align to Function SDK updates (#16)

* align to Function SDK updates * dictionary and link updates * more updates * Update sdk/inference.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update sdk/inference.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update sdk/connection.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * remove Schema note from Manifest * Update deploy.mdx * Update sdk/functions-sdk.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update sdk/connection.mdx * Update sdk/connection.mdx * Update sdk/connection.mdx * Update sdk/connection.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update sdk/functions-sdk.mdx * formatting * Update sdk/connection.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update sdk/connection.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update sdk/connection.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update connection.mdx * Update sdk/connection.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * Update connection.mdx * Update sdk/connection.mdx Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com> * separate response types * edits * Update http.mdx * URL config --------- Co-authored-by: Matt Johnson-Pint <mattjohnsonpint@gmail.com>
hypermodeinc · May 30, 2024 · b13463c · b13463c
1 parent 0c1a7d4
commit b13463c
Show file tree

Hide file tree

Showing 9 changed files with 241 additions and 24 deletions.
diff --git a/.trunk/configs/.markdownlint.json b/.trunk/configs/.markdownlint.json
@@ -1,4 +1,5 @@
 {
   "MD013": false,
-  "MD033": false
+  "MD033": false,
+  "MD034": false
 }
diff --git a/deploy.mdx b/deploy.mdx
@@ -11,7 +11,7 @@ Hypermode leverages a native GitHub integration for the deployment of Hypermode
 
 ## Build
 
-A GitHub Action is available in the [template project](https://github.com/hypermodeAI/sample-project) to handle the build of your project's functions. This Action must complete successfully before your functions deploy.
+The [template project](https://github.com/hypermodeAI/template-project) includes a GitHub Action to build your project's functions. On success, the action deploys your functions to Hypermode when committing or merging to the branch defined in the Hypermode Console.```
 
 ## Deploy
 

diff --git a/mint.json b/mint.json
@@ -59,14 +59,15 @@
       "group": "Write Functions",
       "pages": [
         "sdk/functions-sdk",
+        "sdk/http",
         "sdk/connection",
         "sdk/inference",
         "sdk/search",
         "sdk/console"
       ]
     },
     {
-      "group": "Iterate Constantly",
+      "group": "Observe and Iterate",
       "pages": ["deploy", "function-observability"]
     }
   ],

diff --git a/sdk/connection.mdx b/sdk/connection.mdx
@@ -16,21 +16,64 @@ Connection APIs allow you to securely access an API endpoint at a defined host.
 Make a query or mutation against a GraphQL API endpoint.
 
 ```TypeScript
-invokeGraphqlApi (
+invokeGraphqlApi<TData> (
   hostName: string,
-  query: string,
-  variables?: Map<string, string>
-): string
+  statement: string,
+  variables?: QueryVariables
+): GQLResponse<TData>
 ```
 
+<ResponseField name="TData" type="Type" required>
+  Type of object to use for the data returned from the query.
+</ResponseField>
+
 <ResponseField name="hostName" type="string" required>
   Internal name of your host, as [defined in your manifest](../define-hosts).
 </ResponseField>
 
-<ResponseField name="query" type="string" required>
+<ResponseField name="statement" type="string" required>
   GraphQL operation for execution on your API server to retrieve or write data.
 </ResponseField>
 
-<ResponseField name="variables" type="string">
+<ResponseField name="variables" type="QueryVariables">
   Optional variables added to the query.
+
+Example:
+
+```TypeScript
+const vars = new QueryVariables();
+vars.set("firstName", firstName);
+vars.set("lastName", lastName);
+```
+
+</ResponseField>
+
+### GQLResponse object
+
+A response object from the GraphQL query.
+
+<ResponseField name="data" type="TData">
+  The results of the GraphQL operation.
+</ResponseField>
+<ResponseField name="errors" type="GQLError[]">
+  Any errors incurred as part of your GraphQL request.
+  <Expandable title="properties">
+    <ResponseField name="message" type="string">
+      Description of the error incurred.
+    </ResponseField>
+    <ResponseField name="path" type="string[]">
+      Query path of the error incurred. Each item in the array represents a
+      segment of the path.
+    </ResponseField>
+    <ResponseField name="locations" type="GQLErrorLocation[]">
+      <Expandable title="properties">
+        <ResponseField name="line" type="u32">
+          Specific line number of the request that points to the error.
+        </ResponseField>
+        <ResponseField name="column" type="u32">
+          Specific column number of the request that points to the error.
+        </ResponseField>
+      </Expandable>
+    </ResponseField>
+  </Expandable>
 </ResponseField>
diff --git a/sdk/functions-sdk.mdx b/sdk/functions-sdk.mdx
@@ -8,18 +8,22 @@ The Functions SDK gives your project access to a set of APIs that simplify the d
 
