Support for OpenAI Compatible API embedding model

README.md

@@ -87,6 +87,15 @@ In separate terminal windows from project root:
 
 On first boot and visiting of the homepage, you will be automatically redirected to create your primary admin account, organization, and database connection.
 
+
+### Support for OpenAI Compatible API embedding model.
+Use below variables in your server or docker .env settings.
+```
+OPENAI_BASE_PATH="https://api.openai.com/v1"
+OPENAI_MODEL_NAME="text-embedding-ada-002"
+MODEL_DIMENSIONS=1536
+```
+
 ## Contributing
 - create issue
 - create PR with branch name format of `<issue number>-<short name>`

backend/utils/openAi/index.js

@@ -1,16 +1,23 @@
 const { Configuration, OpenAIApi } = require("openai");
+
 class OpenAi {
   constructor(apiKey = "") {
-    const config = new Configuration({ apiKey });
+    const basePath = process.env.OPENAI_BASE_PATH || "https://api.openai.com/v1";
+    const modelName = process.env.OPENAI_MODEL_NAME || "text-embedding-ada-002";
+    const config = new Configuration({
+      apiKey,
+      basePath,
+    });
     const openai = new OpenAIApi(config);
     this.openai = openai;
+    this.modelName = modelName;
   }
 
   async embedTextChunk(textChunk = "") {
     const {
       data: { data },
     } = await this.openai.createEmbedding({
-      model: "text-embedding-ada-002",
+      model: this.modelName,
       input: textChunk,
     });
     return data.length > 0 && data[0].hasOwnProperty("embedding")
@@ -22,7 +29,7 @@ class OpenAi {
     const {
       data: { data },
     } = await this.openai.createEmbedding({
-      model: "text-embedding-ada-002",
+      model: this.modelName,
       input: chunks,
     });
     return data.length > 0 &&

docker/.env.example

@@ -5,4 +5,9 @@ JWT_SECRET="your-random-string-here"
 INNGEST_EVENT_KEY="background_workers"
 INNGEST_SIGNING_KEY="random-string-goes-here"
 INNGEST_LANDING_PAGE="true"
-DATABASE_CONNECTION_STRING="postgresql://vectoradmin:[email protected]:5433/vdbms"
+DATABASE_CONNECTION_STRING="postgresql://vectoradmin:[email protected]:5433/vdbms"
+
+## Support for OpenAI Compatible API model support.
+# OPENAI_BASE_PATH="https://api.openai.com/v1"
+# OPENAI_MODEL_NAME="text-embedding-ada-002"
+# MODEL_DIMENSIONS=1536

workers/.env.example

@@ -1,3 +1,9 @@
 INNGEST_EVENT_KEY="background_workers"
 INNGEST_SIGNING_KEY=""
-# DISABLE_TELEMETRY="true" # Uncomment to disable telemetry on workers for any ENV.
+# DISABLE_TELEMETRY="true" # Uncomment to disable telemetry on workers for any ENV.
+
+## Support for OpenAI Compatible API model support.
+
+# OPENAI_BASE_PATH="https://api.openai.com/v1"
+# OPENAI_MODEL_NAME="text-embedding-ada-002"
+# MODEL_DIMENSIONS=1536

workers/functions/newWorkspace/index.js

@@ -64,7 +64,7 @@ const newWorkspaceCreated = InngestClient.createFunction(
           workspace.fname,
           {
             vectors: {
-              size: 1536, // TODO: Fixed to OpenAI models - when other embeddings exist make variable.
+              size: parseInt(process.env.MODEL_DIMENSIONS) || 1536, // TODO: Fixed to OpenAI models - when other embeddings exist make variable.
               distance: 'Cosine',
             },
           }