Add embedding support for message.content which is string/object arra…

…y type
Mintplex-Labs · Jan 28, 2025 · a5d6675 · a5d6675
1 parent 4bdd921
commit a5d6675
Show file tree

Hide file tree

Showing 8 changed files with 169 additions and 9 deletions.
diff --git a/server/utils/vectorDbProviders/astra/index.js b/server/utils/vectorDbProviders/astra/index.js
@@ -268,7 +268,27 @@ const AstraDB = {
       };
     }
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const { contextTexts, sourceDocuments } = await this.similarityResponse({
       client,
       namespace,

diff --git a/server/utils/vectorDbProviders/chroma/index.js b/server/utils/vectorDbProviders/chroma/index.js
@@ -355,7 +355,27 @@ const Chroma = {
       };
     }
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const { contextTexts, sourceDocuments } = await this.similarityResponse({
       client,
       namespace,

diff --git a/server/utils/vectorDbProviders/lance/index.js b/server/utils/vectorDbProviders/lance/index.js
@@ -400,12 +400,32 @@ const LanceDb = {
       };
     }
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const result = rerank
       ? await this.rerankedSimilarityResponse({
           client,
           namespace,
-          query: input,
+          query: textInput,
           queryVector,
           similarityThreshold,
           topN,

diff --git a/server/utils/vectorDbProviders/milvus/index.js b/server/utils/vectorDbProviders/milvus/index.js
@@ -311,7 +311,27 @@ const Milvus = {
       };
     }
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const { contextTexts, sourceDocuments } = await this.similarityResponse({
       client,
       namespace,

diff --git a/server/utils/vectorDbProviders/pinecone/index.js b/server/utils/vectorDbProviders/pinecone/index.js
@@ -253,7 +253,27 @@ const PineconeDB = {
         "Invalid namespace - has it been collected and populated yet?"
       );
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const { contextTexts, sourceDocuments } = await this.similarityResponse({
       client: pineconeIndex,
       namespace,

diff --git a/server/utils/vectorDbProviders/qdrant/index.js b/server/utils/vectorDbProviders/qdrant/index.js
@@ -336,7 +336,27 @@ const QDrant = {
       };
     }
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const { contextTexts, sourceDocuments } = await this.similarityResponse({
       client,
       namespace,

diff --git a/server/utils/vectorDbProviders/weaviate/index.js b/server/utils/vectorDbProviders/weaviate/index.js
@@ -380,7 +380,27 @@ const Weaviate = {
       };
     }
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const { contextTexts, sourceDocuments } = await this.similarityResponse({
       client,
       namespace,

diff --git a/server/utils/vectorDbProviders/zilliz/index.js b/server/utils/vectorDbProviders/zilliz/index.js
@@ -304,7 +304,27 @@ const Zilliz = {
       };
     }
 
-    const queryVector = await LLMConnector.embedTextInput(input);
+    let textInput = "";
+    if (typeof input === "string" || input instanceof String) {
+      textInput = input;
+    } else if (Array.isArray(input)) {
+      textInput = input.map((x) => {
+        if (typeof x === "string" || x instanceof String) {
+          return x + "\n";
+        } else if (typeof x.text === "string" || x.text instanceof String) {
+          return x.text + "\n";
+        } else {
+          return "";
+        }
+      });
+    } else {
+      return {
+        contextTexts: [],
+        sources: [],
+        message: `Invalid query - the type of input is not string but ${typeof input} !`,
+      };
+    }
+    const queryVector = await LLMConnector.embedTextInput(textInput);
     const { contextTexts, sourceDocuments } = await this.similarityResponse({
       client,
       namespace,