 <Note>
   The Functions SDK includes support for AssemblyScript, a portable
-  TypeScript-like language. Support for Golang is in development.
+  TypeScript-like language. Additional language support is in development.
 </Note>
 
+For inspiration on getting started, reference the [examples in the Function SDK repo](https://github.com/hypermodeAI/functions-as/tree/main/examples).
+
 ## APIs
 
 <CardGroup cols={2}>
-  <Card title="Connection" icon="server" href="/sdk/connection">
+  <Card title="HTTP" icon="server" href="/sdk/http">
+    Securely call and fetch data from an HTTP endpoint.
+  </Card>
+  <Card title="Connection" icon="database" href="/sdk/connection">
     Augment functions with secure access to data in your databases and APIs.
   </Card>
   <Card title="Inference" icon="message-plus" href="/sdk/inference">
-    Use AI models for classification, summarization, and other generation use
-    cases.
+    Use AI models for classifying, summarizing, and generation use cases.
   </Card>
   <Card title="Search" icon="searchengin" href="/sdk/search">
     Use embeddings for semantic similarity search, clustering, and outlier

diff --git a/sdk/http.mdx b/sdk/http.mdx
@@ -0,0 +1,72 @@
+---
+title: HTTP
+description: ""
+---
+
+The HTTP API allows you to securely call and fetch data from an HTTP endpoint.
+
+<Tip>
+  We're introducing new APIs consistently through ongoing development with build
+  partners. [Let's chat](mailto:help@hypermode.com) about what would make the
+  Functions SDK even more powerful for your next use case!
+</Tip>
+
+### fetch
+
+Invoke an HTTP endpoint to retrieve data or trigger external action.
+
+```TypeScript
+fetch (
+  request: Request,
+): Response
+```
+
+<ResponseField name="request" type="Request" required>
+  A [Request](#request-object) object for calling an HTTP endpoint.
+</ResponseField>
+
+<ResponseField name="response" type="Response">
+  A [Response](#response-object) object from the HTTP endpoint.
+</ResponseField>
+
+### Request object
+
+<ResponseField name="url" type="string">
+  The URL to send the request to. You must have a corresponding
+  [Host](/define-hosts) in your project's manifest.
+
+Example: https://example.com/whatever
+
+</ResponseField>
+<ResponseField name="method" type="string">
+  The HTTP method to use for the request.
+
+Example: "GET"
+
+</ResponseField>
+<ResponseField name="headers" type="Headers">
+  The HTTP headers for the request, as [defined in your
+  manifest](../define-hosts).
+</ResponseField>
+<ResponseField name="body" type="ArrayBuffer">
+  The raw HTTP request body, as an `ArrayBuffer`.
+</ResponseField>
+
+### Response object
+
+<ResponseField name="status" type="u16">
+  The HTTP response status code. Example: 200
+</ResponseField>
+<ResponseField name="statusText" type="string">
+  The HTTP response status text. Example: "OK"
+</ResponseField>
+<ResponseField name="headers" type="Headers">
+  The HTTP response headers.
+</ResponseField>
+<ResponseField name="body" type="ArrayBuffer">
+  The raw HTTP response body, as an `ArrayBuffer`.
+
+  <Tip>
+    The response body isn't normally read directly. Instead, use `response.text()` or `response.json<T>()` functions.
+  </Tip>
+</ResponseField>
diff --git a/sdk/inference.mdx b/sdk/inference.mdx
@@ -35,15 +35,41 @@ generateText (
   Queries for response within the context of the given instruction.
 </ResponseField>
 
