[FEATURE] Google Gemini - Add support for structure JSON output (#14896)

jcortes · web-flow · commit bb0b304ee680 · 2024-12-11T09:29:45.000-05:00
diff --git a/components/google_gemini/actions/common/generate-content.mjs b/components/google_gemini/actions/common/generate-content.mjs
@@ -0,0 +1,92 @@
+import app from "../../google_gemini.app.mjs";
+import constants from "../../common/constants.mjs";
+
+export default {
+  props: {
+    app,
+    model: {
+      propDefinition: [
+        app,
+        "model",
+        () => ({
+          filter: ({
+            description,
+            supportedGenerationMethods,
+          }) => ![
+            "discontinued",
+            "deprecated",
+          ].some((keyword) => description.includes(keyword))
+            && supportedGenerationMethods?.includes(constants.MODEL_METHODS.GENERATE_CONTENT),
+        }),
+      ],
+    },
+  },
+  async additionalProps() {
+    const {
+      model,
+      responseFormat,
+    } = this;
+
+    const {
+      outputTokenLimit,
+      temperature,
+      topP,
+      topK,
+      maxTemperature,
+    } = await this.app.getModel({
+      model,
+    });
+
+    return {
+      ...(responseFormat && {
+        responseSchema: {
+          type: "string",
+          label: "Response Schema",
+          description: "Define the structure of the JSON response. Must be a valid JSON schema object. Leave empty to let Gemini determine the structure.",
+          optional: true,
+        },
+      }),
+      ...(outputTokenLimit && {
+        maxOutputTokens: {
+          type: "integer",
+          label: "Max Output Tokens",
+          description: `The maximum number of tokens to generate in the response. Eg. \`${outputTokenLimit}\`.`,
+          optional: true,
+          max: outputTokenLimit,
+        },
+      }),
+      ...(temperature && {
+        temperature: {
+          type: "string",
+          label: "Temperature",
+          description: `Controls the randomness of the generated text. Lower values make the text more deterministic, while higher values make it more random. Eg. \`${temperature}\`.${maxTemperature
+            ? ` Where max temperature is \`${maxTemperature}\`.`
+            : ""}`,
+          optional: true,
+        },
+      }),
+      ...(topP && {
+        topP: {
+          type: "string",
+          label: "Top P",
+          description: `Controls the diversity of the generated text. Lower values make the text more deterministic, while higher values make it more random. Eg. \`${topP}\`.`,
+          optional: true,
+        },
+      }),
+      ...(topK && {
+        topK: {
+          type: "integer",
+          label: "Top K",
+          description: `Controls the diversity of the generated text. Lower values make the text more deterministic, while higher values make it more random. Eg. \`${topK}\`.`,
+          optional: true,
+        },
+      }),
+      stopSequences: {
+        type: "string[]",
+        label: "Stop Sequences",
+        description: "The set of character sequences (up to 5) that will stop output generation. If specified, the API will stop at the first appearance of a `stop_sequence`. The stop sequence will not be included as part of the response.",
+        optional: true,
+      },
+    };
+  },
+};
diff --git a/components/google_gemini/actions/generate-content-from-text-and-image/generate-content-from-text-and-image.mjs b/components/google_gemini/actions/generate-content-from-text-and-image/generate-content-from-text-and-image.mjs
@@ -1,39 +1,41 @@
 import fs from "fs";
 import { ConfigurationError } from "@pipedream/platform";
-import app from "../../google_gemini.app.mjs";
+import common from "../common/generate-content.mjs";
+import utils from "../../common/utils.mjs";
 
