@ai-sdk/anthropic 2.0.0-alpha.10 → 2.0.0-alpha.11

package/CHANGELOG.md

@@ -1,5 +1,14 @@
 # @ai-sdk/anthropic
 
+## 2.0.0-alpha.11
+
+### Patch Changes
+
+- 25f3454: feat(provider/anthropic): add PDF citation support with document sources for streamText
+- Updated dependencies [c1e6647]
+  - @ai-sdk/provider@2.0.0-alpha.11
+  - @ai-sdk/provider-utils@3.0.0-alpha.11
+
 ## 2.0.0-alpha.10
 
 ### Patch Changes

package/dist/index.js

@@ -58,6 +58,29 @@ var webSearchLocationSchema = import_zod2.z.object({
   country: import_zod2.z.string(),
   timezone: import_zod2.z.string().optional()
 });
+var anthropicFilePartProviderOptions = import_zod2.z.object({
+  /**
+   * Citation configuration for this document.
+   * When enabled, this document will generate citations in the response.
+   */
+  citations: import_zod2.z.object({
+    /**
+     * Enable citations for this document
+     */
+    enabled: import_zod2.z.boolean()
+  }).optional(),
+  /**
+   * Custom title for the document.
+   * If not provided, the filename will be used.
+   */
+  title: import_zod2.z.string().optional(),
+  /**
+   * Context about the document that will be passed to the model
+   * but not used towards cited content.
+   * Useful for storing document metadata as text or stringified JSON.
+   */
+  context: import_zod2.z.string().optional()
+});
 var anthropicProviderOptions = import_zod2.z.object({
   /**
   Include reasoning content in requests sent to the model. Defaults to `true`.
@@ -235,7 +258,7 @@ async function convertToAnthropicMessagesPrompt({
   sendReasoning,
   warnings
 }) {
-  var _a, _b, _c;
+  var _a, _b, _c, _d;
   const betas = /* @__PURE__ */ new Set();
   const blocks = groupIntoBlocks(prompt);
   let system = void 0;
@@ -246,6 +269,26 @@ async function convertToAnthropicMessagesPrompt({
     const cacheControlValue = (_a2 = anthropic2 == null ? void 0 : anthropic2.cacheControl) != null ? _a2 : anthropic2 == null ? void 0 : anthropic2.cache_control;
     return cacheControlValue;
   }
+  async function shouldEnableCitations(providerMetadata) {
+    var _a2, _b2;
+    const anthropicOptions = await (0, import_provider_utils2.parseProviderOptions)({
+      provider: "anthropic",
+      providerOptions: providerMetadata,
+      schema: anthropicFilePartProviderOptions
+    });
+    return (_b2 = (_a2 = anthropicOptions == null ? void 0 : anthropicOptions.citations) == null ? void 0 : _a2.enabled) != null ? _b2 : false;
+  }
+  async function getDocumentMetadata(providerMetadata) {
+    const anthropicOptions = await (0, import_provider_utils2.parseProviderOptions)({
+      provider: "anthropic",
+      providerOptions: providerMetadata,
+      schema: anthropicFilePartProviderOptions
+    });
+    return {
+      title: anthropicOptions == null ? void 0 : anthropicOptions.title,
+      context: anthropicOptions == null ? void 0 : anthropicOptions.context
+    };
+  }
   for (let i = 0; i < blocks.length; i++) {
     const block = blocks[i];
     const isLastBlock = i === blocks.length - 1;
@@ -299,6 +342,12 @@ async function convertToAnthropicMessagesPrompt({
                       });
                     } else if (part.mediaType === "application/pdf") {
                       betas.add("pdfs-2024-09-25");
+                      const enableCitations = await shouldEnableCitations(
+                        part.providerOptions
+                      );
+                      const metadata = await getDocumentMetadata(
+                        part.providerOptions
+                      );
                       anthropicContent.push({
                         type: "document",
                         source: part.data instanceof URL ? {
@@ -309,6 +358,11 @@ async function convertToAnthropicMessagesPrompt({
                           media_type: "application/pdf",
                           data: (0, import_provider_utils2.convertToBase64)(part.data)
                         },
+                        title: (_b = metadata.title) != null ? _b : part.filename,
+                        ...metadata.context && { context: metadata.context },
+                        ...enableCitations && {
+                          citations: { enabled: true }
+                        },
                         cache_control: cacheControl
                       });
                     } else {
@@ -326,7 +380,7 @@ async function convertToAnthropicMessagesPrompt({
               for (let i2 = 0; i2 < content.length; i2++) {
                 const part = content[i2];
                 const isLastPart = i2 === content.length - 1;
-                const cacheControl = (_b = getCacheControl(part.providerOptions)) != null ? _b : isLastPart ? getCacheControl(message.providerOptions) : void 0;
+                const cacheControl = (_c = getCacheControl(part.providerOptions)) != null ? _c : isLastPart ? getCacheControl(message.providerOptions) : void 0;
                 const toolResultContent = part.content != null ? part.content.map((part2) => {
                   var _a2;
                   switch (part2.type) {
@@ -376,7 +430,7 @@ async function convertToAnthropicMessagesPrompt({
           for (let k = 0; k < content.length; k++) {
             const part = content[k];
             const isLastContentPart = k === content.length - 1;
-            const cacheControl = (_c = getCacheControl(part.providerOptions)) != null ? _c : isLastContentPart ? getCacheControl(message.providerOptions) : void 0;
+            const cacheControl = (_d = getCacheControl(part.providerOptions)) != null ? _d : isLastContentPart ? getCacheControl(message.providerOptions) : void 0;
             switch (part.type) {
               case "text": {
                 anthropicContent.push({
@@ -525,6 +579,40 @@ function mapAnthropicStopReason({
 }
 
