the first variant is used. For client interpreters to work correctly, all body variants must have the same raw type (e.g. all are string-based or all byte-array-based)

All possible bodies must have the same type T. Typically, the bodies will vary in the Codecs that are used for the body.

Create a fallback variant to be used in oneOf output descriptions. Multiple such variants can be specified, with different body content types.

Create a one-of-variant which uses output if the class of the provided value (when interpreting as a server) matches the runtime class of T. Adds a fixed status-code output with the given value.

This will fail at compile-time if the type erasure of T is different from T, as a runtime check in this situation would give invalid results. In such cases, use oneOfVariantClassMatcher, oneOfVariantValueMatcher or oneOfVariantFromMatchType instead.

Should be used in oneOf output descriptions.

Create a one-of-variant which uses output if the class of the provided value (when interpreting as a server) matches the runtime class of T.

This will fail at compile-time if the type erasure of T is different from T, as a runtime check in this situation would give invalid results. In such cases, use oneOfVariantClassMatcher, oneOfVariantValueMatcher or oneOfVariantFromMatchType instead.

Should be used in oneOf output descriptions.

Create a one-of-variant which uses output i the class of the provided value (when interpreting as a server) matches the given runtimeClass. Note that this does not take into account type erasure. Adds a fixed status-code output with the given value.

Should be used in oneOf output descriptions.

Create a one-of-variant which uses output if the class of the provided value (when interpreting as a server) matches the given runtimeClass. Note that this does not take into account type erasure.

Should be used in oneOf output descriptions.

Create a one-of-variant which uses output if the provided value exactly matches one of the values provided in the second argument list. Adds a fixed status-code output with the given value.

Should be used in oneOf output descriptions.

Create a one-of-variant which output if the provided value exactly matches one of the values provided in the second argument list.

Should be used in oneOf output descriptions.

Experimental!

Create a one-of-variant which uses output if the provided value matches the target type, as checked by MatchType. Instances of MatchType are automatically derived and recursively check that classes of all fields match, to bypass issues caused by type erasure. Adds a fixed status-code output with the given value.

Should be used in oneOf output descriptions.

Experimental!

Create a one-of-variant which uses output if the provided value matches the target type, as checked by MatchType. Instances of MatchType are automatically derived and recursively check that classes of all fields match, to bypass issues caused by type erasure.

Should be used in oneOf output descriptions.

Create a one-of-variant which uses output if the provided value (when interpreting as a server matches the matcher predicate. Adds a fixed status-code output with the given value.

Should be used in oneOf output descriptions.

Create a one-of-variant which uses output if the provided value (when interpreting as a server matches the matcher predicate.

Should be used in oneOf output descriptions.

A server endpoint, which exposes a single resource available from the given classLoader at resourcePath, using the given path.

resourceGetServerEndpoint("static" / "hello.html")(classOf[App].getClassLoader, "app/data.html")

A server endpoint, which exposes resources available from the given classLoader, using the given prefix. Typically, the prefix is a path, but it can also contain other inputs. For example:

resourcesGetServerEndpoint("static" / "files")(classOf[App].getClassLoader, "app")

A request to /static/files/css/styles.css will try to read the /app/css/styles.css resource.

An fixed status code output.

An output which maps to the status code in the response.

Creates a stream body with a binary schema. The application/octet-stream media type will be used by default, but can be later overridden by providing a custom Content-Type header value.

Creates a stream body with the given schema.

Creates a stream body with a text schema.

Requires an implicit Codec.XmlCodec in scope. Such a codec can be created using Codec.xml.

An empty output.

An input which matches if the request URI ends with a trailing slash, otherwise the result is a decode failure on the path. Has no effect when used by documentation or client interpreters.