 export default {
+  ...common,
   key: "google_gemini-generate-content-from-text-and-image",
   name: "Generate Content from Text and Image",
   description: "Generates content from both text and image input using the Gemini API. [See the documentation](https://ai.google.dev/tutorials/rest_quickstart#text-and-image_input)",
-  version: "0.1.0",
+  version: "0.1.1",
   type: "action",
   props: {
-    app,
-    model: {
-      propDefinition: [
-        app,
-        "model",
-      ],
-    },
+    ...common.props,
     text: {
       propDefinition: [
-        app,
+        common.props.app,
         "text",
       ],
     },
     mimeType: {
       propDefinition: [
-        app,
+        common.props.app,
         "mimeType",
       ],
     },
     imagePaths: {
       propDefinition: [
-        app,
+        common.props.app,
         "imagePaths",
       ],
     },
+    responseFormat: {
+      propDefinition: [
+        common.props.app,
+        "responseFormat",
+      ],
+    },
   },
   methods: {
     fileToGenerativePart(path, mimeType) {
@@ -55,6 +57,13 @@ export default {
       text,
       imagePaths,
       mimeType,
+      responseFormat,
+      responseSchema,
+      maxOutputTokens,
+      temperature,
+      topP,
+      topK,
+      stopSequences,
     } = this;
 
     if (!Array.isArray(imagePaths)) {
@@ -79,6 +88,21 @@ export default {
             ],
           },
         ],
+        ...(
+          responseFormat || maxOutputTokens || temperature || topP || topK || stopSequences?.length
+            ? {
+              generationConfig: {
+                responseMimeType: "application/json",
+                responseSchema: utils.parse(responseSchema),
+                maxOutputTokens,
+                temperature,
+                topP,
+                topK,
+                stopSequences,
+              },
+            }
+            : {}
+        ),
       },
     });
 
diff --git a/components/google_gemini/actions/generate-content-from-text/generate-content-from-text.mjs b/components/google_gemini/actions/generate-content-from-text/generate-content-from-text.mjs
@@ -1,23 +1,25 @@
-import app from "../../google_gemini.app.mjs";
+import common from "../common/generate-content.mjs";
+import utils from "../../common/utils.mjs";
 
 export default {
+  ...common,
   key: "google_gemini-generate-content-from-text",
   name: "Generate Content from Text",
   description: "Generates content from text input using the Google Gemini API. [See the documentation](https://ai.google.dev/tutorials/rest_quickstart#text-only_input)",
-  version: "0.1.0",
+  version: "0.1.1",
   type: "action",
   props: {
-    app,
-    model: {
+    ...common.props,
+    text: {
       propDefinition: [
-        app,
-        "model",
+        common.props.app,
+        "text",
       ],
     },
-    text: {
+    responseFormat: {
       propDefinition: [
-        app,
-        "text",
+        common.props.app,
+        "responseFormat",
       ],
     },
   },
@@ -26,6 +28,13 @@ export default {
       app,
       model,
       text,
+      responseFormat,
+      responseSchema,
+      maxOutputTokens,
+      temperature,
+      topP,
+      topK,
+      stopSequences,
     } = this;
 
     const response = await app.generateContent({
@@ -41,6 +50,21 @@ export default {
             ],
           },
         ],
+        ...(
+          responseFormat || maxOutputTokens || temperature || topP || topK || stopSequences?.length
+            ? {
+              generationConfig: {
+                responseMimeType: "application/json",
+                responseSchema: utils.parse(responseSchema),
+                maxOutputTokens,
+                temperature,
+                topP,
+                topK,
+                stopSequences,
+              },
+            }
+            : {}
+        ),
       },
     });
 
diff --git a/components/google_gemini/common/constants.mjs b/components/google_gemini/common/constants.mjs
@@ -1,7 +1,12 @@
 const BASE_URL = "https://generativelanguage.googleapis.com";
-const VERSION_PATH = "/v1";
+const VERSION_PATH = "/v1beta";
+
+const MODEL_METHODS = {
+  GENERATE_CONTENT: "generateContent",
+};
 