 // src/anthropic-messages-language-model.ts
+function processPageLocationCitation(citation, citationDocuments, generateId3, onSource) {
+  if (citation.type === "page_location") {
+    const source = createCitationSource(
+      citation,
+      citationDocuments,
+      generateId3
+    );
+    if (source) {
+      onSource(source);
+    }
+  }
+}
+function createCitationSource(citation, citationDocuments, generateId3) {
+  var _a;
+  const documentInfo = citationDocuments[citation.document_index];
+  if (!documentInfo) {
+    return null;
+  }
+  return {
+    type: "source",
+    sourceType: "document",
+    id: generateId3(),
+    mediaType: documentInfo.mediaType,
+    title: (_a = citation.document_title) != null ? _a : documentInfo.title,
+    filename: documentInfo.filename,
+    providerMetadata: {
+      anthropic: {
+        citedText: citation.cited_text,
+        startPageNumber: citation.start_page_number,
+        endPageNumber: citation.end_page_number
+      }
+    }
+  };
+}
 var AnthropicMessagesLanguageModel = class {
   constructor(modelId, config) {
     this.specificationVersion = "v2";
@@ -722,9 +810,29 @@ var AnthropicMessagesLanguageModel = class {
     var _a, _b, _c;
     return (_c = (_b = (_a = this.config).transformRequestBody) == null ? void 0 : _b.call(_a, args)) != null ? _c : args;
   }
+  extractCitationDocuments(prompt) {
+    const isCitationEnabled = (part) => {
+      var _a, _b;
+      const anthropic2 = (_a = part.providerOptions) == null ? void 0 : _a.anthropic;
+      const citationsConfig = anthropic2 == null ? void 0 : anthropic2.citations;
+      return (_b = citationsConfig == null ? void 0 : citationsConfig.enabled) != null ? _b : false;
+    };
+    return prompt.filter((message) => message.role === "user").flatMap((message) => message.content).filter(
+      (part) => part.type === "file" && part.mediaType === "application/pdf" && isCitationEnabled(part)
+    ).map((part) => {
+      var _a;
+      const filePart = part;
+      return {
+        title: (_a = filePart.filename) != null ? _a : "Untitled Document",
+        filename: filePart.filename,
+        mediaType: filePart.mediaType
+      };
+    });
+  }
   async doGenerate(options) {
     var _a, _b, _c, _d, _e;
     const { args, warnings, betas, jsonResponseTool } = await this.getArgs(options);
+    const citationDocuments = this.extractCitationDocuments(options.prompt);
     const {
       responseHeaders,
       value: response,
@@ -746,6 +854,16 @@ var AnthropicMessagesLanguageModel = class {
         case "text": {
           if (jsonResponseTool == null) {
             content.push({ type: "text", text: part.text });
+            if (part.citations) {
+              for (const citation of part.citations) {
+                processPageLocationCitation(
+                  citation,
+                  citationDocuments,
+                  this.generateId,
+                  (source) => content.push(source)
+                );
+              }
+            }
           }
           break;
         }
@@ -852,6 +970,7 @@ var AnthropicMessagesLanguageModel = class {
   }
   async doStream(options) {
     const { args, warnings, betas, jsonResponseTool } = await this.getArgs(options);
+    const citationDocuments = this.extractCitationDocuments(options.prompt);