-### computeClassificationLabels
+### classifyText
+
+Invoke a fine-tuned classification model with a text input, resulting in a single classification label that exceeds the provided threshold.
+
+```TypeScript
+classifyText (
+  modelName: string,
+  text: string,
+  threshold: f32
+): string
+```
+
+<ResponseField name="modelName" type="string" required>
+  Internal name of your model, as [defined in your manifest](../define-models).
+</ResponseField>
+
+<ResponseField name="text" type="string" required>
+  Text input for classification amongst labels defined in the fine-tuning
+  process.
+</ResponseField>
+
+<ResponseField name="threshold" type="f32" required>
+  Confidence factor for classification. If no labels exist with a probability
+  higher than this threshold, the model won't produce a result.
+</ResponseField>
+
+### getClassificationLabelsForText
 
 Invoke a fine-tuned classification model with a text input, resulting in an array of labels and probabilities.
 
 ```TypeScript
-computeClassificationLabels (
+getClassificationLabelsForText (
   modelName: string,
   text: string
-): Map<string, number>
+): Map<string, f32>
 ```
 
 <ResponseField name="modelName" type="string" required>
@@ -55,15 +81,59 @@ computeClassificationLabels (
   process.
 </ResponseField>
 
-### embedText
+### getClassificationLabelsForTexts
 
-Invoke an embedding model with a text input, resulting in a vector embedding.
+Invoke a fine-tuned classification model with multiple text inputs, resulting in an array of an array of labels and probabilities.
+
+```TypeScript
+getClassificationLabelsForTexts (
+  modelName: string,
+  texts: Map<string, string>
+): Map<string, Map<string, f32>>
+```
+
+<ResponseField name="modelName" type="string" required>
+  Internal name of your model, as [defined in your manifest](../define-models).
+</ResponseField>
+
+<ResponseField name="texts" type="Map<string, string>" required>
+  Text inputs for classification amongst labels defined in the fine-tuning
+  process.
+</ResponseField>
+
+### getClassificationProbability
+
+Invoke a fine-tuned classification model with a text input, resulting in the probability of the provided label.
 
 ```TypeScript
-embedText (
+getClassificationProbability (
   modelName: string,
   text: string,
-): number
+  label: string
+): f32
+```
+
+<ResponseField name="modelName" type="string" required>
+  Internal name of your model, as [defined in your manifest](../define-models).
+</ResponseField>
+
+<ResponseField name="text" type="string" required>
+  Text input for classification based on the label provided.
+</ResponseField>
+
+<ResponseField name="label" type="string" required>
+  The label to classify the provided text against.
+</ResponseField>
+
+### getTextEmbedding
+
+Invoke an embedding model with a text input, resulting in a vector embedding.
+
+```TypeScript
+getTextEmbedding (
+  modelName: string,
+  text: string
+): f64[]
 ```
 
 <ResponseField name="modelName" type="string" required>
@@ -73,3 +143,22 @@ embedText (
 <ResponseField name="text" type="string" required>
   Text input for embedding.
 </ResponseField>
+
+### getTextEmbeddings
+
+Invoke an embedding model with multiple text inputs, resulting in an array of vector embeddings.
+
+```TypeScript
+getTextEmbeddings (
+  modelName: string,
+  texts: Map<string, string>
+): Map<string, f64[]>
+```
+
+<ResponseField name="modelName" type="string" required>
+  Internal name of your model, as [defined in your manifest](../define-models).
+</ResponseField>
+
+<ResponseField name="texts" type="Map<string, string>" required>
+  Text inputs for embedding.
+</ResponseField>
diff --git a/sdk/search.mdx b/sdk/search.mdx
@@ -9,4 +9,4 @@ description: "Add search into your functions"
   Functions SDK even more powerful for your next use case!
 </Tip>
 
-The [embedText](/sdk/inference#embedtext) API creates text embeddings for use in semantic similarity search use cases.
+The [getTextEmbedding](/sdk/inference#getTextEmbedding) API creates text embeddings for use in semantic similarity search use cases.
diff --git a/styles/config/vocabularies/general/accept.txt b/styles/config/vocabularies/general/accept.txt
@@ -2,9 +2,16 @@
 
 APIs
 inferencing
+repo
 
 invokeGraphqlApi
+fetch
 generateText
-computeClassificationLabels
-embedText
-log
+classifyText
+getClassificationLabelsForText
+getClassificationLabelsForTexts
+getClassificationProbability
+getTextEmbedding
+getTextEmbeddings
+log
+GQLResponse