 export default {
   BASE_URL,
   VERSION_PATH,
+  MODEL_METHODS,
 };
diff --git a/components/google_gemini/common/utils.mjs b/components/google_gemini/common/utils.mjs
@@ -0,0 +1,24 @@
+import { ConfigurationError } from "@pipedream/platform";
+
+function emptyStrToUndefined(value) {
+  const trimmed = typeof(value) === "string" && value.trim();
+  return trimmed === ""
+    ? undefined
+    : value;
+}
+
+function parse(value) {
+  const valueToParse = emptyStrToUndefined(value);
+  if (typeof(valueToParse) === "object" || valueToParse === undefined) {
+    return valueToParse;
+  }
+  try {
+    return JSON.parse(valueToParse);
+  } catch (e) {
+    throw new ConfigurationError("Make sure the custom expression contains a valid object");
+  }
+}
+
+export default {
+  parse,
+};
diff --git a/components/google_gemini/google_gemini.app.mjs b/components/google_gemini/google_gemini.app.mjs
@@ -34,7 +34,11 @@ export default {
       type: "string",
       label: "Model",
       description: "The model to use for content generation",
-      async options({ prevContext: { pageToken } }) {
+      reloadProps: true,
+      async options({
+        prevContext: { pageToken },
+        filter = (model) => model,
+      }) {
         if (pageToken === null) {
           return [];
         }
@@ -46,13 +50,16 @@ export default {
             pageToken,
           },
         });
-        const options = models.map(({
-          name: value,
-          displayName: label,
-        }) => ({
-          label,
-          value,
-        }));
+
+        const options = models
+          .filter(filter)
+          .map(({
+            name: value,
+            displayName: label,
+          }) => ({
+            label,
+            value,
+          }));
 
         return {
           options,
@@ -62,6 +69,14 @@ export default {
         };
       },
     },
+    responseFormat: {
+      type: "boolean",
+      label: "JSON Output",
+      description: "Enable to receive responses in structured JSON format instead of plain text. Useful for automated processing, data extraction, or when you need to parse the response programmatically. You can optionally define a specific schema for the response structure.",
+      optional: true,
+      default: false,
+      reloadProps: true,
+    },
   },
   methods: {
     getUrl(path) {
@@ -99,7 +114,7 @@ export default {
         ? model
         : `models/${model}`;
       return this.post({
-        path: `/${pathPrefix}:generateContent`,
+        path: `/${pathPrefix}:${constants.MODEL_METHODS.GENERATE_CONTENT}`,
         ...args,
       });
     },
@@ -109,5 +124,13 @@ export default {
         ...args,
       });
     },
+    getModel({
+      model, ...args
+    } = {}) {
+      return this.makeRequest({
+        path: `/${model}`,
+        ...args,
+      });
+    },
   },
 };
diff --git a/components/google_gemini/package.json b/components/google_gemini/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@pipedream/google_gemini",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "description": "Pipedream Google Gemini Components",
   "main": "google_gemini.app.mjs",
   "keywords": [
@@ -13,6 +13,6 @@
     "access": "public"
   },
   "dependencies": {
-    "@pipedream/platform": "^3.0.0"
+    "@pipedream/platform": "^3.0.3"
   }
 }
diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml

Original file line number	Diff line number	Diff line change
`@@ -1,6 +1,6 @@`
`1`	`1`	`{`
`2`	`2`	`"name": "@pipedream/google_gemini",`
`3`		`- "version": "0.2.0",`
	`3`	`+ "version": "0.2.1",`
`4`	`4`	`"description": "Pipedream Google Gemini Components",`
`5`	`5`	`"main": "google_gemini.app.mjs",`
`6`	`6`	`"keywords": [`
`@@ -13,6 +13,6 @@`
`13`	`13`	`"access": "public"`
`14`	`14`	`},`
`15`	`15`	`"dependencies": {`
`16`		`- "@pipedream/platform": "^3.0.0"`
	`16`	`+ "@pipedream/platform": "^3.0.3"`
`17`	`17`	`}`
`18`	`18`	`}`