     const body = { ...args, stream: true };
     const { responseHeaders, value: response } = await (0, import_provider_utils3.postJsonToApi)({
       url: this.buildRequestUrl(true),
@@ -1036,6 +1155,13 @@ var AnthropicMessagesLanguageModel = class {
                     return;
                   }
                   case "citations_delta": {
+                    const citation = value.delta.citation;
+                    processPageLocationCitation(
+                      citation,
+                      citationDocuments,
+                      generateId3,
+                      (source) => controller.enqueue(source)
+                    );
                     return;
                   }
                   default: {
@@ -1104,7 +1230,26 @@ var anthropicMessagesResponseSchema = import_zod3.z.object({
     import_zod3.z.discriminatedUnion("type", [
       import_zod3.z.object({
         type: import_zod3.z.literal("text"),
-        text: import_zod3.z.string()
+        text: import_zod3.z.string(),
+        citations: import_zod3.z.array(
+          import_zod3.z.discriminatedUnion("type", [
+            import_zod3.z.object({
+              type: import_zod3.z.literal("web_search_result_location"),
+              cited_text: import_zod3.z.string(),
+              url: import_zod3.z.string(),
+              title: import_zod3.z.string(),
+              encrypted_index: import_zod3.z.string()
+            }),
+            import_zod3.z.object({
+              type: import_zod3.z.literal("page_location"),
+              cited_text: import_zod3.z.string(),
+              document_index: import_zod3.z.number(),
+              document_title: import_zod3.z.string().nullable(),
+              start_page_number: import_zod3.z.number(),
+              end_page_number: import_zod3.z.number()
+            })
+          ])
+        ).optional()
       }),
       import_zod3.z.object({
         type: import_zod3.z.literal("thinking"),
@@ -1243,13 +1388,23 @@ var anthropicMessagesChunkSchema = import_zod3.z.discriminatedUnion("type", [
       }),
       import_zod3.z.object({
         type: import_zod3.z.literal("citations_delta"),
-        citation: import_zod3.z.object({
-          type: import_zod3.z.literal("web_search_result_location"),
-          cited_text: import_zod3.z.string(),
-          url: import_zod3.z.string(),
-          title: import_zod3.z.string(),
-          encrypted_index: import_zod3.z.string()
-        })
+        citation: import_zod3.z.discriminatedUnion("type", [
+          import_zod3.z.object({
+            type: import_zod3.z.literal("web_search_result_location"),
+            cited_text: import_zod3.z.string(),
+            url: import_zod3.z.string(),
+            title: import_zod3.z.string(),
+            encrypted_index: import_zod3.z.string()
+          }),
+          import_zod3.z.object({
+            type: import_zod3.z.literal("page_location"),
+            cited_text: import_zod3.z.string(),
+            document_index: import_zod3.z.number(),
+            document_title: import_zod3.z.string().nullable(),
+            start_page_number: import_zod3.z.number(),
+            end_page_number: import_zod3.z.number()
+          })
+        ])
       })
     ])
   }),